0:00 We manage the paid media for one of the 0:01 largest fashion portfolios in Australia, 0:03 including Kougy, Blue Illusion, Tony 0:05 Biano, Pepper Mayo, Steve Madden, 2XU, 0:08 Caliber, Zoe Craftsman, and another 5 to 0:11 10 8 figure fashion brands. That 0:13 includes 9 figure omni channel 0:14 retailers, 8 figure omni channel 0:16 retailers, eight and nine figure pure 0:18 play direct to consumer only. We have 9 0:20 figure direct to consumer with a focus 0:22 in the US. We have 8 figure direct to 0:25 consumer with focuses in the UK and 0:26 Europe. Then we have a bunch that are 0:28 focused here in Australia working with 0:30 so many brands. I can tell you with 0:32 confidence that most fashion brands are 0:34 running their paid media completely 0:35 wrong. Not because their agencies are 0:37 incompetent, but because they're 0:38 applying generic e-commerce playbooks to 0:41 a vertical that doesn't work that way. 0:42 This is the video that I wish existed 0:44 when we first started managing fashion 0:46 accounts. Everything I'm about to share 0:48 comes from real audits, real accounts, 0:50 and real money spent on the platforms. 0:52 The core difference in fashion and every 0:54 other e-commerce vertical is that you're 0:56 appealing to emotion and you're not 0:58 agitating on a problem most of the time. 1:00 In supplements, you end up identifying a 1:02 health problem. You present a solution. 1:04 Whereas in fashion, the purchase is 1:06 typically driven through aspiration, 1:08 identity, and how the product makes 1:10 someone feel. And so this dichotomy ends 1:12 up existing within fashion that there 1:14 are brands that go way too far in one 1:16 direction, which is pretty rare, and 1:17 then the others go way too far in this 1:19 direction. And so you see these new fast 1:21 scaling startups like House, like 1:23 Comfort, I would almost put True Classic 1:26 in here as well. These were all fashion 1:28 brands that were built off of a focus on 1:31 function through performance marketing 1:32 ads. But the reality is is that most 1:34 fashion brands sit all the way over here 1:36 on this end of the spectrum where all 1:38 their content is just campaign shoots, 1:40 which honestly aren't very friendly to 1:42 the platforms. And you're not going to 1:43 drive scale through meta just by having 1:46 campaign shoots constantly coming in. 1:47 And so the brands that are winning right 1:48 now, the brands that are scaling very 1:50 fast are typically the ones that sit 1:53 somewhere in the middle. Now, it's not 1:54 to say that you can't lean all the way 1:56 into function, but I would argue even 1:58 all of these guys sit in the middle, 1:59 which is that they lean into the 2:00 functionality of the clothing within all 2:03 of their marketing, or at least a big 2:05 portion of their marketing, but they 2:07 have this brand component, too, that 2:09 makes people want to have association 2:11 with the actual product. There are seven 2:13 core structural differences in fashion 2:15 that's worth calling out right at the 2:16 start of this video. Number one is 2:18 generally you're going to be appealing 2:20 to emotion rather than a problem. Now, 2:22 this isn't always necessarily true. For 2:23 example, you can make a piece of 2:25 creative that's going to agitate on a 2:27 specific problem like having a day dress 2:29 that turns into a night dress when 2:31 you're traveling. Okay? So, you are 2:33 problem agitating on a problem that 2:35 someone probably has. But ultimately 2:37 what we're going to appeal to in that 2:39 creative is more so the emotional 2:41 feeling of being able to have confidence 2:44 through not having to change clothes 2:45 throughout the day or whatever the angle 2:46 might be. So generally we're leaning 2:48 hard into emotion in the creative rather 2:49 than problems which lends itself to an 2:51 entire different creative approach and 2:53 creative strategy which we will dive 2:54 into in this video and I'll show you 2:56 exactly how you need to be approaching 2:57 creative development and design. Number 2:59 two is seasonality. Seasonality is 3:01 structurally built into all fashion 3:03 brands. Okay? You need to constantly 3:04 rotate through clothes depending on the 3:06 season. This becomes even more complex 3:07 when you have a northern hem and 3:08 southern hem component of the business. 3:10 Number three is returns. Returns 3:13 significantly skew every core metric 3:15 that you look at in the business. A 3:17 rorowaz number, although rorowaz is not 3:19 a great number to use, a rorowaz number 3:21 in a fashion ad account is very 3:23 different from a rorowaz number in any 3:25 other ad account. And the reason being 3:27 is that this fashion brand might have 3:29 50% returns. So this substantially 3:31 changes the actual rorowaz number that 3:33 we're seeing in the platform. We thought 3:35 we drove 10 purchases or five of them 3:37 will get returned. So we're actually 3:38 driving half of the impact. So our 3:40 acquisition me profit contribution all 3:42 of these core metrics that I'll run you 3:44 through in this video. They all need to 3:46 be adjusted for returns. We then have 3:48 the size curve. The size curve ends up 3:51 killing performance silently because if 3:52 a few of your best selling sizes go out 3:55 of stock, your ad performance tanks 3:57 overnight. Shopping and meta performance 3:59 will always suddenly decrease because 4:02 high sell to rate sizes are gone. We 4:04 then of cross purchasing. This occurs 4:06 more in fashion than any other vertical 4:08 which is someone clicks on an ad of 4:10 let's say a t-shirt and then they go and 4:12 they browse around on the website and 4:13 then they end up buying a dress. Now we 4:15 think that the ad of the shirt or 4:17 whatever it was got us the purchase and 4:19 it actually drove a purchase of the 4:20 shirt but it did nothing like that. It 4:22 went and drove something in a completely 4:24 different category. Happens a lot in 4:26 fashion. Number six is omni channel. 4:28 Most large fashion retailers at least in 4:30 Australia have an omni channel presence 4:32 and it's due to the nature of the 4:34 product. Generally people want to try 4:36 the product on. So therefore people are 4:38 always going to want to purchase and 4:40 have an in-person experience. Now this 4:42 creates even further complexity within 4:45 the strategy because we need to factor 4:46 in dduplication of instore customers 4:49 verse online. We need to understand what 4:51 is the actual impact of paid on revenue 4:54 realization in the store. Because as you 4:56 go up higher in the funnel in terms of 4:58 the channel that you're advertising on, 5:00 let's say you go all the way up to 5:01 something like YouTube or TV, you will 5:02 have more revenue capture in your omni 5:05 channel places than in.com. So you won't 5:08 see all your YouTube ad spend just 5:10 result in revenue. You will end up 5:12 seeing this captured on marketplaces 5:14 that you're on, on retail stores that 5:15 you have. anywhere else that people can 5:17 buy your products, you will end up 5:18 seeing a lift in revenue. And then 5:20 number seven is inventory and the 5:23 balance sheet play nature of fashion, 5:26 which is that unlike really any other 5:28 business model, there is so much 5:29 turnover of product in fashion that you 5:31 pretty much always inevitably end up 5:34 with grade C inventory. Inventory that's 5:36 just not moving or is moving very slowly 5:38 to that the days on inventory is like 5:40 300 days. And so as a product of this, 5:43 you do need to factor in either organic 5:45 strategies to be able to move this grade 5:47 C inventory. But more applicable here is 5:50 to tie this into the paid media strategy 5:52 to where we may want to push grade C and 5:54 grade D inventory through paid to be 5:57 able to get it out of the business and 5:58 get cash back. But it's going to be a 6:00 low rorowaz exercise. It's not going to 6:02 be profitable, but it's going to improve 6:04 the health of the business. And so this 6:05 is where you start playing the game of 6:08 do we want to optimize for profit 6:09 necessarily in the next week or do we 6:11 want to optimize for cash to put the 6:13 business in a healthier position. Now if 6:14 we just zoom in on this for a second. 6:16 This concept of a balance sheet play 6:18 gets really interesting when you take 6:20 into consideration the virality 6:21 coefficient in fashion which is that in 6:24 fashion there are trends and products 6:25 usually go viral. It gets seen 6:27 everywhere. Influencers wear it. It 6:28 blows up on organic. And when you look 6:30 at the ad account and you look at the 6:32 spend distribution across all of your 6:33 different ads and all of the different 6:35 products in here, what you end up seeing 6:36 is one particular ad with this product 6:38 is crushing it. Might have a four times 6:40 higher efficiency than all of the other 6:42 ads. And so what naturally happens is 6:44 all of the spend goes here. And the 6:46 agency redirects additional spend here 6:49 and says, "Hey, make more creative for 6:50 this product that's doing so well." But 6:52 this product probably would have sold 6:55 anyway. and we are just eroding the 6:57 contribution margin on that purchase 7:00 order. So, let me actually run you 7:01 through an example. Let's say that you 7:03 had product A, product B, and product C. 7:06 You bought $10,000 units of each. And 7:09 then let's say that you have $100,000 in 7:13 budget to spend this month to sell 7:15 through these products. Now, product A 7:17 has an incredible natural sell-through 7:19 rate. The product has gone viral to some 7:21 degree, and so sellrough rates are 7:23 incredibly high. Maybe if we're looking 7:24 at monthly sell through rates, you're 7:26 sitting at something like 55%. So in 7:29 terms of days on inventory, how long 7:30 will it take to sell through this 7:32 product, maybe it's about 50 days? Then 7:35 if we look at days on inventory of the 7:37 next product, this product isn't selling 7:40 through very quickly. And so we're 7:41 looking at maybe 75 days. And then the 7:44 bottom product was a flop, isn't doing 7:46 well at all, and maybe we're sitting at 7:48 240 days of inventory. Now, as I said 7:50 before, what will naturally happen 7:52 within the ad account is that the agency 7:55 or the inhouse performance marketer will 7:57 go and put all of the spend here because 7:59 it will have the highest rorowaz. So, if 8:02 we just go and make a rorowaz column 8:03 here as well, this is probably going to 8:04 have something like a 7x. This might 8:06 have something like a 4.4x. And down 8:09 here down the bottom, this might have a 8:11 2.5x. And so, naturally, all the spend 8:13 will go here. So, what happens when you 8:15 go and just start putting 90k of ad 8:17 spend? So let's say 90% of this spend 8:20 goes up here. Well, the days on 8:22 inventory goes down because we increase 8:24 sellrough rate. We start selling a bunch 8:26 of this product and maybe days on 8:28 inventory after a couple days drops to 8:30 something like 9. Now days on inventory 8:32 here remains relatively the same. 8:34 Obviously a few days pass. So let's take 8:36 a little bit off it. But we're not 8:37 really getting much sellrough rate cuz 8:38 we're not putting really any budget 8:39 here. Now you can probably tell what the 8:41 issue is is that we just took all of our 8:43 media budget and put it behind a product 8:45 that would have sold anyway. it would 8:47 have just taken 50 days and we now 8:49 compressed it down to nine. If we now 8:51 look at the contribution margin on this 8:54 particular purchase order, let's say 8:56 that this was like a million in revenue, 8:58 50% gross margin to $500,000 in gross 9:01 profit. We had 500K in gross profit. And 9:05 in case these numbers aren't making 9:07 sense to you, I'd recommend going and 9:08 watching our finance e-commerce video, 9:10 which is a 1 and a half hour master 9:11 class like this that breaks down 9:13 financial metrics. 500,000 gross profit. 9:15 And then if we go to contribution 9:17 margin, we just need to minus off all 9:18 this ad spend, which is 410 because we 9:21 wanted to spend about 90k here. So the 9:22 main point is that we were going to make 9:24 500k if we didn't spend on ads. We spent 9:26 on ads, we sped up how quickly we sold 9:29 the product. Amazing. But now we make 9:30 less money. So why did we even do this 9:32 in the first place? And this is the 9:33 dilemma of fashion because wouldn't we 9:36 have just been better off actually 9:38 putting no ad spend behind that winning 9:40 product, made sure that we made half a 9:42 million dollars in contribution margin, 9:44 and we didn't actually erode this margin 9:46 at all. And instead, we go and put all 9:49 of our ad spend here and here. Not good 9:51 returns, but these products aren't going 9:53 to sell. We need them to sell or else 9:55 we'll have to take them to sale, which 9:57 is a bad look in itself. So, we have our 10:00 own disadvantages of taking products to 10:02 sale. And then if we do this like yes 10:04 let's say the rorowes is bad it's at a 10:05 4.4 and a 2.5 but it will drop the days 10:08 on inventory down. So rather than days 10:11 on inventory being 75 maybe this dynamic 10:13 changes to 30 meaning we will actually 10:15 sell through this product. Now this 10:16 product maybe it's terrible. Okay even 10:18 with ads it's performing at a terrible 10:20 row as people aren't buying it but we 10:21 managed to drop the days on inventory to 10:23 100. Well amazing because now when we do 10:25 take this product to sale if we decide 10:27 to there's hopefully not much product 10:29 left and we've done a good job at moving 10:31 most of it. Now, the contribution margin 10:32 here is probably going to be $0. 10:34 Contribution margin here is probably not 10:36 going to be a lot. Maybe it's like 50K. 10:38 But overall, we're in a much better 10:40 position because if we didn't do this, 10:42 these products wouldn't have sold 10:43 anywhere. And so, our contribution 10:44 margin would have actually been negative 10:47 because we would have had to have taken 10:48 these products to sale and likely taken 10:50 a loss on them, which would have brought 10:51 down the contribution margin of this PO. 10:54 And now this will always be the case 10:55 forever in every fashion ad account, not 10:57 just at a product level, but at an ad 10:59 level, at every single level in the 11:00 business, because this is fundamentally 11:02 just Pareto's principle, which is that 11:04 20% of whatever we're talking about 11:06 drives 80% of results. And so 20% of 11:09 products will always drive 80% of 11:11 revenue or 80% of performance. 20% of 11:14 ads will always drive 80% of 11:15 performance. And where this gets a lot 11:17 crazier is that Pareto's principle 11:19 applies into itself. And so of the top 11:21 20%, the top 20% of the 20% drives 80% 11:25 of the 80%. Right? And so what this ends 11:27 up equaling out to when you times them 11:28 together is that 4% of products, 4% of 11:31 ads drive 64% of total revenue. This is 11:34 really critical to understand because 11:36 you will have always have majority of 11:38 your ad spend over half of your ad spend 11:40 pull into just a couple ads within the 11:42 and so you want to be sure that if you 11:44 are going to go and push that much spend 11:45 behind an ad you want enough days on 11:47 inventory to be able to actually support 11:49 it or are you just eroding the 11:51 contribution margin on that PO and 11:53 there's no point in even pushing it 11:54 anywhere. Another dynamic that exists 11:56 within fashion is surfacing products at 11:58 scale. A large multi-figure retailer 12:00 that we were talking to put this really 12:01 well. They called it the trench coat 12:03 problem. And this is that they sell 12:05 trench coats, but nobody even knows 12:07 about it because they might have 2,000 12:09 products active on the website at all 12:11 times and people aren't navigating the 12:13 website at least in enough detail to 12:15 ever know that they sell trench coat. So 12:17 ultimately, how do we use paid media as 12:19 a vehicle to be able to drive up 12:20 awareness across the entire product 12:22 portfolio, especially when 80% of spend 12:25 will only go into 20% of these products? 12:27 So Prito's principle will automatically 12:29 put you on the back foot because it will 12:31 want to be distributing spend to just a 12:33 few products and just those products 12:34 will surface onto Facebook, onto Google, 12:37 etc. So how do we structure the accounts 12:39 in a way that we can better surface 12:41 products with lower sellrough rates and 12:42 I'll tell you how to do it later in the 12:44 video. So the structure from here is 12:46 we'll dive into the financial model. 12:48 What makes fashion different from KPIs 12:50 to how you need to think through 12:51 financial modeling and growth? We'll 12:53 then go into measurement and attribution 12:54 and how this is fundamentally unique in 12:56 the fashion niche. We'll then go into 12:58 account structure. I'll show you how to 12:59 think through account structure on Meta 13:01 and Google and what works best. We'll 13:02 then dive into creative strategy. We'll 13:04 then go into seasonal strategy. We'll 13:06 then go into how to scale particularly 13:07 into international markets. We'll then 13:09 wrap it all into the playbook by ad 13:11 spend level. So based on where your 13:13 current ad spend is, you will have core 13:15 action items by the end. And then 13:16 lastly, we're going to do some live ad 13:18 reviews. I want to show you ads that 13:19 actually work well in fashion so that 13:21 you can understand the core components 13:23 within ads that makes them do well. So 13:25 with that being said, let's dive into 13:26 the financial model. So the best metric 13:29 for indexing the performance of paid 13:31 media in fashion is going to be a 180day 13:35 LTGP to capac. And the reason for this 13:37 is that gross profit growth within the 13:39 first 180 days actually changes a lot 13:42 brand to brand. Now on average, what you 13:44 want to be KPIing against to give you 13:46 some retention benchmarks is at day 90 13:48 good fashion brands will see 15% lift in 13:51 gross profit from their existing 13:53 customers. Then by day 365, good fashion 13:56 brands will see about 55% left in gross 13:59 profit. If you're not hitting these 14:01 benchmarks, you need to be way more 14:02 efficient on the front end of actually 14:04 acquiring a customer. Now, a big mistake 14:06 in fashion, a very very big mistake is 14:09 to measure just against first purchase 14:12 gross profit against CAC, which is a 14:14 very common exercise. So you will look 14:16 at your average order value. Let's say 14:17 that your average order value is $200. 14:19 You'll say on average our gross profit 14:21 is 50%. So, we make $100 in gross 14:23 profit. Therefore, we want a 2LTGP to 14:26 CAC or a first purchase GP to CAC. So, 14:28 we want our cost to acquire to be $50. 14:31 Therefore, we're at a two, which is not 14:33 bad. Now, the issue here, and this is 14:35 where a lot of people make a mistake, is 14:37 that this doesn't encapsulate the 14:39 returns. And so you may have a 50% 14:42 return rate in which case your CAC is 14:45 actually 2x higher because yes you are 14:47 driving 100 customers let's say but 50 14:50 of them return their product and then 14:52 don't come back and buy a second time. 14:54 And so as a product of that your CAC 14:55 actually isn't 50. Your CAC is 100 which 14:58 means your first purchase to CAC ratio 15:00 is a one. And this is not a good 15:03 position to be in. So you always want to 15:04 extend your lifetime value slightly to 15:07 fashion because you want to encapsulate 15:09 the time period in which a return or a 15:11 refund might actually occur. Now not 15:12 only that, a big issue within fashion 15:14 actually lies in this gross profit 15:16 calculation. And what's wrong about it 15:19 is that most brands rely on taking 15:21 products to sale when they don't sell 15:23 them. And so you will have your grade C, 15:25 arguably even your grade B and C 15:28 inventory that ends up not selling 15:30 through fast enough during that 15:31 particular season. So you go and spin it 15:33 off onto a sale page and you take it to 15:35 sale to try to sell through it. Now the 15:37 issue here obviously is that all of 15:38 these sales erode your average gross 15:41 margin. And so the gross margin number 15:43 that you're using to calculate this on 15:45 is this based on the average gross 15:47 margin inclusive of all of the sale 15:49 items going on. Is this non-inclusive 15:51 and this is just your full price GP? 15:53 Like gross profit percentages get 15:55 complex once you start to introduce the 15:57 fact that you're always taking products 15:58 to sale. And so you need to understand 16:01 what the gross margin numbers are 16:02 inclusive and exclusive of all of these 16:04 items always going to sale. And then if 16:06 you don't take products to sale and you 16:08 just burn them, so you donate product or 16:10 you simply just don't sell through it 16:12 and you do something with it to get rid 16:14 of it, well then you also need to take 16:15 that into consideration as a cost 16:18 against your average GP percentage. We 16:20 then have inventory economics within 16:22 fashion. If you're actually working in a 16:24 fashion brand, you obviously know all 16:26 about this. This is more so for the 16:28 performance marketers that might be 16:29 working in house or any agencies that 16:31 might be watching this, which is that 16:32 your products fall into four different 16:34 categories of inventory. This is based 16:36 on sellrough rates. Grade A, it's 16:37 selling very quickly. You should 16:38 actually be worried about them selling 16:40 out. Grade B, it's moving slow. Grade C, 16:42 it's pretty much not moving at all. 16:43 Grade D, it's not moving at all. You're 16:45 selling none. This is like a 0% 16:47 sellthrough rate. And some people don't 16:49 even have grade D in the way that they 16:50 rank products. Now, the fundamental 16:52 thing to understand here, whether it's a 16:54 seven, eight, or nine figure business, 16:55 is that you can end up pooling a bunch 16:57 of inventory here, and you can end up 16:59 having like $10 million in inventory 17:01 that's grade C that isn't moving. And 17:03 even though the business might be cash 17:04 flowing $10 million in profit this year, 17:06 because $10 million is held in 17:08 inventory, they don't actually have this 17:09 cash. And so, it's a very terrible 17:11 position to be in. They need to figure 17:12 out how to sell through this inventory. 17:14 Now, you can often try to sell through 17:16 inventory using a paid strategy. Okay? 17:18 You can prioritize these products in a 17:20 shopping campaign. And you can 17:21 prioritize these products and make more 17:22 creative for meta. You could prioritize 17:24 them in Tik Tok on a new channel. You 17:25 could move them into a different country 17:28 based on seasonality. So there's a lot 17:30 of different stuff and there's a lot of 17:31 different strategies that you can 17:32 employ, but often this is just limited 17:33 to organic strategies rather than 17:35 actually tying in paid. Now, the 17:37 interesting dynamic to be across here is 17:38 that if you do go and start pushing 17:40 gradec inventory through ads, there is a 17:42 potential branding component that needs 17:45 to be considered, which is that if we're 17:47 just taking our bad product that people 17:48 don't want to buy and we're surfacing 17:50 that in paid ads, well, is that really 17:53 the front-end visuals that we want to 17:54 give to new customers? Is that hey, our 17:56 brand is not these excellent products 17:58 that everyone likes. Our brand is all 18:01 these terrible products. And so that is 18:02 where sometimes brands can be a bit iffy 18:04 with pushing products more aggressively 18:06 in paid. But this is where just 18:07 understanding budget allocations across 18:09 the three uh grades of inventory and 18:11 understanding that we could just 18:13 prioritize grade C and existing customer 18:15 retargeting. Like there's a lot of 18:16 different strategies that could be 18:17 employed here. But it is really 18:18 important for agencies to be across this 18:20 dynamic or else they end up just over 18:22 prioritizing winners. Another really 18:24 important component that exists within 18:26 fashion is cohort quality over time. And 18:28 so as you're acquiring customers in Jan, 18:31 Feb, March, these customers end up being 18:34 super high quality. They often have 18:35 lower return rates and they often have 18:38 much better LTV uplift. But then as you 18:40 move through into November, December, 18:43 you end up having considerably higher 18:45 return rates and considerably lower LTV 18:48 off these customers. And is because they 18:50 are discount orientated. Number one. 18:52 Number two, there's a lot of gifting 18:53 going on. And number three, is it 18:56 dependent on your returns policy? This 18:58 is the best opportunity of all time for 19:00 consumers to just buy literally 10 units 19:03 of your product, try them all on, and 19:05 the ones that they don't like, go and 19:07 refund them. And this is this whole 19:08 mentality within fashion, which is that 19:11 people will buy a whole of products, try 19:14 them all on, and then just return the 19:16 ones that they don't like and keep one 19:17 or two. And so this ends up 19:19 substantially changing what the 19:20 front-end efficiency looks like on paid 19:23 or on the business as a whole. If you're 19:25 not taking into consideration how much 19:27 returns you expect to actually occur or 19:30 if you have a returns policy that allows 19:31 for this, obviously you're in a pretty 19:33 bad position and you're going to get 19:34 taken advantage of. I have actually seen 19:36 brands in international markets with 19:39 return rates from 80 to 100% during 19:42 discount periods. So that means 100% of 19:45 their revenue gets refunded, which is 19:47 crazy. So their efficiency is a zero. 19:49 And this is because people just buy a 19:51 ton of the product and then they end up 19:52 returning it. And because of the way 19:53 that their returns policy is structured 19:55 as well as I think the positioning of 19:57 the brand in that market over time, they 19:58 have been attracting a customer that 20:00 does this a lot. And so this 20:02 substantially changes then how we need a 20:04 KPI paid media because if we're looking 20:06 at like a 4X rorowaz, we need to take 20:08 into consideration that this needs to be 20:10 discounted by 80 to 100%. This is also 20:13 really important when you are KPIing 20:15 November December and so what a lot of 20:17 agencies or internal teams will do is 20:19 let's say in Jan Feb we are happy with 20:22 an acquisition me for those that haven't 20:24 watched any of our other videos and this 20:25 is your first time acquisition me is new 20:28 customer revenue divided by ad spend so 20:30 this is a great measure of efficiency on 20:32 paid media let's say in Jan Feb your 20:34 acquisition me is a four and then in 20:36 November December your acquisition me is 20:37 a four the issue here is this is way 20:41 better than this because you're going to 20:44 have way higher return rates here that 20:46 aren't encapsulated in this front-end 20:48 metric and you're going to have worse 20:50 LTV. So, this acquisition me needs to 20:52 actually be higher during Black Friday 20:54 and typically substantially higher 20:57 dependent on what your expected changes 20:58 in return rates and LTV. And so, you 21:00 want to have these numbers leading into 21:03 this period so that you can be prepared 21:05 accordingly and so that you know what 21:06 kind of efficiency targets you want to 21:08 hit. You always need to back propagate 21:10 from the bottom of the P&L when it comes 21:12 to setting KPIs. So ultimately for this 21:14 period, what is our if we want to really 21:16 take this all the way down to the 21:17 bottom, what is our net profit target 21:19 for November, December. From there, we 21:21 then layer on our opex and we get a 21:23 contribution profit target and then from 21:25 there we now know what kind of revenue 21:28 volume do we need to hit and this will 21:30 be constrained based on inventory at 21:32 what ad spend efficiency. And then if we 21:34 know what revenue and ad spend 21:35 efficiency we need to hit, we can then 21:36 back propagate this into the actual 21:39 efficiency target in the month and 21:40 understand that it's inclusive of return 21:42 rates and the lower LTV dynamic. Another 21:45 important dynamic in the financial model 21:47 is spend towards existing customers. Now 21:49 in a lot of industries, our internal SOP 21:52 is to allocate close to 0% to existing 21:55 customers in paid media. And the reason 21:57 being is that paid media often isn't 21:58 very incremental at all at targeting 22:00 existing customers unless there's some 22:02 kind of new product launch that we want 22:04 to surface in front of them because a 22:05 lot of these people just might not see 22:07 the email or might not see the organic 22:08 socials and so we want to get it in 22:10 front of their screens. Now the unique 22:13 dynamic about fashion is obviously that 22:15 there are product drops every 1 2 3 22:18 weeks depending on the brand and so 22:19 because there is so much surfacing of 22:21 new product we do need to get it in 22:22 front of existing customers and as a 22:24 function of that paid media does become 22:27 incremental for fashion brands and we 22:30 actually learned this the hard way about 22:31 3 years ago which is that internally 22:34 we're very much so of the opinion with 22:35 all of the incrementality tests and lift 22:37 tests that we've run that spending on 22:38 existing customers is not a very good 22:40 use of capital. You don't need to spend 22:42 a lot. Most people are overspending. And 22:44 so we brought that idea into a lot of 22:46 our fashion accounts, too. Until we 22:48 started running inverse holdout tests to 22:49 be able to validate it. And what we 22:51 found out was that actually no, existing 22:53 customer spend is incremental in 22:56 fashion. So you want to be careful here, 22:57 especially if you watch any of our other 22:58 content where I always say don't target 23:00 existing customers. Don't target 23:01 existing customers. The one caveat to 23:03 that is in fashion. So the question 23:05 becomes, okay, well how much spend as a 23:07 percentage do we allocate to existing 23:09 customers? And the answer is thinking 23:11 through this problem as a percentage is 23:13 a terrible way to think about. Instead, 23:14 what you want to be doing is using this 23:16 formula. CPM times frequency times 23:19 number of existing customers that you 23:22 have. And so, just as a worked theory, 23:24 if you have $20 CPMs on targeting 23:27 existing customers, you want a frequency 23:30 over a 7-day, sorry, this is over a 23:32 30-day period. Uh, so if you want a 23:33 frequency at a six, I wouldn't 23:36 recommend, just as a side note, having a 23:38 frequency above a seven. Complete waste 23:39 of money. I also wouldn't recommend 23:41 having it below a three. And so in this 23:43 range is where you want to sit it. It's 23:44 going to be up to you. It's going to be 23:45 up to how much creative refresh you have 23:47 going into this campaign as well as what 23:48 your product drop cadence looks like. 23:50 Let's say you want a six. And then let's 23:52 say you have 500,000 existing customers. 23:55 And because this CPMs, we just need to 23:57 do a divide by a,000 at the bottom here. 24:00 And that will give us $60,000 24:03 a month. So if we want to target half a 24:05 million existing customers six times a 24:07 month and our CPMs is $20, we need to 24:09 spend 60k. And this is way higher than 24:11 most people because these CPMs are often 24:13 way lower and you can pull frequency 24:15 down a little bit. So this is the way 24:16 that you modulate existing customer 24:18 spend within a fashion brand. 24:21 Measurement and attribution within 24:23 fashion. Fashion is probably the most 24:24 notorious category other than CPG for 24:27 overattribution within the platforms. 24:29 And the reason being is that if you have 24:31 a fashion brand that's been around for a 24:32 while, your returning customer revenue 24:34 as a percentage of total revenue 24:36 generally tends to be pretty high. It's 24:38 at least 30% sometimes all the way up to 24:40 70 75% if new customer growth is slowed 24:42 down. And so because of that, you end up 24:44 with massive overattribution in the 24:46 platform simply from view through 24:47 conversions or from ads that yes serve 24:49 to people. Yes, they click, but it isn't 24:51 actually causally related to the 24:53 purchase. And so, I'll give you some 24:54 real examples of 9 figure fashion brands 24:57 that I've audited. Example, a meta 24:59 return on ad spend. When you just look 25:01 at the last 30 days when I did an audit 25:03 was 36x. Now, right away, this should be 25:06 ringing alarm bells. Like, no one's at a 25:08 36x return on Meta. It's a few dollars 25:11 in terms of a CPA. Then when you 25:12 delineate down to 7-day click row for 25:16 those that don't know how to do this in 25:18 meta if you've got it pulled up right 25:19 now you go to columns compare 25:21 attribution settings and you click on 25:22 7-day click and it will give you the 25:24 7-day click read which takes out all the 25:26 view through conversions and this number 25:27 dropped to a 17 act still pretty high. 25:30 Then do that exercise again. Columns, 25:33 compare attribution settings, but this 25:34 time click on incremental attribution. 25:37 And what this is going to do is use hold 25:38 out groups to be able to identify the 25:40 actual causal relationship of paid by 25:43 seeing if people are buying even when 25:44 they're not seeing ads. And what 25:46 happened then is the row dropped to an 25:48 8.22x. And the kicker here, why this 25:51 matters so much is that the acquisition 25:53 me in the business is an 8x. And so 25:56 suddenly you can see we build congruency 25:58 once we go down to incremental 26:00 attribution against acquisition me. This 26:02 is so over inflated and is misleading 26:05 you in terms of decision-m and so is 26:07 this. I'll give you one more example 26:08 just cuz you believe me and I could give 26:10 you 10 examples of this. This is another 26:11 multi-figure fashion brand. Their 26:13 rorowaz in the platform is a 13.6. When 26:16 you go to 7-day click it drops to an 26:18 8.6. And then once you open up 26:20 incremental attribution, it drops to a 26:24 5.16. 26:25 Their acquisition me in the business is 26:28 a 4.4. 26:30 Now you can see that if you're using 26:32 this number to make decisions, it 26:34 doesn't really make much sense when it's 26:36 3x off the actual financial reality. And 26:40 so you want to be looking for as tight 26:41 congruency as possible between the 26:43 numbers that you're actually looking at 26:44 in the platform and the numbers that are 26:46 occurring within the P&L. It's also 26:47 really critical in fashion due to the 26:49 refund and return rates that all of your 26:52 metrics and we actually had to build out 26:53 custom metrics so that this was possible 26:55 within our internal analytics software 26:57 is that we have two definitions of every 27:00 core metric that we're tracking against. 27:02 For example, with acquisition me 27:05 acquisition me and then acquisition me 27:08 adjusted. So am by itself is net new 27:11 customer revenue divided by ad spend 27:13 total ad spend across all platforms. Now 27:15 the reason why this ends up being a good 27:18 metric but not necessarily always 27:20 contextually relevant in fashion is 27:22 because of the refund. What will 27:24 actually happen in quite a few brands is 27:25 refunds will sometimes spike on a 27:27 particular day when the team decides to 27:29 process a lot of the refunds. And so as 27:30 a function of that when you look at 27:32 acquisition me over time it might look 27:34 relatively stable and then suddenly you 27:36 have this drop. And now looking at just 27:38 this metric here you look at this dip 27:40 and you try to troubleshoot and you go 27:41 okay why did this dip? What happened in 27:43 the platforms? What change happened? 27:44 what was the difference in product 27:45 sellrough rates? Did conversion rates 27:47 drop or sessions drop? Like what 27:48 actually occurred? But instead, if you 27:51 just go and look at net new customer 27:53 revenue plus adback returns, you take 27:56 returns out of the graph, what you 27:58 actually end up finding is that 27:59 acquisition me is stable and that the 28:01 kick in revenue was actually being 28:03 caused by returns being processed. And 28:05 so having both of these available to you 28:07 and having a KPI on the adjusted value 28:10 and then the unadjusted value is really 28:12 critical because you might not actually 28:14 hit your efficiency for the week, but it 28:16 might not have anything to do with the 28:17 actual efficiency you achieved on the 28:19 platform, but it might be to do with 28:20 refunds spiking because of a particular 28:23 product portfolio that's currently 28:24 getting pushed. And so having insight 28:26 into that and being able to understand 28:27 what's our unadjusted uh target and then 28:29 what's our adjusted target and then if 28:31 we hit one but not the other, we know 28:32 that returns ultimately were the issue. 28:34 And then we need to troubleshoot deeper 28:35 into what actual products led to those 28:37 returns. And then in omni which means 28:40 that you have retail stores there 28:42 becomes the problem of dduplication 28:46 which is that let's say someone goes 28:47 into a retail store of yours over here 28:50 and then they purchase and then after a 28:52 while they come to the online store. 28:55 This is a PDP and then they end up 28:56 buying. Well on Shopify this is going to 29:00 be tagged up and considered as a 29:02 firsttime customer. if we're not using 29:05 Shopify POS over here and these are 29:07 sitting on two different systems and so 29:08 this person actually gets registered as 29:10 first time twice. They get registered as 29:13 first time in the store and then first 29:15 time online. Ideally, we want to 29:17 dduplicate this so that this gets 29:19 referenced as a returning customer. 29:21 Number one, it's going to give us a more 29:22 accurate representation of our CAC and 29:24 acquisition M and all of these metrics 29:26 that are based on this first time 29:28 tagging. But number two, this becomes 29:30 important as well for targeting because 29:32 if we're targeting campaigns purely for 29:34 new customer acquisition, we don't 29:36 really want to be targeting this person. 29:38 Now, yes, there can be a strategy to try 29:40 to bring people from offline to online 29:42 for sure, but if we're trying to go 29:44 truly cold, we don't want this person 29:46 included. And so if a client has a CDP 29:49 that brings together retail customers 29:52 and online customers and we can 29:54 dduplicate what we will do is have our 29:57 regular ENCAC calculation which is just 30:00 online Shopify firsttime new customers 30:03 divided by ad spend and then we'll have 30:05 what we call our true NCAAC which is 30:08 true new customers that have been 30:09 dduplicated against offline divided by 30:12 ad spend. And what you end up finding 30:13 obviously is that the true ENCAC is much 30:16 higher than the ENCAC that we think it 30:18 is. And so we might see a $70 ENCAC when 30:21 we're measuring against Shopify's 30:22 firsttime customers, but once we 30:24 dduplicate the data, this jumps to $100. 30:26 Now, ideally, what we want to start 30:28 seeing is we want to start seeing more 30:30 true new customer acquisition and a 30:32 decrease in this number. And so we 30:33 actually have a case study that we'll 30:35 put out at some point where we did this 30:37 exact exercise and we started excluding 30:39 all of the retail new customers with the 30:41 true new customer list. And what we saw 30:43 was these two numbers converge. And so 30:45 true rank was up here at like $100 and 30:48 our ENAC down here was at $70. And then 30:51 what we saw was our ENAC started to 30:53 creep up. it was becoming more expensive 30:55 to acquire customers or it was seemingly 30:58 looking like it was more expensive. But 31:00 our true new customer came down and 31:04 actually converge to the same number 31:05 which was at around about 80. And so 31:07 when we look at omni channel dduplicated 31:09 data and then we look at just online 31:11 data, it was starting to tell the exact 31:13 same picture. The reason being is we 31:14 were able to substantially increase the 31:16 volume of true new customer acquisition 31:19 which started to blend out the 31:20 difference between these two numbers. So 31:22 for account structure in meta there's 31:23 two issues that we typically see in 31:25 audits. Number one is there's over 31:27 segmentation and this is based on the 31:28 fact that there are so many products in 31:30 fashion and so people believe that they 31:32 need a segment to the level of the 31:34 product width which ends up meaning that 31:36 accounts with like a 50k a month budget 31:38 or 100k a month budget have like 21 31:40 campaigns. They have reach campaigns, 31:42 traffic campaigns, they have all these 31:43 different products segmented. They have 31:44 their sales that take it further off 31:46 campaigns. It ends up just being so much 31:48 oversegmentation that it substantially 31:50 decreases performance. Number two, and 31:52 this isn't a mistake, this is something 31:53 you need to keep in mind, is that there 31:54 is the product launch cadence that 31:56 exists within fashion. So, how do we 31:58 constantly introduce products without 31:59 resetting learning phases, disrupting 32:01 campaigns, but then also not creating 32:03 hundreds of different campaigns. So, how 32:04 do we structure to facilitate this? 32:06 Well, at sub50k a month, I always 32:09 recommend keeping it very very simple. 32:11 This isn't enough budget to be able to 32:13 introduce heavy levels of segmentation 32:15 or complexity into the account, and 32:17 neither should you need to. And so I 32:19 would have one testing campaign. This 32:21 can be a CBO or an It honestly 32:23 doesn't matter. And there's reasons for 32:25 one, there's reasons for the other. This 32:26 is going to come down to personal 32:28 preference and how you want to structure 32:29 the ad account. But the core premise 32:31 here is that you have separate adsets 32:33 and every time you have new product 32:35 launches, it goes in its own ads set. 32:37 You can also segment this based on 32:38 concepts. You can segment it based on 32:40 margin, on product type, etc. There's a 32:42 lot of different structure that you need 32:43 to think through here, which I'll give 32:44 you more clarity on in a moment. Then 32:46 one scaling CBO. Now, this is only going 32:49 to be applicable and there's only going 32:51 to be something that you put into the 32:52 account if this is running as an and 32:54 if you can run assets for a long period 32:56 of time. A big issue in fashion is that 32:59 ads will only have one product in them. 33:01 Why that's an issue is that if this is a 33:02 really good ad, if it's a winning ad, it 33:04 does incredibly well. Well, the second 33:06 this product goes out of stock or even 33:08 like core sizes within the size curve go 33:10 out of stock, you need to go and turn 33:12 this ad off, which is not ideal because 33:14 this is a winning ad. This could 33:15 probably hold $100,000 in spend. now you 33:17 need to turn it off. And so you actually 33:19 don't really have the capability to 33:21 scale a lot of creative in fashion 33:23 accounts if it is not built to be 33:25 evergreen and run for an extended period 33:27 of time. And so there is a lot of 33:28 different elements that you need to 33:30 introduce into creative in fashion that 33:32 allows the ad to be scalable that allows 33:34 you to run it for longer and that would 33:36 even allow the introduction of a scaling 33:38 campaign to be a possibility within the 33:40 account structure. Now we'll dive into 33:42 all of this when we talk about creative 33:44 strategy. Then you have a retargeting 33:45 campaign. Typically these days this is 33:48 just existing customer retargeting. Now 33:50 website visitor retargeting will be done 33:52 in your topfunnel campaigns anyway. So 33:54 this is purely for isolating out 33:56 existing customers showing them DPAs 33:58 surfacing new arrivals in front of them 34:00 so that we can continue to turn over 34:02 returning customer revenue and have the 34:04 budget delineated accordingly. As we 34:06 move into 50 to 250k recommend multiple 34:10 different advantage plus campaigns. Now, 34:11 how you want to be thinking through the 34:13 different advantage plus campaigns and 34:14 how you want to be thinking through the 34:16 ad structure is the same as how you 34:18 would actually think through it in any 34:19 other different business. And how that 34:21 is is that at the campaign level, 34:23 campaign level segmentation should be 34:26 based on business units. And so if the 34:28 business has fundamentally different 34:30 units that it's KPIing on, it's looking 34:32 at revenue separately, it's forecasting 34:34 goals separately, then this needs to be 34:36 introduced at the campaign level so that 34:38 you can have the segmentation of budgets 34:39 and KPIs. As an example, you may have 34:42 all of the clothes within the fashion 34:44 brand and then you may have denim. You 34:46 actually end up segmenting this out and 34:47 looking at it as a separate business 34:49 unit because it drives maybe 40% of 34:51 revenue. It has maybe different LTV 34:54 dynamics since you want to look at it as 34:55 a different category entirely. And so 34:57 because of that, you'll have a campaign 34:58 for everything and then you'll have a 35:00 campaign for denim and that'll be 35:01 separated out. Same thing applies for 35:03 maybe your basics. Okay? A lot of people 35:05 don't push basics cuz they sell anyway, 35:07 which is a good idea as we talked about 35:08 earlier. You don't need to push stuff 35:10 that's doing well organically, but if 35:11 you do, you would have maybe basics in 35:13 one campaign and then you would have all 35:14 your seasonal products in another. And 35:16 so you want to be splitting campaigns 35:18 based on business units. When it comes 35:19 to the adset level, you want to be 35:21 splitting the adset level based on the 35:23 question that you want the answer to. So 35:25 the advantage of adset level structuring 35:27 is that you can get feedback at the 35:29 adset level on what's working and what's 35:31 not. And so you might do adset level 35:33 structuring based on format. So you have 35:35 adsets with UGC in it and then you have 35:36 adsets with campaign shoots. And the 35:38 idea is that you want to be able to go 35:39 back to your seuite or your marketing 35:41 your CMO or whoever's in the business 35:43 and say hey look at this data UGC 35:46 outperforms campaign shoots by 2x. Hence 35:48 we should take some of the campaign 35:49 shoot budget and we should redirect that 35:51 into user generated content whatever it 35:53 might be. Now do I think splitting by 35:54 format is the best call? No. I actually 35:56 don't think it's a very good idea at 35:58 all. But if you need that data to be 35:59 able to pass through to someone then it 36:01 is a good call. You want to go collect 36:02 that data, you want to structure like 36:03 that at the adset level, then you can 36:05 roll into a different adset structure 36:06 over time as you want to answer 36:08 different questions. So, I think one 36:09 mistake or misconception that people 36:11 have with adset level structuring is 36:13 that they think that it has to be the 36:14 same forever. Just like the campaign 36:16 level structure has to be split by 36:17 business units and don't touch it. 36:19 That's how it is. But the adset level 36:20 structure can change all the time. If 36:22 you want to learn about formats, you 36:23 split by formats. If you want to learn 36:24 about concepts, you split by concepts. 36:26 If you want to learn about the product 36:28 level targeting, you split by products. 36:29 So this ends up being very accustom to 36:32 the questions and feedback on data that 36:34 you want answered within the business. 36:35 And then as you move into 250k plus ad 36:37 spend, all that occurs here is that just 36:39 additional complexity likely exists 36:41 within the business, which then 36:43 translates into the account. And so the 36:44 core difference between spending 50k a 36:46 month and 250k a month isn't that you 36:48 build out a more complex structure 36:50 because of the level of spend. You 36:52 actually build out a more complex 36:53 structure because of the level of 36:55 revenue in the business. because the 36:57 revenue makes the business more 36:59 complicated. And so we're a 50k a month 37:01 fashion brand, we might just have 10 37:02 ads, 20 ads a week, plus there's not a 37:06 huge wide product portfolio. Maybe they 37:07 drop like 15 products every two weeks or 37:09 something like that. Then once we go to 37:11 this kind of ad spend, the brand is 37:12 dropping weekly. It's 60 to 70 products 37:16 and we're getting 200 new ads a week. 37:18 And so because of how much additional 37:20 complexity exists here, that needs to 37:22 start getting mirrored within the 37:23 account structure based on the business. 37:25 One thing you want to be very careful of 37:27 in fashion is DPA over reliance. We 37:31 actually have an entire YouTube video on 37:33 this called something like DPA of the 37:35 death spiral for fashion brands. And the 37:37 reason is is that DPAs, in case you 37:39 don't know what this is, this is dynamic 37:41 product ad. It's those cataloges where 37:43 each of the slides in the catalog is a 37:45 dynamic product that's pulling through 37:47 from your feed from the website. The 37:49 idea is it will present the product that 37:50 the person looked at on the website in 37:52 the ad. So really good retargeting ad. 37:54 Now, what will happen with DPAs is the 37:56 rorowaz will be very good. And the 37:58 rorowaz will be very good because it 38:00 sits really far at the bottom of the 38:01 funnel to where just before someone's 38:03 about to buy the DPA comes in and serves 38:05 them an ad. And so it often takes credit 38:07 for a lot of purchases. So once you take 38:09 view through conversions out of this and 38:11 you just look at 7-day click, the 38:12 rorowaz ends up dropping enormously, but 38:14 it's still pretty good. Still pretty 38:15 high. Then once you go to incremental 38:16 attribution, the rorowaz here ends up 38:18 being terrible comparative to the 38:20 original number. And so most of the time 38:22 you'll see your DPAs and you'll be like, 38:23 "Oh, it's at a 7x or it's at a 10x." But 38:25 then if you went and delineated down to 38:27 incremental attribution, the rorowise is 38:28 probably closer to a two. Now, because 38:30 most people don't do this, they end up 38:32 substantially overspending on DPAs cuz 38:34 they just put more and more and more 38:36 budget hair because it looks good on the 38:37 surface. And that's what you have to be 38:38 really careful of. I have audited 38:40 multiple fashion brands that have just 38:42 looked at DPAs and gone, "Pas are 38:43 incredible. Why are we even making more 38:45 creative? Just put more spend behind 38:46 DPAs. Put more spend behind DPA." So 38:48 they actually decrease their creative 38:50 volume. they end up putting more and 38:51 more spend into DPAs. Issue is because 38:53 DPA sit at the bottom of funnel, you 38:55 aren't scaling the business because all 38:57 of the new customer traffic that should 38:59 be coming up here, you're actually 39:00 taking budget away from it and moving it 39:02 down to a stage in the funnel which is 39:04 meant to convert this traffic. And so if 39:05 you're not generating topunnel traffic 39:07 or top ofunnel awareness in any way, you 39:09 can't scale because the funnel isn't 39:11 getting filled up, which is kind of 101 39:12 marketing, but because people are so 39:14 obsessed with just distributing spend to 39:15 the highest return on ad spend product, 39:17 they end up forgetting these things. So 39:18 now account structure for Google number 39:20 one the priority is to saturate Google 39:23 shopping before you go into Google 39:26 search. Reason being is that shopping 39:28 ads show the product image and because 39:31 fashion is very visual obviously the 39:34 product photo matters more than anything 39:35 else. It then shows the product price 39:37 down here. It shows the product title, 39:39 reviews, shipping and so the user ends 39:42 up self-qualifying themsel before they 39:44 even click on the ad. They also get five 39:47 competitors next to your listing that 39:49 they could also click on. So the fact 39:50 that they ended up clicking on your 39:52 listing self-qualifies them enormously. 39:55 And so conversion rates end up being 39:57 super high through clicks on shopping 39:59 ad. Search ads on the other hand, you 40:01 only get a few listings and it's text. 40:03 Because once again, fashion is visual. 40:05 What's text really going to do for 40:06 selling me on the garment that I haven't 40:08 even seen yet? And so search ads often 40:09 end up substantially underperforming 40:11 shopping. And for the brands that search 40:13 campaigns do do very well on, it's 40:15 normally because of the pre-existing 40:16 brand recognition. Not to say this isn't 40:18 incremental, like you should run them if 40:20 you can, if they're profitable. But for 40:22 a lot of our clients that almost 40:23 everyone knows in Australia, when you're 40:24 searching for a dress and then their 40:26 search ad pops up and then someone 40:27 clicks and buys, they're not clicking 40:29 and buying if they don't know who this 40:30 person was. And so it's actually all of 40:32 the existing brand awareness that even 40:33 causes this ad to work in the first 40:35 place. So if you're a small business, 40:36 you definitely probably won't be able to 40:37 see success with a search ad in fashion. 40:40 Now the second thing here is feed 40:42 optimization. This might be the largest 40:45 needle mover when it comes to 40:46 performance within shopping for fashion. 40:49 So first big mistake is that most uh 40:51 e-commerce websites in fashion have 9x6 40:54 imagery. 9x6 doesn't cleanly flow 40:56 through into the feed. Google will 40:58 autocrop it. Sometimes it doesn't do a 41:00 very nice job of it. And so you want 41:02 your images flowing through in one by 41:04 one or 4x5. 4x5 is actually the 41:07 preference these days so that it doesn't 41:09 autocrop and it fits the feed nicely. 41:11 Second thing is that you do want to 41:13 always build be building contrast in 41:15 some way on shopping. And so if you 41:17 think through the shopping experience 41:18 when you go and search for something 41:19 right now and you get that banner up the 41:20 top, what draws your eye is often the 41:22 listing that has contrast. And so if 41:25 everything is just a white background of 41:26 a product and then you have a model 41:28 outside suddenly that will pull your 41:31 attention. Now will it necessarily drive 41:32 a higher conversion rate etc? Probably 41:34 not. But if we're looking at just 41:35 pre-click intent and we want to get as 41:37 many clicks as possible and drive up 41:39 clickthrough rate, we want to be 41:40 building contrast in some way. And so 41:42 often testing some kind of lifestyle 41:44 imagery here, and you can do this just 41:45 on select products and CTR goes up, will 41:48 generally drive higher CTRs cuz we've 41:50 ran this test quite a bit. So this is 41:51 the image component. The next is titles. 41:54 Most fashion brands will just pull in 41:56 the default Shopify title, which will 41:58 look something like this. It'll just be 41:59 brand product title. You want to change 42:01 this to a product title that is more 42:04 descriptive. So rather than just because 42:06 typically products are titled with some 42:08 arbitrary name, you want to build the 42:09 name out a little bit of more detail so 42:11 that there's actual context being 42:12 provided here. Then you want some kind 42:14 of feature of the product. Then you 42:16 either want a colorway or a category in 42:18 here. And then finally at the back you 42:21 want the brand. This is immediately 42:22 going to outperform this as more context 42:24 is being provided to Google as to the 42:26 actual keywords that this product should 42:28 place on. Often these brand key terms 42:30 get prioritized in like everyone's 42:31 titles, but it restricts your cold key 42:34 term discovery volume because Google 42:36 will prioritize your quality or match 42:39 scores against keywords based on how far 42:41 forward they are in the titles. And so 42:42 if you're using the most important real 42:44 estate in the title on your brand name, 42:46 what's the point, right? There's almost 42:47 no point. You want it all the way at the 42:49 back. The brand's already mentioned 42:50 anyway in the shopping listing down the 42:51 bottom of it. Next, images, titles, and 42:54 this is in order of importance. We then 42:56 have product type. So you just want your 42:58 product type tagged up within the feed. 43:00 And then the last down the bottom here 43:01 is size curve. Now this is getting a 43:04 little bit more complicated, but this is 43:05 a really good little gem that you can 43:07 apply, which is that you want dynamic 43:09 feed rules so that when your highest 43:11 seller rate size, let's say that your 43:13 highest seller rate size is a medium as 43:15 an example, when this medium goes out of 43:17 stock, this entire product gets pulled 43:20 out of the feed, all the sizes. And the 43:21 reason being is that the return on ad 43:23 spend, the efficiency of this product 43:24 through paid media once the primary size 43:27 curve goes out of stock will drop 43:28 precipitously, probably by like 50%. And 43:30 so to avoid that decline from even 43:32 occurring, let's just pull the product 43:33 out of the feed the moment the highest 43:35 sellrough rate size curves drop out. The 43:37 third thing on account structure in 43:39 Google is brand search. Now, this is 43:41 primarily related to large retail 43:43 fashion brands, which is that pretty 43:45 much every single large retail fashion 43:46 brand I've ever seen, which is a good 43:48 probably 20 plus, they all overspend on 43:51 brand search like crazy. Brand search 43:53 just always, every single time, ends up 43:55 taking up an enormous amount of spend 43:57 irrespective of the fact that they have 43:58 no competitors bidding on their branded 44:00 search term. And so, just make sure that 44:02 if you're in fashion, don't spend as 44:04 much on brand because you don't need to. 44:06 And then four, when it comes to 44:08 structure and actually setting up the 44:10 account, it is a similar approach to 44:12 meta, which is that you want to be 44:13 splitting your campaigns based on 44:16 business units. And the most important 44:17 thing here is there just needs to be a 44:18 very very good reason that you can steal 44:21 man as to why the campaigns are 44:22 structured in the way that they are. 44:24 Because every time you segment campaigns 44:26 in Google, you will decrease performance 44:28 because you are segmenting out the 44:30 conversion data and the campaigns work 44:32 better when they have more conversion 44:34 data to model against. And so how many 44:36 campaigns you have is generally a 44:38 product of your ad spend. If you're 44:39 spending under 50k a month on Google, 44:41 you probably don't want more than 3, 44:43 four, five campaigns at a maximum. If 44:45 you're spending way less than that, if 44:46 you have a 10k budget, you ideally don't 44:48 want more than two campaigns. And then 44:50 obviously that flexes with scale. Now, 44:52 it still always is the case, even if 44:53 you're spending a million dollars a 44:55 month, that less campaigns will 44:56 generally yield better performance 44:58 unless they're not aligned with the 44:59 commercial objectives of the business. 45:01 Now what segmentation will typically 45:03 look like is by category and there's a 45:06 few reasons why. Number one, categories 45:08 have different gross margin profiles. 45:10 They also have different search volume 45:12 and so they will have different limits 45:13 of scalability on platform. Number two 45:15 is just splitting by gross margin 45:17 percentage if you want a better KPI and 45:19 have different target return on ad 45:20 spends across them. Number three is 45:22 splitting new arrivals 45:25 versus more of your evergreen 45:28 products. And then number four is just 45:30 having a bunch of campaigns that are 45:32 super dynamic where you have like best 45:34 sellers 45:36 and then you're modulating against like 45:37 zombie products 45:39 or you can quite literally just title 45:41 this grade A B C D inventory and have 45:44 rules that are automatically moving 45:45 products between here. So there's a lot 45:47 of different options and it's going to 45:48 be ultimately down to you. I can't just 45:50 recommend and I never will recommend 45:51 just one account structure that everyone 45:53 should run because every fashion brand 45:55 we work with has a slightly different 45:57 structure based on their commercial 45:59 objectives. Now we've touched on Meta 46:01 and Google here. We can put a quick note 46:03 on Tik Tok and Pinterest. On Tik Tok, 46:08 Tik Tok is an incredible platform for a 46:11 few of the fashion brands that we work 46:12 with. It is also a terrible platform for 46:15 a few of the others. And the big decider 46:18 on whether Tik Tok is going to work 46:19 really well for you is going to be 46:21 focus, but more so specifically focus on 46:24 creative. And so if you have an 46:27 incredible Tik Tok organic strategy, you 46:28 have a lot of content going up. It's 46:30 really good content. It's going to work 46:31 organically. It's going to work on paid, 46:33 you should be running Tik Tok ads. Super 46:35 underlever, really good opportunity. If 46:38 you don't have good native Tik Tok 46:39 content that you're pumping out at 46:41 volume, Tik Tok will not work for you. 46:43 It will not be good because you need 46:44 native content. you need a lot of it cuz 46:45 it fatigues very quickly. So, this is 46:47 really the core decider whether you run 46:49 Tik Tok or not. Tik Tok, if you have all 46:50 this, it's great. We've ran to date, I 46:53 believe, five lift experiments on Tik 46:55 Tok and four of them have been 46:57 unbelievably positive. And that's 46:59 fashion specifically. And then outside 47:01 of fashion, we've ran another four 47:03 inverse holdout tests and all of them 47:04 have been terrible. And so, Tik Tok has 47:07 shown pretty much no incrementality and 47:09 we've pulled clients out of the platform 47:11 entirely. And so in fashion, it seems to 47:13 be better from the data that we have and 47:15 the lift tests that we've run. And then 47:17 in conjunction to that, it will pretty 47:19 much always perform as long as there's a 47:20 really good creative organic strategy. 47:23 Then when it comes to Pinterest, 47:24 Pinterest is a really like tertiary 47:26 channel. And just for the sake of this 47:28 video, I'd recommend 99% of people just 47:30 don't run it. Uh Pinterest also 47:32 substantially overattributes like crazy 47:34 because a view through conversion on 47:36 Pinterest is very easy to get because if 47:38 you open up Pinterest, right, you have 47:40 all of these different tiles, you see 47:42 like 10, I guess you call it pins all at 47:44 once. And if one of these is an ad and 47:47 then you buy from that brand within 24 47:48 hours, Pinterest claims a conversion. 47:50 And so you end up with massive view 47:52 through overattribution because it will 47:54 just serve pins to anyone who's an 47:56 existing website visitor. And then when 47:57 that person inevitably goes and buys, 47:59 Pinterest claims a conversion. And so 48:01 when you look at click-through purchase 48:03 value or click-through rorowaz on 48:04 Pinterest, it ends up being way worse 48:06 than what you actually think it is on 48:08 the surface. And it's also generally not 48:10 a very scalable platform. And so when 48:12 you try to push spend up, particularly 48:13 in the Australian market, which is very 48:15 small, uh you can't get much spend out 48:17 of this platform at all. You can get 48:18 like maybe $10,000 a month at a maximum 48:21 in the Australian market. And the 48:22 question becomes, do you want all the 48:24 time, resourcing, and effort going into 48:25 campaign builds and optimization on a 48:27 platform that's only going to spend 10K 48:28 a month when you could just go and drop 48:30 an extra 10K into Meta and probably see 48:31 a better incremental return and way 48:34 better efficiencies and operating 48:35 expenses and focus within the business. 48:37 And I would always choose consolidation 48:39 with the platforms that I'm advertising 48:41 on. So, in our 3hour creative master 48:44 class, we went through everything 48:45 regarding creative strategy. So, I'd 48:47 highly recommend you go and watch that 48:48 video as well if you're interested in 48:50 creative. In this video, we're going to 48:52 be going through the specific nuances 48:54 behind all of that creative strategy and 48:56 how it changes inside of fashion. Now, 48:59 one way that we ideulate creative is 49:02 using what's called concepts. So, 49:04 concepts is the intersection of an 49:06 angle, an offer, and a persona. Now, 49:09 people often think that they can't build 49:10 out concepts in fashion because we're 49:13 talking to one persona, which is the 49:14 target demographic of whoever the brand 49:16 is. angle is here's clothes and then the 49:20 offer is just the product. And so how do 49:22 we actually use this concepting 49:24 framework to be able to generate 49:26 concepts within fashion? And so the 49:28 offer is fairly obvious. This is just 49:30 going to be the product. Now yes, this 49:32 could be a discount or a bundle or some 49:34 kind of gift with purchase, whatever you 49:36 want it to be. I generally really don't 49:38 like when fashion brands do that because 49:40 they become very offer driven. And we 49:42 actually internally uh when we're doing 49:43 audits, we call this offerdriven fashion 49:45 brands, which is that they're just 49:47 always doing like a buy three t-shirts, 49:49 get one free or like a gift with perses 49:51 or some kind of bundle. And that's the 49:52 only way that they keep the business 49:53 aloat. These offer-driven fashion brands 49:56 typically end up having really bad 49:57 margin. I've seen a lot of their P&Ls 50:00 and it's like 5 to 10% IBIDA is because 50:02 they're super reliant on just always 50:05 rotating through offers all of the time 50:06 and that's the only reason they get 50:07 people to buy and their LTV is horrific 50:09 because people are buying due to the 50:11 offer. They're not buying due to the 50:12 brand and the affinity that they want to 50:14 make and so therefore it's very 50:15 transactional becoming a loyal customer. 50:18 So if you are an offer-driven fashion 50:20 brand I'd kind of recommend trying to 50:21 pivot out of it because I haven't seen 50:23 that model do very well at scale. Now, 50:25 when it comes to the persona, there is 50:26 flexibility here and any large retail 50:29 business will have four or five 50:31 different personas built out that their 50:33 four to five different customers. They 50:34 give them names and they say, "This is 50:35 who actually buys from us." This is 50:37 effectively your persona breakout. So, 50:39 if you're planning on working with us or 50:41 if you're working with any agency, I'd 50:42 strongly recommend that you give them 50:44 this internal uh branding document that 50:46 breaks out the personas cuz then when 50:48 the agency or you internally are 50:50 building out concepts, you can just 50:51 build it out based on those personas 50:53 that you've already built internally. 50:54 And they're normally pretty in-depth 50:56 anyway. If you don't have these or if 50:57 you haven't built them, I strongly 50:59 recommend doing it. Now, how you 51:00 actually do this as an exercise, it's 51:02 probably a 20-minute tutorial in itself, 51:04 but the quick TLDDR is that you want to 51:06 survey your actual customers and you 51:08 want to figure out, okay, who are all 51:09 the people that are purchasing from us? 51:10 Why are they purchasing from us? Where 51:12 are they at their stage in life? Are 51:13 they a trends setter or are they just 51:15 hopping on translate? Like, what kind of 51:17 person is this within fashion? And then 51:19 how do we start to categorize them into 51:20 multiple different groups? So, personas 51:22 is straightforward. Offer is relatively 51:24 straightforward and restrictive. And 51:26 then we have the angle. Now, this 51:28 ultimately comes down to how you are 51:31 going to present the product and then 51:34 also what problem you're going to 51:37 agitate on and solve. And now, this 51:39 doesn't have to be super feature- 51:40 driven, but it needs to at least 51:42 insinuate towards a feature. So, a 51:45 creative that does incredibly well in 51:47 fashion is something that accentuates a 51:50 feature that is caused by the product. 51:52 So, if the product cinches on your waist 51:54 really well, if you can accentuate that 51:56 within the actual video in the creative, 51:59 then it ends up doing incredibly well 52:00 because it's orientating around a 52:02 problem and it's presenting a really 52:04 good solution. Another example of how to 52:06 think through this very well is to think 52:07 through a very specific situation that a 52:10 customer will be in that we can speak 52:12 to. So the angle might be that if you 52:15 are a mom with age 10 years old and you 52:18 want to have a work wear that also 52:22 allows you to go straight from work to 52:24 home to date night without ever having 52:26 to change once. Well, this particular 52:29 product is the solution to this 52:31 situation that you'll be in. Now this 52:33 situation probably a lot of people are 52:35 in this to some capacity, right? People 52:36 go from work and they go to date night 52:38 and they have kids. So this is talking 52:39 to a very specific person and we can 52:41 obviously go more specific regarding the 52:43 persona when we think through the 52:44 language and framing here and this can 52:46 be done through a piece of user 52:47 generated content of this target 52:49 demographic. Right? So you can start to 52:51 script this out around I really wanted 52:53 something where I could jump straight to 52:55 work into date night cuz I didn't have 52:56 enough time at home to get changed 52:58 because blah blah blah blah blah blah. 52:59 Therefore this product is so helpful in 53:01 this particular circumstance. And that's 53:03 how you frame the entire created. 53:04 Obviously, you don't just frame around 53:06 this solution, but you frame around all 53:07 of the benefits of that particular piece 53:09 of clothing. And even better, if you 53:10 want to increase the total addressable 53:12 market and make it a more scalable ad, 53:14 you rotate between a few different 53:15 pieces or you sell a set as well. So 53:17 rather than just selling, let's say, the 53:19 dress, you're selling the dress plus a 53:21 bag. And so when you start thinking 53:22 through angles in more specificity like 53:24 this, you'll start realizing that you 53:26 can actually generate concepts in 53:29 fashion at a very large scale. Okay, you 53:32 can probably generate hundreds of these, 53:33 then rotate them through all the 53:34 different personas, then rotate them 53:36 through tons of products. And this is 53:38 one of your biggest benefits in concept 53:40 generation in fashion is that the 53:41 products are always rotating over and 53:43 changing. And so, very quickly, throw 50 53:45 products in here as 50 different offers. 53:47 Throw your personas. You might have six 53:48 core personas. And then angles, go and 53:50 ideulate on 50 angles. In fact, let's 53:52 just cut that down. Let's say only 20 53:54 angles. Well, now all of a sudden, 20 * 53:56 50* 6. So, if you mix and match all of 53:59 these different options, you will have 54:01 6,000 concepts. All right, that is 54:03 obviously crazy. This is enough concepts 54:05 to get you 6,000 ads. And really, for 54:07 each concept, you can make 10 to 20 ads 54:09 per concept easily. And so, we're 54:10 looking at like 60 to 120,000 ads off 54:14 just sitting down, getting 50 products, 54:16 getting your personas ready, and 54:18 thinking of angles. And so, there should 54:19 be no limit in terms of creative volume 54:22 within fashion once you're approaching 54:24 it through this framework. Now the 54:26 question I get all the time is how 54:27 should we think through the production 54:29 split and honestly I don't think this is 54:31 a great exercise because for every brand 54:33 this is going to be very different and 54:34 for some brands just based on 54:35 performance data you want to lead into 54:37 whatever is performing well however if I 54:39 have to give percentage allocations 54:41 campaign shoots which is like hi-fi 54:43 production set at 25% of volume UGC and 54:46 creator content sits at 40% of volume 54:49 now this can include by the way like 54:50 founderled content as well as EGC etc 54:53 organic repurposing so taking organ 54:55 organic content, repurposing it into ads 54:57 sits at about 20%, and then we've got 54:59 statics sitting at about 15% of 55:02 production. So, as a rough split, this 55:03 is what you should be looking at. 55:04 Obviously, these numbers are going to 55:06 change brand to brand based on 55:07 performance, based on what does well. 55:08 Some people will zero this out entirely. 55:10 Some people will bring this up a little 55:11 bit, etc., etc. So, let's give a bunch 55:13 of nuances on UGC and creators within 55:16 fashion. Number one, fashion UGC does 55:19 not need talking, which is honestly one 55:22 of the huge benefits of it. So you 55:23 actually do not need any voice. Some of 55:25 the best performing UGC that we've 55:27 actually had has not had any talking. 55:29 This also has an added benefit that it 55:31 becomes incredibly scalable globally 55:33 because accents will restrict 55:35 performance in foreign countries. If you 55:37 don't have any talking, this becomes a 55:39 very scalable asset. And so I recommend 55:40 if you're getting UGC out of a creator, 55:42 just get a non-talking version as well. 55:43 It'll be easy to shoot. It'll be 55:44 quicker. They just don't need to voice 55:46 over. And it will give you a second 55:48 asset that you can test, but also an 55:49 asset that might be able to scale 55:50 globally. the format close up to full 55:52 body does incredibly well. So you start 55:54 really close up to camera and then you 55:55 pull back. This is likely because it's 55:57 an organic format that's translating 55:58 well into paid right now. This might not 55:59 be the case at the time that you're 56:00 watching this video. I think the better 56:02 thing to understand here rather than the 56:04 fact that this is a format you should be 56:05 testing is that formats will continue to 56:08 rotate over time and you need to be on 56:10 top of what formats are actually working 56:12 organically and translating them into 56:14 paid as well as you need to be testing 56:16 new formats. If you want to be first to 56:18 market on something new that's actually 56:19 going to work. The only way that you do 56:20 that is you test new formats. You 56:22 generally want to add size data on the 56:24 screen or on the primary text somewhere 56:26 so that it reduces returns. You want to 56:29 use trending audio obviously if you have 56:32 the rights to don't want to claim being 56:34 put on you. But this is incredibly 56:36 effective and works well on Tik Tok. Uh 56:39 Tik Tok did a wide meta analysis on this 56:42 which they saw that performance was 30 56:44 to 40% better on all ads that had some 56:47 kind of trending audio that was overlaid 56:49 on the content. And that's obviously a 56:50 product of the ad just feels more 56:51 organic. The next is that you want to 56:53 show as many products as possible. And 56:56 so the best way to do this normally is 56:58 like a curated haul. And the reason 57:00 being is that it makes the asset more 57:02 scalable and it can last for longer. If 57:03 you pay a UGC creator for a piece of UGC 57:06 and there's just one product in it, when 57:08 that product inevitably goes out of 57:09 stock, you're dumb and you need to roll 57:12 this ad out of the account and it's 57:13 over. You're never going to use that 57:15 piece of content again unless you 57:16 restock that product next year. So 57:17 instead, you always want to have 57:19 multiple different products so that you 57:20 can get more longevity out of the asset. 57:22 So these are some good pointers when it 57:24 comes to UGC creators. Two quick notes 57:26 here, EGC and fitting room content. 57:28 These are two really underutilized 57:29 content types at the moment. EGC is 57:31 employee generated content. You take 57:33 staff members, they create in store 57:34 content behind the scenes, etc. Feels 57:36 really authentic. A lot of people are 57:37 doing this and putting it onto Tik Tok, 57:39 but they're not actually repurposing 57:40 into ads. And then fitting room content 57:42 is off the back of AGC, which is raw, 57:44 authentic fitting room tryon content. 57:46 Uh, this has done unbelievably well for 57:49 a couple Australian brands. It has 57:50 allowed them to scale internationally. 57:52 So, this is a format that is great. 57:54 Ultimately, if we're talking about 57:55 creative, we can't not mention the brand 57:57 aesthetic versus direct response duality 58:00 that exists within fashion, which is 58:02 that you're going to have performance 58:04 within an ad account sits somewhere 58:06 here. But most teams pull too far into 58:09 brand aesthetics or there's other people 58:10 that pull way too hard into direct 58:12 response to where it's probably damaging 58:14 to the brand. This is rarer in fashion, 58:16 at least with the large retailers that 58:17 we work with. But I think the real key 58:18 here is that you need to let the data 58:20 feed into your decision-m. So, if we go 58:23 all the way over to a brand that maybe 58:25 sits here that's super direct response 58:26 and has no brand aesthetics. So, let's 58:28 say everything they run is like 58:30 affiliate super scrappy UGC that's 58:33 pushing heavy benefits and features of 58:35 the product. That's fine. But because 58:37 they're never investing any any like 58:39 brand level production because they're 58:41 not investing in any brand level 58:42 production, it'll ultimately cap their 58:45 brand perception to a degree because 58:47 their brand is just all of this content 58:49 and the summation of it which is super 58:52 features super features driven. On the 58:54 other side of things, we have 58:56 traditional retail businesses who 58:57 haven't needed to run performance 58:59 marketing based ads who are only just 59:01 putting campaign shoots in their ad 59:03 account. Now the issue is this stuff is 59:04 not native at all to the platform and it 59:06 doesn't convert anywhere near as well as 59:08 this content converts. And so we need to 59:10 find a middle ground ideally somewhere 59:12 here where we have a combination of 59:14 direct response advertising that stays 59:17 on brand and the format doesn't infer 59:19 the branding right so like you can stay 59:21 on brand for something that's super 59:22 premium like an example would be like 59:24 Apple like Apple could run an ad that's 59:26 super premium but it could be a piece of 59:28 userenerated content as long as it's a 59:30 little bit more polished and it stays 59:31 within the brand guidelines. They don't 59:33 have to just run campaign shoots over 59:35 here. And honestly, it's fine to just 59:36 not take my word for this and just stay 59:38 over here and just keep pumping out 59:39 campaign shoots all day long because the 59:41 reality is is that anyone that actually 59:43 gets to move themselves more into the 59:45 the middle are the ones that'll actually 59:46 perform well on the platform and 59:48 ultimately won't go under. And so you 59:50 can stay over here all you want if you 59:52 don't believe me and you don't think you 59:53 need to move a little bit more towards 59:54 direct response type marketing and 59:56 advertising. Uh but this is ultimately 59:57 how the platforms work. This is how they 59:58 are rewarded. All the brands that we 1:00:00 work with that are scaling the fastest 1:00:01 are the ones that balance this duality 1:00:04 really well where they can keep direct 1:00:05 response marketing on brand and continue 1:00:08 to pump out high volumes of content as 1:00:10 well. A subcomponent of this brand 1:00:12 versus performance argument and to 1:00:14 actually give you something material to 1:00:15 think about is the percentage of content 1:00:18 that's between desirebased content and 1:00:20 painointbased content. Most brands are 1:00:22 weighted to about a 9010 and so 90% of 1:00:24 their content will be desire 1:00:26 aspirational content and then 10% will 1:00:28 be painpoint driven. This mix should be 1:00:30 a lot further towards 30 to 40% over 1:00:33 here. The ad accounts that we're saying 1:00:35 the best performance on lean further 1:00:37 into painpointbased content. Now what's 1:00:38 an actual example of this? Well, a hook 1:00:41 here could be I never know what to wear 1:00:43 to and then just put an occasion in 1:00:45 here. Or let me give you a gifting angle 1:00:47 one which is my partner is impossible to 1:00:49 buy for. These are opening with pain 1:00:51 points and then obviously moving down 1:00:53 the stages of awareness into selling. 1:00:54 This will generally drive better 1:00:56 performance in the platform and you can 1:00:58 stay on brand particularly by you just 1:01:00 run this stuff as partnership ads. So it 1:01:02 doesn't even run through your handle. 1:01:04 Due to the seasonal nature of fashion, 1:01:06 creative planning is often done like 1.5 1:01:09 to 2.5 months out from the product being 1:01:12 available. Now, one mistake I think a 1:01:14 lot of people make within the seasonal 1:01:15 rotation of creative inad accounts is 1:01:17 that when all these new products come 1:01:19 in, so there's a lot of newness, the old 1:01:21 products, if they are still driving 1:01:23 performance, now it might not be at a 1:01:25 large spend level, but it might be 1:01:26 efficient. We might be turning over 1:01:28 stock. What will often happen is people 1:01:29 will kill the old and they'll move all 1:01:31 the spend into the new. This is a 1:01:33 mistake because you end up just 1:01:34 decreasing sellrough rates on these old 1:01:36 products, which the seller rates on 1:01:37 these products is only going to decline 1:01:39 over time because they're going out of 1:01:40 season. And so we actually want to 1:01:42 either maintain spend here or increase 1:01:44 it to be able to sell through and then 1:01:45 move on to new. Right? This is where 1:01:47 inventory visibility once again becomes 1:01:49 incredibly critical because if you are 1:01:50 an agency with no visibility into 1:01:52 inventory, with no visibility into any 1:01:54 of the wider strategy, what do you do? 1:01:56 And this is what I see in faction 1:01:57 accounts all the time is just, oh, new 1:01:58 stuff comes in, old stuff off, old stuff 1:02:00 off, old stuff off. And then you just 1:02:02 end up with turning off ads that were 1:02:04 performing okay and turning over product 1:02:06 and just rep prioritizing new. and you 1:02:08 end up in this cyclical position 1:02:09 continuously where you're just moving 1:02:11 products to old and taking them out of 1:02:13 the ad account despite the fact that we 1:02:14 could keep spending on them and push 1:02:16 through the remainder of the stock. Now, 1:02:17 due to the way that this planning uh 1:02:19 setup is structured as well, we want to 1:02:21 be taking creative learnings and we want 1:02:23 to be pushing them into future creative 1:02:25 production so that we have these one to 1:02:27 two month cycles of new content coming 1:02:29 through that's based on the data from 1:02:31 the last product drop. In some fashion 1:02:32 brands, there's pretty much always some 1:02:34 kind of tradeoffer going. So there's 1:02:36 some kind of discount, there's some kind 1:02:38 of take a further off, etc. Um, with 1:02:40 that, you want to be thinking about 1:02:41 budget allocations across the ad account 1:02:43 into these discounts. Generally, this is 1:02:45 obviously going to be based on targets 1:02:47 and budgets, etc. But we would generally 1:02:49 like to cap this at 20 to 30% of budget 1:02:51 so that we can maintain a 70 to 80% 1:02:54 budget allocation into what I would call 1:02:56 our evergreen campaigns that are going 1:02:58 to continue to run and we're not 1:02:59 disrupting performance here. All right. 1:03:01 So, I want to talk about scaling 1:03:02 internationally for a second. I want to 1:03:04 give you how we've done this really 1:03:06 effectively specifically for one brand 1:03:08 and then all the cautionary tales around 1:03:10 it. There won't be much drawing at this 1:03:11 segment. I'll kind of just be walking 1:03:13 you through our experience on 1:03:15 international scale. So, the first one 1:03:16 is that people try to scale 1:03:17 internationally too quickly. If you're 1:03:19 an Australianbased brand or if you're 1:03:20 based in a country that's a similar 1:03:22 population size to Australia, you really 1:03:23 want to go to 10 million AUD a year 1:03:26 before you even consider international 1:03:28 expansion. Reason being is that anytime 1:03:30 you expand internationally, you will 1:03:32 just increase the operational complexity 1:03:34 of the business and you likely won't see 1:03:36 as good results in the new country as 1:03:38 you do in the primary market right now. 1:03:39 So, you're just going to decrease 1:03:40 efficiency for no reason when you can 1:03:42 just do more scale locally. Another 1:03:43 component of scaling fashion, I'll give 1:03:45 you a case study here, is that I did an 1:03:47 audit on a brand a while ago who was 1:03:49 inhouse. So, I did it just out of 1:03:51 goodwill cuz I really like the guys and 1:03:52 they do about 10 to 20 million a year in 1:03:55 Australia and New Zealand. And they were 1:03:56 asking me, okay, we're planning for this 1:03:58 US launch. We're thinking about all the 1:04:00 operational logistics and should we have 1:04:02 a separate ad account. How should we 1:04:04 think through the paid strategy 1:04:05 changing? And I said, guys, I've seen 1:04:06 your ad accounts, right? You have like 1:04:08 four to five elite ads that are 1:04:11 evergreen that run all day long just on 1:04:13 one or two SKs and they have spent in 1:04:15 the Australian market. I think each of 1:04:17 them had spent nearly a4 million dollars 1:04:18 in lifetime spent. And I went, this is 1:04:20 four to five actual winning ads that you 1:04:23 can take and you can run them in any 1:04:25 English-speaking international market 1:04:27 and they will crush. If you have an ad 1:04:29 that's done that well in the Australian 1:04:31 market, it will do well elsewhere. And I 1:04:33 said just make sure you have a onetoone 1:04:35 pricing strategy. So whatever your AUD 1:04:37 pricing is, keep that number the same 1:04:39 and just change it to USD. So if you're 1:04:41 selling 70 AUD, you now sell 70 USD. So 1:04:44 you get a margin expansion and you ride 1:04:46 the 4x advantages here. And then they 1:04:48 did it. They did it one to two months 1:04:49 later. They launched into the US market. 1:04:51 They saw better efficiency on their ads 1:04:52 in the US than in AU because they had 1:04:54 fatigued the AU market so heavily. And 1:04:56 now the business has, I believe, nearly 1:04:58 tripled and 50% of revenue comes from 1:05:00 the US market, which is crazy. And the 1:05:02 way that they were so successful there 1:05:03 was because they had winning ads in the 1:05:05 Australian market that had performed 1:05:07 incredibly well that they could 1:05:08 translate over. If they didn't have 1:05:10 those winning creatives, the expansion 1:05:12 wouldn't have worked. And I have seen so 1:05:14 many brands try to expand to the US as a 1:05:17 crutch to try to save Australia. But 1:05:19 that is the wrong approach to going 1:05:21 international. You do not go 1:05:22 international to save the primary 1:05:24 market. You need the primary market to 1:05:25 be working. You need to have winners and 1:05:27 then you can translate it 1:05:28 internationally. And the nice thing 1:05:30 about fashion is you do have a 1:05:31 structural advantage which is it's 1:05:32 fairly easy to ship clothes 1:05:34 internationally. Now when you are going 1:05:35 to these international markets, there 1:05:37 are a few considerations within fashion. 1:05:39 Number one, this is a little bit of a 1:05:41 hack. This might just be right now, 1:05:42 might not apply into the future, but 1:05:44 lowfi content performs better in the US 1:05:47 in fashion accounts than it does in AU. 1:05:49 So, if you're in fashion, you want lowfi 1:05:51 content, at least in your mix when you 1:05:53 do the US launch. If you're just like 1:05:54 campaign shoots only, good luck. Number 1:05:56 two is you want accent localization. 1:06:00 So, unfortunately, Australian accents in 1:06:03 videos don't do as well as they do in 1:06:05 Australia in international markets like 1:06:07 the US. Now, UK to AU does decently 1:06:10 well. You can take these accents and run 1:06:12 them in both countries and there's not 1:06:13 much of a performance drop off, but 1:06:15 there is a performance drop off. And so, 1:06:16 ideally, if you want to maximize and 1:06:18 keep performance where it is, you want 1:06:19 the accents localized. And then the 1:06:21 advantage obviously that you have here 1:06:22 in fashion is that you actually uh don't 1:06:25 need audio running on the ad, right? In 1:06:27 fashion, if you just have statics or you 1:06:28 have GIFs or you have UGC with no 1:06:30 talking overlay, you don't need this 1:06:32 accent localization. You can just run 1:06:34 stuff internationally. If we want to get 1:06:35 more specific to US launches because I 1:06:37 know a lot of my audience is in 1:06:39 Australia and they're wanting to launch 1:06:42 into the US as the primary market. Well, 1:06:44 number one, keep your pricing one to one 1:06:46 from AUD to USD. A lot of a big mistake 1:06:49 that people will make is they'll 1:06:50 decrease their US pricing substantially 1:06:52 against Australian. So their $70 AUD, 1:06:55 they'll convert that to like 52 USD, 1:06:57 which is a big mistake. US dollar has 1:06:59 higher purchasing power. You can 1:07:00 actually get margin expansion in the US. 1:07:02 Number two is you need catalog or 1:07:04 product alignment, right? And so if you 1:07:06 don't have dedicated buying for Northern 1:07:09 Hem, well then you need the website to 1:07:11 be filtered down to just products that 1:07:14 are transseasonal. Otherwise, people 1:07:16 will hit the website, they'll browse 1:07:18 around, they'll see a bunch of products 1:07:19 that have nothing to do with the season 1:07:21 that they're in, and they'll bounce and 1:07:23 conversion rates will be incredibly low. 1:07:24 And so even though you might be driving 1:07:26 traffic to one particular product 1:07:28 landing page that has a transseasonal 1:07:30 product, if they then click out to the 1:07:32 wider website, which they always do cuz 1:07:34 they want to see the wider collection 1:07:35 mix and these products aren't relevant 1:07:37 to the season that they're in, you'll 1:07:38 see much lower conversion rates. And 1:07:40 then lastly, we have state selection. So 1:07:42 this is going to be based on weather and 1:07:43 it's going to be based on the 1:07:44 demographic within the state. When you 1:07:45 do a US launch, you do generally want to 1:07:48 specify down to a particular state. 1:07:50 States in the US are as big as the 1:07:51 entire Australian population. So you 1:07:53 don't have to worry about a TAM issue 1:07:54 and this is just going to allow you to 1:07:56 be more contextually relevant to meet 1:07:58 the right audience where they are. Now 1:07:59 there's a lot of different ways you can 1:08:00 go about state selection and how to 1:08:02 think about it. I will leave that for a 1:08:04 separate video. Now before we go into 1:08:05 the final section of the video wrapping 1:08:07 everything together and giving you the 1:08:09 playbook I want to talk about 1:08:11 particularly omni channel brands that 1:08:13 have retail uh presence and a 1:08:15 measurement system called geolyft. Now 1:08:19 the reason why this becomes extra 1:08:20 applicable for omni or very large 9 1:08:23 figure retailers is because number one 1:08:25 attribution becomes very unreliable once 1:08:27 you achieve a large scale because you 1:08:29 have so many different channels that are 1:08:30 driving revenue that to understand the 1:08:32 actual impact of meta is very difficult. 1:08:34 Number two, when you have Omni, what you 1:08:35 actually really want to understand is 1:08:36 what is the impact of instore revenue 1:08:38 against the media spend that we have 1:08:40 across these platforms. And geo lift 1:08:42 experiments is the best most applicable 1:08:44 way to be able to do this. And so when 1:08:46 it comes to test design in the US, you 1:08:49 do this at a statebystate level and you 1:08:51 use a clustering algorithm that takes 1:08:53 all of your historical revenue and order 1:08:55 data in each state and it clusters them 1:08:57 together and pairs them based on the 1:08:59 states that move together correctly. So 1:09:01 what you end up seeing when you graph 1:09:02 this in the pre-est design period is 1:09:05 that let's say you select two states as 1:09:06 the control and two states as the 1:09:08 treatment. If you take these and you add 1:09:10 them together and you look at revenue 1:09:11 over time the control may look like this 1:09:13 which for the sake of this let's call 1:09:15 this maybe like Arizona and Texas. And 1:09:18 then if you look at the treatment it may 1:09:20 look like this. So they move together. 1:09:22 They're very tightly correlated over 1:09:24 time. And this could be like LA plus 1:09:27 maybe New York. And so the way that this 1:09:29 works is that because these are tightly 1:09:31 correlated, they can predict each other 1:09:33 on any day in this past set of data, I 1:09:35 could remove Arizona and Texas and I 1:09:38 could look at just New York and LA and I 1:09:40 would be able to predict with very high 1:09:42 accuracy what the revenue is in these 1:09:44 states because it's always like $100 1:09:46 lower or 95% of whatever the revenue 1:09:49 level is here. And so what we then do is 1:09:51 we go into one of these uh clusters. So 1:09:54 we'll go into the treatment which is LA 1:09:55 and New York. And maybe what we do is we 1:09:57 go and introduce Tik Tok and we 1:09:58 introduce Tik Tok at a 100K a month 1:10:00 spend. Then through the treatment window 1:10:03 what we want to see is does LA and New 1:10:05 York suddenly diverge and then the 1:10:07 difference in the divergence is 1:10:08 obviously the lift that we've seen. Now 1:10:10 the issue for my Australian audience is 1:10:13 that state level targeting doesn't 1:10:15 really work in Australia cuz there isn't 1:10:16 enough state selection. And so you 1:10:18 actually need to refine this down to 1:10:20 commuting zones, which is relatively 1:10:22 self-explanatory in the language here, 1:10:24 which is zones in which people commute 1:10:26 in. And so we want to isolate areas in 1:10:28 Australia where people stay within. So 1:10:31 they work there and they live there. 1:10:32 They don't cross borders. And then we do 1:10:34 our lift tests at a commuting zone 1:10:36 level. This gives us greater selection. 1:10:37 We have about 60 plus commuting zones in 1:10:39 Australia. And so we can do effectively 1:10:41 the same thing as we do over here, but 1:10:42 rather than doing it based on states, 1:10:43 it's based on more regional areas. Now 1:10:45 the added advantage of commuting zones 1:10:48 in Australia with omni channel is that 1:10:50 we can get more specific to stores. And 1:10:53 so most uh large retail fashion brands 1:10:55 will have a very high density of stores 1:10:57 in like Vic. So there might be seven or 1:11:00 eight stores in Vic. There might be 1:11:01 seven or eight stores, let's say 10 in 1:11:02 New South Wales, then you've got like 1:11:04 six in Queensland. And then you have 1:11:06 like one in Western Australia, one in 1:11:08 South Australia, one in Tazzay, right? 1:11:10 And so all the density is here. And so 1:11:11 when you do a lift test at a state level 1:11:13 like let's say in Vic, you have to 1:11:15 measure revenue lift across all of these 1:11:17 different stores which actually isn't 1:11:18 super applicable because storebased 1:11:21 sellrough rates are going to be 1:11:22 indicative of the product ranging and 1:11:24 the season and you might have better 1:11:26 performance in a particular state in a 1:11:28 particular week because the product 1:11:29 ranging is more conducive to the actual 1:11:31 weather at the time. And so when you can 1:11:33 pull down to a commuting zone level 1:11:35 instead, commuting zones will typically 1:11:37 only house one maybe two stores. And so 1:11:39 you can more tightly measure revenue 1:11:41 realization in the store against the 1:11:43 spend increase in the platform. Now this 1:11:45 doesn't just apply into Omni. We 1:11:47 actually ran this uh for Cookie when we 1:11:51 first took over and we're marketing into 1:11:52 the US market which is direct to 1:11:54 consumer only. There's no retail stores. 1:11:56 The reason why we even ran this 1:11:58 experiment was because they had a base 1:12:00 level of revenue within this market that 1:12:03 was being driven seem seemingly without 1:12:05 any causal impact to ads. And so the ads 1:12:08 just didn't seem very effective in 1:12:10 driving revenue. And so therefore, we 1:12:12 made an assumption that the ads probably 1:12:14 weren't causal to revenue, at least in 1:12:16 the way that they were being run at the 1:12:17 time. And so we obviously restructured 1:12:18 the account. And then before being 1:12:20 confident in just arbitrarily increasing 1:12:22 budgets within the market, we put 1:12:24 together state selection based on 1:12:26 clustering. Then we did a large budget 1:12:28 increase in one particular state. It 1:12:30 actually wasn't multiple in this case. 1:12:31 And then we measured the increase in 1:12:33 revenue against the baseline revenue of 1:12:36 the control. And what we saw was 1:12:38 fortunately massive lift in the test 1:12:40 state and so we proved causal impact of 1:12:43 metas-pend into the region and that gave 1:12:45 us the confidence to continue scaling 1:12:47 the US market and we scaled the US very 1:12:49 quickly. This then also applied into Tik 1:12:52 Tok once we layered Tik Tok in. We 1:12:54 obviously weren't confident at the 1:12:55 beginning that we should rapidly 1:12:57 increase Tik Tok in conjunction with the 1:12:58 rapidly increasing spend in Meta. So to 1:13:00 isolate the causal impact, we started 1:13:02 running lift tests on Tik Tok as well to 1:13:04 validate it as we continue to scale. All 1:13:05 right. So now running through the 1:13:07 playbook by spend. If you're spending 1:13:08 between $30 to $75,000 a month in ad 1:13:11 spend, you want to be running all your 1:13:13 campaigns on 7-day click. You want to be 1:13:15 using 180day altgic hack as the measure 1:13:18 for whether you're in a good position or 1:13:19 not. And you want this number to be 1:13:20 between a two to a three. In terms of 1:13:22 structuring meta, you want just two to 1:13:24 three campaigns. You want to be 1:13:25 consolidating Australia and New Zealand. 1:13:27 If you're an Australianbased brand, run 1:13:28 them all in the same campaigns. Don't 1:13:30 segment them out. And you want to be 1:13:31 capping existing customer spend. On 1:13:33 Google, you want shopping to be 80% plus 1:13:35 of your budget. You want to be focusing 1:13:36 on fixing your shopping feed. That's the 1:13:38 highest leverage activity within Google 1:13:40 for you. And then in terms of creative, 1:13:41 you want 30 to 50 new ads a month. You 1:13:43 want to be leveraging partnership ads. 1:13:45 And you want 80% of your creative to be 1:13:46 evergreen and 20% to be seasonal 1:13:48 related. If you're spending 75k to 200k 1:13:51 a month on measurement, you want to be 1:13:53 using incremental attribution for being 1:13:55 able to make decisions within meta. You 1:13:57 also want to be leveraging lift studies, 1:13:59 particularly if you're omni channel. You 1:14:01 want to be using acquisition me and 1:14:03 acquisition me adjusted to add back 1:14:05 returns as a daily measurement of 1:14:07 success. And you want to be looking 1:14:09 closely at your cohort analysis to 1:14:11 understand how repeat rates and return 1:14:12 rates are changing over time to be able 1:14:14 to layer that into your forecasting and 1:14:16 your KPIs against paid media. On meta, 1:14:18 you want four to six campaigns. At the 1:14:20 adset level, you want this to be split 1:14:22 based on concept testing. You want your 1:14:24 existing customer frequency to be capped 1:14:26 at a seven. And you want your DPAs to be 1:14:28 capped. This is the spend area where 1:14:30 people start making the mistake of 1:14:31 scaling up dynamic product ads. On 1:14:33 Google, you can have a little bit more 1:14:34 flexibility to other channels here now 1:14:36 with shopping going down to 60 to 70%. 1:14:38 And the big leverage here becomes size 1:14:40 curve management. So making sure 1:14:42 products are getting pulled out of the 1:14:43 campaign if the core sizes go out of 1:14:45 stock. Then in regards to creative, you 1:14:47 want 65 to 200 plus creatives per month 1:14:50 for this spend level. You want concept 1:14:52 mapping. So all of your creatives are 1:14:53 mapped against a particular concept 1:14:55 that's been built up here. And then you 1:14:57 want balance sheet. aka inventory 1:14:59 visibility for the creative that you're 1:15:01 producing to be able to turn over grade 1:15:03 BCD inventory rather than just blindly 1:15:06 making creative irrespective of the 1:15:08 amount of inventory that we need to sell 1:15:09 through across all of the different 1:15:11 product portfolio. Ideally, your 1:15:12 creative production here in terms of 1:15:14 quantity one for one matches the amount 1:15:16 of inventory that you have under those 1:15:18 SKs and then the SKs that obviously 1:15:20 aren't moving, we increase creative 1:15:22 production for them. on measurement at 1:15:24 the 200k plus ad spend level. We want 1:15:26 geolift testing in place for certain. We 1:15:28 want omni channel dduplication. We want 1:15:30 P&L reconciliation on a day-to-day and 1:15:32 month-to-month basis. So all of our 1:15:34 decisions within the ad account are 1:15:35 reconciling up into P&L outcomes. And 1:15:37 then we want to be tracking contribution 1:15:39 margin on firsttime customers as well as 1:15:41 existing customers to be able to once 1:15:43 again see platform impact on this 1:15:44 number. With regards to meta, the ad 1:15:46 account quite frankly becomes complex at 1:15:48 this level because at an ad account 1:15:49 spending over 200k a month, there is 1:15:51 likely complexity in the business, the 1:15:53 revenue and the business units that 1:15:55 exist. And so this needs to be 1:15:56 translated into the ad account. What 1:15:58 this actually looks like is going to be 1:15:59 very custom brand to brand. You also 1:16:01 want to be thinking about at this stage 1:16:03 introducing Tik Tok if it is a viable 1:16:05 channel for you and if you have the 1:16:06 creative production to support on 1:16:08 Google. you're now at the spend level 1:16:10 where you actually can consider 1:16:11 beginning going into YouTube as a more 1:16:13 top offunnel campaign. The issue with 1:16:15 YouTube is measurability is very poor. 1:16:18 And so you need the ability to run 1:16:19 geolyft tests and you need the ability 1:16:21 to allocate large budgets to it as a 1:16:23 campaign for it to actually work which 1:16:24 is why it's a viable campaign type at 1:16:27 this level because at this level you 1:16:28 should already be doing lift tests and 1:16:29 you should already have a lot of ads and 1:16:30 so because you have both of those things 1:16:32 YouTube becomes a viable next channel in 1:16:34 conjunction with Tik Tok to start 1:16:36 continuing to scale. In regards to 1:16:38 creative we want 200 plus ads a month 1:16:41 there. And this is going to be 1:16:42 indicative based on the expected value 1:16:44 per ad. We have a creative calculator 1:16:46 that allows you to actually calculate 1:16:47 this. Reach out to us if you want to 1:16:48 know your exact creative volume 1:16:50 requirements. You want to make sure at 1:16:51 least 10% of ad spend is going into 1:16:54 production. This is where people start 1:16:56 underweighting and not proportionally 1:16:58 increasing creative production with ad 1:17:00 spend. This needs to continue to go up 1:17:01 and scale because the creative demands 1:17:03 at this level of spend become really 1:17:04 important. And then for scaling, you 1:17:06 want to look at going international as 1:17:08 long as your localized revenue is above 1:17:10 10 million online. After working with so 1:17:12 many fashion brands and continuing to 1:17:14 audit fashion brands on a weekly basis, 1:17:16 if I was to generalize a few core 1:17:18 pillars of success that I see across the 1:17:20 ones that do unbelievably well and are 1:17:22 very profitable and growing and the ones 1:17:24 that are in a very different position is 1:17:25 the ones that are doing really well 1:17:27 generally have an incredible ability to 1:17:30 drive brand within the market and not 1:17:32 need to discount products and not need 1:17:35 to move a lot of products to discount. 1:17:36 So they end up being the brands that 1:17:38 build a reputation of not discounting, 1:17:40 investing in brand perception and then 1:17:42 using paid media to drive genuinely net 1:17:45 new customers rather than recycling 1:17:47 existing ones and overspending on their 1:17:49 existing customer base. Now you would 1:17:50 have noticed I didn't even mention the 1:17:52 word AI in this entire video. And the 1:17:54 reason for that is that we leverage AI 1:17:56 enormously in every other category 1:17:58 except for fashion created. And the 1:18:00 reason for that is that if someone can't 1:18:02 get clear visibility on exactly what the 1:18:03 garment actually looks like in real 1:18:05 life, then return rates end up shooting 1:18:06 up. And so if you're going to use AI 1:18:08 generated content, you're often going to 1:18:10 diverge from the reality of what the 1:18:12 product looks like. And that becomes an 1:18:13 issue when it's very visually driven 1:18:16 rather than benefit driven like 1:18:17 supplements are. So, as the very final 1:18:19 section in this video, what I wanted to 1:18:21 do was cut out to live ad breakdowns 1:18:23 that I have done of multiple different 1:18:25 fashion brands so you can get more of an 1:18:27 intricate understanding of the tactical 1:18:29 application into what makes a good 1:18:31 fashion creative, what doesn't, and how 1:18:33 you can go and translate that into your 1:18:34 own brand. Before we cut out to those ad 1:18:36 reviews, if you're a performance 1:18:37 marketer and you've gotten this far in 1:18:39 the video, please reach out to us at 1:18:40 hiringbluensedigital.com.au. 1:18:42 We are always hiring for A+ talent and 1:18:45 we would love to have a chat about any 1:18:47 opportunities that we currently have in 1:18:48 the team. And if you are a brand and 1:18:51 this was helpful to you, you feel it was 1:18:53 very contextualized to fashion, then 1:18:55 also feel free to reach out to us. 1:18:56 There'll be a link in the description 1:18:57 below to get a free audit. You need to 1:18:59 be doing at least $5 million a year in 1:19:01 revenue. And we need to have capacity 1:19:03 currently to be able to actually do the 1:19:05 audit and take on more clients. But as 1:19:06 long as those things are true, we'll be 1:19:07 able to give you an in-depth road map on 1:19:09 exactly how to apply everything that 1:19:11 we've done in this video directly into 1:19:13 your ad accounts. Zara's got 630 ads 1:19:16 running, but there's still a lot of 1:19:17 opportunity in their creative strategy. 1:19:19 Right now, they have zero videos 1:19:21 running. They have 26 images. They have 1:19:22 nine carousels. They have about 300 1:19:25 dynamic creatives active. So, that 1:19:27 probably has the videos in it. And then 1:19:29 they have 200 DPAs. 200 DPAs. They've 1:19:32 really weighted in there. I would 1:19:34 probably be pulling down budget 1:19:35 allocation to DPAs considering they have 1:19:37 such high brand awareness and brand 1:19:40 recognition. The DPAs are going to be 1:19:42 overcrediting but not actually 1:19:43 delivering an incremental impact at 1:19:45 least to the point that they think. Now 1:19:46 in the dynamic creatives they have 300 1:19:49 here. Normally you start running dynamic 1:19:50 creatives because you've hit your page 1:19:52 limit. I don't think they would have hit 1:19:53 their page limit. About 630 ads normally 1:19:56 you can go higher than this uh in the 1:19:58 reported ad library before you hit a 1:20:00 limit. And even so, worst case scenario, 1:20:02 they could spin out pages for different 1:20:05 regions. Now, they're a very big 1:20:06 business, so getting that over the line 1:20:08 and actually approved is a different 1:20:10 story. But going this hard into Dynamic 1:20:13 Creatives just means that you're not 1:20:14 collecting as much data on what's 1:20:16 actually working and what's not. So, 1:20:17 when you're looking at building creative 1:20:18 feedback loops into the internal 1:20:20 creative team or the agencies that 1:20:22 they're using, it's going to add a 1:20:23 little bit of complexity. They wait 1:20:26 really hard into professional shoots. As 1:20:28 you can see, pretty much all of the 1:20:30 content here is all professional shoots. 1:20:32 Either videos, either life style images, 1:20:35 either direct product shots spun into 1:20:38 carousels and DSAs. It's all very 1:20:41 similar in actual content format and 1:20:44 style. Obviously, the variation here is 1:20:47 just the actual product and the people 1:20:49 in the shoot. probably a very large 1:20:50 opportunity tier to start to introduce 1:20:52 more low-fi content, start to introduce 1:20:54 partnership ads, start to bring in 1:20:56 influencers if that's the direction that 1:20:58 the brand wants to go. Because when you 1:20:59 start talking about diversifying formats 1:21:01 within creative types on a business is 1:21:03 big, it really comes down to whether the 1:21:05 internal brand team is willing to make a 1:21:07 bet like that. But it's pretty likely to 1:21:09 improve performance just because they're 1:21:11 really restricting their ability to have 1:21:13 diverse content in the account. So, it's 1:21:15 how do they start to introduce more 1:21:17 creative types while still staying on 1:21:19 brand and keeping the visual aesthetic 1:21:21 consistent across every piece of content 1:21:24 that hits the end consumer, which I 1:21:26 think is more than achievable, right? 1:21:27 Meshki does an incredible job at this. 1:21:30 Cook eye does an incredible job at this. 1:21:31 You could list a bunch of brands that 1:21:33 have large diversity within the creative 1:21:35 set that still stay on brand and keep 1:21:38 consistency across all the creatives in 1:21:40 terms of tone, in terms of visuals, 1:21:41 whilst being able to drive further 1:21:43 performance through the Meta account. 1:21:44 White Fox Boutique only 42 ads running 1:21:47 right now. Let's break down what they're 1:21:48 doing well, but where there's a massive 1:21:50 opportunity right now in their paid ad 1:21:51 strategy. Now, they have a nice mix of 1:21:53 hi-fi and lowfi content. So, you have 1:21:55 hi-fi content here. You have a 1:21:56 relatively lowfi content here. Issue is 1:21:59 they have no partnership ads running, 1:22:01 and they have a very limited amount of 1:22:03 UGC or organic content running through 1:22:05 ads. Right now in fashion, partnership 1:22:08 ads, organic style content is what's 1:22:10 performing really well. then hopefully I 1:22:12 don't get their agency fired or an 1:22:13 in-house person in trouble here. But 1:22:15 from what I can see, their Australian 1:22:17 ads are driving to the UK website. You 1:22:20 can see that right there, UK URL. And 1:22:22 when I land on this, it wants to give me 1:22:24 a redirect. Now, fortunately, only half 1:22:26 their ads have that issue. The other 1:22:27 half are working and they're driving to 1:22:29 an Australian website on a collection 1:22:30 page, which is good. The collection page 1:22:32 is irrelevant. Last big opportunity 1:22:34 missing from the current creative set is 1:22:36 there's no audiobased hooks. You might 1:22:38 think fashion really, do we want 1:22:39 audiobased hooks in here? Surely that's 1:22:41 not aligned with the category. It's very 1:22:43 aligned with the category. All of the 1:22:45 best performing ads within fashion are 1:22:47 generally organic pieces of content run 1:22:49 through partnership ads with relatively 1:22:51 well-known influencers where there's a 1:22:53 strong audiary hook that actually pulls 1:22:54 people into want to watch the video. For 1:22:56 a multi- nfigure business that probably 1:23:00 spend a lot on meta at the moment, 1:23:02 there's a lot of opportunity here. 1:23:04 They're underweighted on userenerated 1:23:06 content. They're driving to the wrong 1:23:07 landing pages. They don't have any audio 1:23:09 hooks. and they have no partnership ads 1:23:11 from what I can tell from the outside 1:23:13 looking in running right now, which 1:23:15 would be a huge unlock for them. Zo is 1:23:17 doing a pretty good job right now in 1:23:18 their ad library. Let's break down 1:23:19 exactly why. Number one, they have 740 1:23:21 ads running, a lot of volume, 374 1:23:24 videos. They have a bunch of dynamic 1:23:25 creative. This is probably because 1:23:27 they're hitting up against their page 1:23:28 limit. If it isn't and you're watching 1:23:29 this Sabo, I would recommend leaning 1:23:31 away from DCOS's because it just limits 1:23:33 your ability to get data and continue to 1:23:35 iterate creatively. Moving through the 1:23:37 actual creative set. So they have a lot 1:23:38 of lowfi shots here. They have a lot of 1:23:41 UGC. They're obviously leveraging famous 1:23:43 people on Instagram that are wearing 1:23:44 their clothes. High production shoots. 1:23:46 They're also got back in stock what I 1:23:48 imagine is on their highest performing 1:23:50 products and they're driving straight to 1:23:51 the product pages. They're using audio 1:23:52 hooks on a lot of their videos. This is 1:23:54 actually underrated. A lot of faction 1:23:56 accounts that I've taken a look at do 1:23:58 not utilize audio at all. Potentially 1:24:00 people think it's off brand, but it's an 1:24:02 easy way to get better performance 1:24:03 through ads. When we look at landing 1:24:05 pages, here's their top landing page. 1:24:06 They're driving straight to this 1:24:07 particular product. Must be a top 1:24:09 performer. They're very So, they're 1:24:10 going and leaning into it. They've then 1:24:12 got a Valentine's Day landing page, 1:24:14 which is great. They're leaning into 1:24:15 this angle. And looks like they're going 1:24:16 pretty hard on the US cuz third biggest 1:24:18 landing page is the US and it's their 1:24:19 top seller in Australia. This is 1:24:21 generally the best way to crack into the 1:24:23 US market as a fashion brand is that you 1:24:25 take whatever your best sellers are in 1:24:26 Australia and you translate them 1:24:28 straight into the US. They generally 1:24:29 translate pretty well as long as they're 1:24:31 organically selling because it's 1:24:32 actually a good product that people 1:24:34 want. As we start to move through here, 1:24:35 they then have a bridals collection, 1:24:37 which is great. So, they've got 1:24:38 collections based on specific events. 1:24:40 So, it seems like an event heavy based 1:24:42 business that then got more individual 1:24:43 products. So, these I imagine would be 1:24:45 more bestsellers and they're going very, 1:24:46 very heavy at a product level. Now, do 1:24:48 they necessarily need to go this heavy 1:24:49 at a product level in Australia? No. 1:24:51 They probably have enough of a brand 1:24:52 presence. They probably have enough of a 1:24:54 following that they can go 1:24:55 collectionbased. And this is generally 1:24:57 what you see as you start to bridge into 1:24:58 8 fashion businesses in Australia. 1:25:00 In the US, however, I really do like 1:25:02 that they're leaning into individual 1:25:04 product pages because you don't have a 1:25:05 brand and so you need to crack the 1:25:06 market and you need to crack the market 1:25:08 through the best product that you 1:25:09 currently have. Now, there's no 1:25:10 partnership ads running. I say this in 1:25:11 almost every real I make these days, but 1:25:13 partnership ads really big opportunity. 1:25:15 So, I would be leaning into that right 1:25:17 now if I was in their internal marketing 1:25:18 team. And then they could probably do 1:25:19 some more instore stuff as well, 1:25:21 considering that they do have a retail 1:25:22 presence. actually leaning into in store 1:25:24 content is generally a very easy quick 1:25:26 win to be able to get a different 1:25:28 content format into the current creative 1:25:30 set. Country Road is a $1 billion 1:25:32 business and they're currently missing 1:25:33 the mark on Facebook. Firstly, only 160 1:25:36 ads live for a business that's doing a 1:25:38 billion dollars in revenue. Not a lot of 1:25:40 creative volume at all. 12 videos, 1:25:42 majority of them are DCOS, dynamic 1:25:44 creative. You should not be running 1:25:46 dynamic creative unless you are maxing 1:25:47 out your page limit. They're obviously 1:25:49 nowhere near that. So there's no reason 1:25:51 why they're running that as a creative 1:25:52 type. All it's going to do is limit the 1:25:54 amount of learnings that they can 1:25:55 actually get. Now let's take a look at 1:25:56 the actual diversity set of their ads. 1:25:59 So they have a bunch of spend and save. 1:26:01 Then they have a bunch more spend and 1:26:02 save. Then they have a bunch more spend 1:26:04 and save. Then they've got a few videos 1:26:06 in here. We can click play on one of 1:26:08 them. 1:26:17 So there's a few things I do like about 1:26:18 this. switching between a bunch of 1:26:20 different products is really good for 1:26:21 topfunnel. It makes the ad a lot more 1:26:23 scalable because it obviously resonates 1:26:24 with a larger audience. What I don't 1:26:26 like about it is that it's the same kind 1:26:28 of content type as everything else. It's 1:26:30 just a professional hi-fi shoot. But 1:26:32 let's continue to scroll through. We 1:26:33 have a bunch of statics here from 1:26:35 shoots. 1:26:37 Let's click play on this one. 1:26:44 Continue going through. So, it's all 1:26:46 very, very similar content. Yes, they're 1:26:48 different shoots. Yes, they're different 1:26:50 products here. Okay, we've got a mixup. 1:26:53 This looks to be an instore video, which 1:26:54 would be great. 1:26:55 >> Hi, I'm Leanne. Welcome to Country Road. 1:26:57 Come with me. I'll show you some of our 1:26:59 new collection. Something that I'm 1:27:01 really excited about, our new gingham 1:27:03 wrap jacket. This piece is really trans. 1:27:05 >> Great. So, it looks like they are taking 1:27:07 a little bit of inspiration from other 1:27:08 retailers and having people walk through 1:27:10 the stores and actually show the 1:27:11 product. That's amazing. It's a really 1:27:13 long piece as well, a minute and 52 1:27:15 seconds. So, it's good to see that they 1:27:16 have that in there. As we continue to 1:27:19 scroll through, then it's just it's all 1:27:20 the same stuff. So, there's really like 1:27:22 no diversity in here whatsoever, except 1:27:25 for that one walk through. Okay, we've 1:27:27 got another walkthrough over here. 1:27:28 Amazing. So, they're they're putting a 1:27:30 little bit in. Then, they've just got 1:27:31 statics with the sale amount over there, 1:27:34 which I'm sure this is crushing for 1:27:36 them, which is why they're probably 1:27:37 putting a lot of spend in it. It's been 1:27:39 active for a while, but it's because 1:27:40 it's just basic sales messaging. And 1:27:41 then we continue to scroll through and 1:27:43 it's all kind of the same stuff over and 1:27:45 over and over again. They're way 1:27:46 underwrited on every other content type 1:27:49 outside of hi-fi professional shoots. 1:27:51 Now look at this. The top performing 1:27:52 hook that's been live for 22 days. It is 1:27:55 a piece of user generated content. 1:27:57 >> Okay, Country Road has just released 1:27:59 their autumn winter homeware collection. 1:28:01 So come with me and check it all out. 1:28:03 >> Amazing. So triple down here. They 1:28:05 should be launching way more of this. 1:28:07 They should also be running it as 1:28:08 partnership ads. They shouldn't be 1:28:09 running this through the primary page. 1:28:11 And then lastly, there's a very low 1:28:13 volume of ads going live. They have 1:28:14 about 10 ads going live every 3 to 4 1:28:16 days. Not a lot at all for the size of 1:28:18 this business. I love ugly. 110 ads 1:28:20 running. What are they doing well? 1:28:22 Where's this opportunity for 1:28:23 improvement? 30 videos running, three 1:28:25 images. They're heavily leaning into 1:28:27 DCO's dynamic creative. This would be 1:28:29 because they're coming up against the 1:28:30 page limit on their ad account. They 1:28:32 have no DPAs running, which I love. 1:28:34 That's nice. They could probably put one 1:28:36 in, but none is better than hundreds. In 1:28:39 terms of the actual content that they're 1:28:40 running, they're leaning heavily into 1:28:42 partnership ads, which is amazing. This 1:28:44 is a huge opportunity right now. I've 1:28:46 been reviewing a lot of fashion brands 1:28:48 ad libraries, and most people aren't 1:28:50 leaning into them. They had to shut down 1:28:51 the US due to tariffs, but they've 1:28:53 recently fixed that. So, they're 1:28:54 notifying everyone in their creatives 1:28:55 that you can now shop tariffree. Where 1:28:57 is their opportunity right now? There's 1:28:58 not much lowfi content in here. All of 1:29:00 it is really high production, as you can 1:29:03 see. if I stop anywhere. All very high 1:29:04 production stuff, high production 1:29:06 shoots, very clean text over the top. I 1:29:08 really like the content, but nothing is 1:29:10 feeling very organic. Now, there's 1:29:12 obviously the partnership ads that feel 1:29:14 a little bit organic. We can watch this 1:29:16 one right here. 1:29:27 I don't mind it. Once again though, it's 1:29:29 very much so within the bounds of the 1:29:30 content style that they're trying to 1:29:32 maintain across all of the ads. Here's 1:29:34 another one that looks a little bit more 1:29:35 lowfi. 1:29:44 So this would be a really interesting 1:29:46 one to know performance and understand 1:29:47 how this is standing out from the rest. 1:29:49 To be fair, if we sort by the oldest 1:29:51 running ads, so these are the ads that 1:29:52 have run the longest ever in this ad 1:29:54 account. This one for example, ran for 1:29:55 nearly 3 years. They are all this style. 1:29:57 So it doesn't seem like they have had 1:29:59 great success if they have done testing 1:30:01 of lowfi more organic based content 1:30:04 because all of their long-standing 1:30:06 content has been all of these high 1:30:07 production shoots with offers on them. I 1:30:10 think one opportunity is they're not 1:30:11 using any auditory hooks on any 1:30:14 creative. So that's something that they 1:30:15 could look into, but that generally 1:30:17 lends itself to lowfi content. Then with 1:30:18 regards to landing pages, they're going 1:30:20 to collections, which is exactly what we 1:30:22 do 80% of the time. the top collection 1:30:24 being active wear, which is really 1:30:25 interesting. This seems to be an up 1:30:26 andcoming category for them that they're 1:30:28 leaning into, particularly with 1:30:29 partnership ads, too. So, I think active 1:30:32 wear can lend itself a little bit more 1:30:33 towards creating more organic content 1:30:35 and introducing that into the account. 1:30:36 So that's probably a pillow that I'll be 1:30:38 trying separated