How to Turn Amazon Into a DTC Profit Powerhouse

[00:00:00] Richard Gaffin: Hey folks, welcome to The Ecommerce Playbook Podcast. I'm your host, Richard Gaffin, Director of Digital Product Strategy here at Common Thread Collective. And I'm joined today as I usually am by Mr. Taylor Holiday. Second week in glasses, looking good. How are you doing Taylor?

[00:00:14] Taylor Holiday: Yep. Still disorienting. Like, when I take them off, I feel dizzy. I don't know if that's a good thing or bad thing, but it's we're, we're making it through, man. We're making it through.

[00:00:21] Richard Gaffin: Gotcha. All right. Well, so let's press on then through our topic for today, which what we want to talk about is how do we incorporate Amazon into forecasting your entire digital enterprise? And I should frame that by saying We're expanding our profit system, which is, of course, the operating system that we put all of our clients through to build daily expectations for profitability.

We're bringing Amazon into that. So what we want to talk about today is some specifics around what it takes to include, incorporate Amazon into your forecasting. So Taylor, why don't you give us a little bit more context around why we're bringing this in and then kind of how forecasting Amazon works.

[00:01:01] Taylor Holiday: Yeah. If you are a business that has a large and growing Amazon and DTC business, this is a critical episode for you. As we are doing going through the process of building 2025 forecasts for many customers, one of the big topics of conversation is how we begin to move out of viewing these as separate siloed channels with separate demand creation profiles into one digital, one view of the digital enterprise, like where it's one channel with the media spend driving incremental demand in both places.

And we believe this is a mission critical change to the underlying ways that businesses are viewing their performance online. That has to happen or the businesses will end up suffocating their direct business by isolating all the media performance expectation into one channel while creating a simultaneous competition for the demand that they're creating in a separate channel.

So it is so important and complicated and hard. But really, really critical to the success of the future growth.

[00:02:09] Richard Gaffin: Gotcha. So let's dig a little bit then into maybe the mechanics of how people are doing this wrong now, which is to say, why do the two channels end up cannibalizing each other?

[00:02:20] Taylor Holiday: So in many cases, we see the business sort of bifurcate DTC as the. com revenue and marketplace as separate revenue with separate PNLs. And what they do is they take all the media budget that exists on the channels where the top of funnel demand is being created. So think of that as like meta or YouTube or television or these other channels, and they isolate all that media impact.

Onto the assessment of the. com PNL and then only put Amazon media spend. So that's buying within the Amazon platform itself. So search ads, buy box, et cetera onto the Amazon PNL. But the evidence is definitively clear that that media spend on the YouTube, on the Facebook, on the, all those channels is driving incremental demand in both places.

And especially for those brands that have duplicitous merchandise in both places, you have absolutely created competition for the realization of the demand on your. com with Amazon. Some people just like it more. They're always going to shop there if they prefer it. And so when you create demand with a Facebook ad, it's going to get satiated in both places.

If you don't figure out how to account for that, it's going to look like your. com performance is falling off. When it's really just about the growth of the overall business and the ability to allocate the media against the performance in both channels.

[00:03:42] Richard Gaffin: Gotcha. So how, so how does our system then help you tease these things apart or determine where each is coming from?

[00:03:49] Taylor Holiday: So we have to begin by building all of our forecasting models, both at the new customer level and the returning customer level. With data from both channels. So you have to build your cohort of new customers to include Amazon customers. You have to build your overall spend and new customer acquisition efficiency, inclusive of the Amazon spend and the Amazon revenue and all of the spend so that you get to an overarching budget to grow new customer revenue across both channels.

And then you have to build your cohort specific LTV forecast using both channels to get to an initial modeled expectation. Of where the overall revenue growth and the budget is going to come from. So we're going to treat, treat it as one revenue, one spend and one revenue to begin. Then we're going to go into a detailed layer that where we begin to create expectations in both places, but the media budget and revenue has to be total from both places.

[00:04:43] Richard Gaffin: essentially, it sounds like the beginning of this exercise is to bucket everything together. So to what extent is teasing them apart or determining where the final purchase is coming from? To what extent is that part of this? Or is that important?

[00:04:59] Taylor Holiday: So if we think about right now, what we do is we build what we call like a spend an AMER model. So it's, it's spend against modeled against the efficiency of new customer acquisition. And we use all of your media spend against your. com revenue and your. com new customer revenue over time. And we build a relationship between those two variables to start.

What we're doing is now also introducing new customer revenue from Amazon. Against your overall spend so that we can begin to see the relationship between the spend and your revenue in both places. And we'll add in the Amazon spend as well from there. We then want to get to the question of the more specific allocation and expectations in both places.

So, if we understand, and that this requires a layer of measurement testing that has to proceed this process, which is, you have to run a geo holdout. Of the media impact on both. com and Amazon so that you can get to an incremental ROAS target for your meta spend for your YouTube spend that includes the impact on Amazon.

This is the key. Otherwise you won't be able to understand that if I spend a dollar on meta or if I spend a dollar on YouTube, what is my return going to be on. com? What is my return going to be on Amazon and be able to model it out. But if you have that information, you can actually begin to understand that, oh, when I spend a dollar on meta, my incremental ROAS on my.

com exclusively might be a 1. 2, but it turns out there's an additional 40 cents captured on Amazon. So my actual incremental ROAS is 1. 6. That will transform the way you think about your media allocation and measurement. Of the overall business and will fuel more growth because what I see happening right now is that as people grow their Amazon business, their.

com business efficiency of acquisition begins to degrade. Why does that happen? It's because they functionally created competition for themselves. Of the efficiency of that media by offering on Amazon, the same product in many cases on prime with a better fulfillment to customers who are so patterned to purchase there.

And so they look at it and go, why is my. com media efficiency declining? And in the danger is that you actually start to pull back and then you end up sort of on this death spiral, which is that by not. Allocating the media against both places. You actually makes you think that your spend is not working and you pull back.

And it's only because you created actually competition for the assessment of that media and it becomes a death spiral.

[00:07:33] Richard Gaffin: Gotcha. So, so the issue is that like the, the, let's say the, not cannibalization, but I know, but the, the inefficiency or your DTC inefficiency going down is sort of inevitable because Amazon just is this juggernaut that is going to do what it's going to do. The issue is. Perhaps like maybe over correcting or thinking that DTC becoming more inefficient is overall a bad thing when actually you need to think about it holistically and say like your, your business is growing.

So, so what's the problem here? Is

[00:08:05] Taylor Holiday: That's well, right, exactly. And like the way you have to think about this is that when you create demand for your product, and this is even more true and impression based environment, so even more true on television and YouTube is that the satiation of that demand, in other words, the realization of that demand or where the customer ends up purchasing does not follow the same way.

Direct linear path as a click does it follows. It shows up whenever the person then encounters the brand in retail, in marketplace in anywhere else, there is an impact in all of those places. And the data that we have seen, Olivia Corey from house has published some great stuff is that there are cases that their research would say that in every case, there is incremental lift on Amazon.

What, like almost, I think that it was like 85 percent of cases. There's incremental lift on Amazon and half of the time it's over 50%. So imagine that you're running your Facebook media and you're like, ah, I'm barely breakeven. It's like a 1. 2. But then you find out that there's a 50 percent incremental impact on the Amazon revenue.

Well, now all of a sudden your Ross went to 1. 8. That may completely change the way you think about scaling your ad dollars, because suddenly bad performance became good performance by understanding the impact that it's having in other places. And so I think that that's what people need to understand is that there is, there is impact on Amazon.

The idea that Amazon is generating all of its own incremental demand is not true. Amazon in many ways is like a search engine. It's just a place people go to look for a thing that they've already sort of, seen elsewhere in the world. And so in many ways, you are driving that demand with the advertising that you're doing.

And so I think businesses need to start to assess these things together and the bifurcation of them is one of the biggest limiters in the way that businesses are structuring the view of their own results.

[00:09:58] Richard Gaffin: Gotcha. So actually, so one way to think about it maybe is that in a sense, Amazon as a marketing mechanism is kind of more like Google search almost like it's just sort of a alternative platform or like a Google shopping alternative. Even though Google shopping is more of an Amazon alternative, let's say, but that that's kind of how you need to think about it.

[00:10:16] Taylor Holiday: Exactly. And you'll see data, right? That says like half of product searches begin on Amazon. It's yeah, well, people go there with shopping intention. And don't get me wrong. It's not to say that there isn't potentially incremental ad dollars to be spent on categorical search on Amazon.

Of course there is, but that, that doesn't change the fact that there is some amount of demand that you're creating off platform. That people are showing up looking for your brand or looking in your product category with an eye for the ad that they saw on Instagram or otherwise. And that's just the reality of it.

That happens. People price compare though. They may even get the ad click to your website and see if it's available next day shipping. Like I think about one, I'll give you an example. One of our customers skull candy. Okay. Let's imagine you're a headphones. Customer, and you decide that you need headphones.

Well, oftentimes there are sort of two different categories of purchase for headphones, right? There's like, I'm going on vacation next week. And I want headphones for the plane. Like the value proposition there is almost entirely the immediacy of getting the product. I need it right now. I'm going on this trip tomorrow.

I need my headphones today. Like Amazon is just going to be a better solution for that because your website likely can't compete with the pace of delivery for that problem. So it's like an acute need problem. That's very different than like, Oh, you know what? I'd like a new set of headphones. That's the best value.

And you're going to shop it over time. Those are different mechanisms, but there's this whole category of purchasing that's like, I need this thing now that Amazon is going to satiate better than your website. I keep saying that word. I'm not going to do it again, but I think that understanding that for your business and realizing that some of the especially when you create the duplicitousness of your merchandise, meaning the same product is available in both places and one has free shipping and next day delivery, and one is 5.

99 for shipping and comes in a week. Like if you create demand that can be realized in those two different ways, the customer is going to choose the better fulfillment option a lot of times. And so that's cool. It's actually, if you're choosing to go into Amazon, we don't want to fight that fact necessarily, but we just want to make sure that we're clear that there is an impact that our advertising is happening on that purchase behavior.

[00:12:21] Richard Gaffin: Gotcha. So a couple of questions here. So one is this, the way that you're framing this, this approach seems like pretty, pretty obvious. So my question is like, why? Why, why are we only doing this now maybe? And why hasn't this been a way of thinking about it for the

[00:12:37] Taylor Holiday: Organizational design in almost every case, the person who's in charge of the website, is it, isn't necessarily in charge of Amazon, right? Or at least that's how we see it a lot. So there's just, if I'm in charge of the Amazon P& L, I don't want any of the media dollars on my P& L. It's great for me.

My margin looks awesome. I look like the darling of the business, right? So I think part of it is like who has authority and do we view these things as one revenue opportunity? And then the other is just like people don't usually like They'll try to, to, to model them both as if they'll just keep growing at the same level of efficiency.

Like that's what everybody wants to happen. But the, but the reality is, is that there's some tension that as you introduce more and more and you've grow the Amazon business and make that better and more awesome, that there is a trade off, these things aren't just purely incremental to one another, there is some competition that gets created Against those outcomes.

And so I think there's, there's reasons why, based on how the organization is expecting performance in both of these places, whether it's because individuals have different incentives or the organization has just poorly modeled the growth and efficiency of the future. com revenue, independent of what they were doing on Amazon, as if that was going to stay the same, because last year when Amazon didn't exist, my efficiency was here, but now Amazon does exist, but I still want my efficiency to comp to last year because year over year comps are like.

The ultimate degrading idea inside of an organization, but they're so pervasive, even though the environment might be completely different. So all of those things contribute to a reality where people don't combine them in the same way.

[00:14:02] Richard Gaffin: So what about the, the pushback maybe from especially like smaller brands around the, I don't know, maybe the right word is the inefficiency of Amazon in that Amazon takes a bigger cut of what you're selling relative to DTC, where more of the money goes to you. How does that dynamic play into this? If at all,

[00:14:20] Taylor Holiday: Yeah, for sure. You have to model the gross margin in both places and understand then how that gets to like an incremental marginal ROAS, right? Like ultimately the expectation based on the revenue realization is different. And this is, this is all where like channel expansion creates increased complexity.

And there's an understanding and trade off that has to be considered in that process, which is that the more you make your data harder to understand for yourself, The more challenging it becomes to make good decisions and the more likely you are to make bad ones. And so this is where it's like, I would just encourage brands.

I think Amazon is not thought through enough in terms of what the strategy is in its rollout and how it's considered in its interaction with DTC. So this is an opportunity, I think, as you head into 2025 to stop and go, okay, what is my growth expectation in both channels? How am I looking at the media dollars impact in both places?

Do I have a read on how meta in particular, it's likely your largest spend bucket is affecting Amazon. If not, you need that answer. And then I would start to say like, as you begin to want to expand into more impression based advertising. So again, I'll use YouTube as a channel that we're seeing great, some better incrementality reads.

When this is included than others, like this is really important because the click based assessment is not going to work for you. It's not going to pencil. And you'll be like putting your thumb in the wind without having a clear understanding of how it's impacting every place that dollar shows up.

And look, there's a whole third category here, which for many brands is the biggest one, which is retail, where this just act. This is actually does apply to but for the sake of. Our customer base, what we're primarily trying to focus on first is the, well, we're calling the digital enterprise, which is their.

com and Amazon businesses and how those things work together. To produce outcomes.

[00:16:01] Richard Gaffin: Yeah, that's cool. No, I just sort of had a stray thought that it's interesting to see us moving into the world of, like, like you were saying, more impression based marketing or almost awareness based marketing, kind of like a little bit of a throwback to, I don't know, the nineties or something like that, where you showed a TV ad and then you showed sort of a lift study on the side and said like, this probably worked because this went up and this went up and it's, it's getting a little bit closer to like, it's not one click, one purchase, perfect attribution.

It's a little bit of us thinking about how do we incorporate. this necessary phase of expansion into something that's as precise as the profit system, I guess. Does that make sense?

[00:16:38] Taylor Holiday: Yeah, I think the promise of DTC and really what's still true is like the simplicity of advertising and measurement. But there are real limitations for many brands about getting tranche of growth because. And e commerce is not growing that fast. Like e commerce is a growing category, but let's say it grows 15 to 18 percent every year.

My experience is that most brands have a growth expectation for themselves. That's bigger than that. Now there's probably a whole separate conversation of whether it should be bigger than that or not, or if everybody should slow the hell down, which is like, I think a different conversation for a different day.

But for today, my experience is that nobody's writing down 2025 to be 15 percent growth. And so if you're going to outpace the growth of e commerce. Then you're going to have to do that in some unique way. And it's probably not all going to persistently come from just growing your. com revenue, 50 percent year over year, at some point, that isn't the plan that works.

And so businesses, especially as we move into larger customer bases, who have come off of years of growing a hundred percent, then 70%, then 60%, then 30 percent are now going. How are we going to get more growth? And the. com doesn't seem to be the place where maybe we could write down 10 percent in that channel next year, but that'll be a grind even.

And now instead, what we have to figure out is expansion into other channels. It often comes with increased business complexity though. And you have to be ready for it. You have to be, make sure you have good information to make those decisions.

[00:17:57] Richard Gaffin: All right. So let's give let's give the folks just like one in lieu of joining us for the profit system. What's one simple step that people could take to start down the path to more Amazon clarity.

[00:18:06] Taylor Holiday: Well, so I do want to, I do want to plug why I think you actually, this is a thing where it is complex and I can't make it too simple for you. And I, I want to tell you a little bit about how we're doing this. So one is we've built statless to have an Amazon integration, which is, is complicated in and of itself, because one of the things that's natural about the Amazon API is that it doesn't give you very long history of data.

So without specific relationship with Amazon, you're only going to get like three or four. 30 days of ad spend data that makes it really hard to build these kinds of models. It's also not super simple to get new and returning customer revenue cohorts and other things. So there is some complexity around the Amazon data integration.

They're not the friendliest with sharing all of their data. Let's just say that. It's also, we have to acknowledge like the inability to de dupe customers between Amazon and your. com because. They don't give you any PII. They're like, there's no identifiable customer information. And they're really protective about that.

So that's just a limitation as well. So again, increased data complexity here requires some sophistication, but we've built that Amazon integration to be able to look at these different models. And then you have to do the good data science. And this is where I think you could. Do some simple work, which would be when you go to model the relationship between your spend and revenue to forecast next year, include your Amazon revenues to like, that's like the simplest step export all of your Amazon revenue, as well as your um, com revenue.

If you can get it new and returning, then great. Do that. And then try and build. That relationship between those numbers so that you can start to get an estimate of impact. Right. We like to prescribe to this idea. We call best available truth at CTC. If you don't have a geo holdout study, the best you could probably do is run a correlation between the spend on meta and the revenue on.

the spend on YouTube and the combined revenue between the two. You could look at all of those things on sort of a regression analysis about the historical relationship between those things. That's a good sort of substitute that I would put below an incrementality study to give you an indication of impact.

Cause that's what you just want. You need to, you need to get the evidence for yourself that there is impact in a. com with your advertising. The other thing you could do is a simple exercise would be to just amortize The spend between the two channels. So if you are keeping separate P and L's and Amazon is 20 percent of your revenue and DTC is 80 percent of your revenue, you could put 20 percent of the media spend onto the Amazon P and L and sort of burden them appropriately so that you can actually get a sense of the margin of each channel.

So I would say if you're leery, right, if you look right now and your Amazon P and L looks way better than your. com P and L, but there's no ad dollars on it. That would be my red flag. Like stop, stop, don't go into next year with that situation because that is lying to you. That's deceiving you into thinking one channel is producing amazing growth while the other is sort of stuck in inefficiency.

Or if like your Amazon margin showing up and being reported as way better than your. com revenue, red flag. Those are all things that I would just stop and begin to reassess how you're looking at the channel.

[00:21:01] Richard Gaffin: Gotcha. Cool. And you heard the man. If you want somebody to help clarify the complexity for you, reach out to us. CommonThreatCode. com. Smash that hire us button. Let us know you want to talk and we'd love to chat about bringing you into this. Taylor, anything else that you want to hit on this topic?

[00:21:15] Taylor Holiday: Yeah, we're going to give you a full forecast for 2025, inclusive of a spending AMER to include Amazon and your. com business. We're going to give you a measurement roadmap to get those incrementality holds, holdout studies to, so that you have those definitive results. We're going to put those results into action in terms of setting and executing it against the channel targets in a way that's going to bring this all to life for you.

So if it feels overwhelming to create that process, It is overwhelming. It's fricking hard. We've spent a lot of time trying to think through this problem. But we've got the tooling updated. We've got the integrations in place. We're building the models. So we'd love to help put that to put that to use for you.

[00:21:50] Richard Gaffin: All right. Cool. All right, folks. Thanks for listening. We will chat with you next week. Goodbye.