A (2:48)

For most people listening, I'm going to assume that not everyone has an advanced SQL knowledge here. Yes, because there's two Ways that you can kind of use this tool for both audiences and analytics and one is using their pre built audiences or analytics and reporting or to create your own SQL query. That said, they do have an AI agent that helps you write SQL queries now, so that's at least more helpful than it was prior.

B (3:23)

Give me an idea of a SQL query that you'd write with AI or if you could write SQL like this would be to, you know, find repeat purchasers or give me an exam, give me something tangible that we'd use it for.

A (3:34)

Yeah. So they already have these pre built audiences essentially. So you can go in, click that, fill in a template. But let's say that doesn't satisfy your need. Say you actually want to go out and target customers that browse certain nodes or specific products. That's where you can then use the AI agent to try and get it to do something more specific like that. That said, you will still have to know a little bit of SQL or you'd have to just improve your ability to prompt it and maybe you can even take it out, bring it into Gemini, try and prompt again from there to try and optimize it. But I would say anything where a pre built audience doesn't help you, that's where you're going to have to use this AI agent.

B (4:22)

And how are some of the ways we the examples you know, chat came up with here to retarget high intent shoppers, to exclude recent purchasers who aren't likely to purchase again. Build lookalikes based on high LTV cohorts. If we've used all three of those use cases.

A (4:36)

Yeah. So I think with every audience it's nuanced to your brand essentially. So there's a lot of considerations I would say before you start applying audiences to certain campaigns that you'll have to consider just because you can start to apply some of the things that you just said. But it actually might limit your entire audience and your campaign now goes too narrow and you're actually not converting on an audience base that you previously were converting on. So it's super important to keep at least net new campaigns that you build separate from your older campaigns.

B (5:17)

Yeah.

A (5:18)

And then you can test in isolation.

B (5:20)

I think that makes sense. Cause I guess with all, all ad platforms they generally want you to buy more and then you know, find out what, what works. They're not always so sometimes when you come in with too many narrowing things in the beginning it might not work for you. And also like do you find that when you put more custom Work into an audience, it's more expensive so you.

A (5:42)

Don'T actually have to pay for audience creation.

B (5:45)

I mean, to reach them.

A (5:46)

Yeah, to reach them. So, so that's, that's where it gets interesting because really you're using like that ACOS measure in your campaigns to really understand how much you're spending to reach that audience. And Amazon's, it tries to be really good at putting your product in front of the right audience, but in a lot of cases it may not do that efficiently for you. So that's where you need to intervene by creating your own audience and one that you have a high confidence will actually convert for specific keywords or asins, et cetera, which is cool because that.

B (6:23)

Allows you to maybe have a more targeted approach when it comes to growing your rank on some of those more competitive keywords maybe.

A (6:30)

Exactly. And you know, not every brand is going to be able to compete on the most generic keywords. Let's say I want to use an example of a brand that we're working with, but I'll, I'll take their product and now it's a cooler. So let's say it's like a, a rugged cooler that you want to bring outside. You don't want to bid on just cooler for that type of product because it's got additional features, it's rugged, it's meant for the outdoors. Some people might be searching for that, but you may actually be able to overlay an audience where you're still targeting generic keywords. So let's say it's just cooler. But then you overlay an audience that you selected based on customers that browsed certain nodes and those nodes were camping and hiking or you know, sports and outdoors. So then you're targeting this audience that might be already qualified for that type of product and then now you're willing to bid a higher bid on that broader or more generic type keyword.

B (7:33)

That's interesting. Talk to me about the analytics, analytics side of things. I know there's so much power that can be unlocked when you dive in there. What are some of the instances that we're sort of, again, I assume it's a thing with Amazon where there's going to be some pre built templates for reports as well as some ones that we like to build. What do we like to do with the analytics?

A (7:51)

Yeah, so analytics, I would say for most people you'll have to stay to the no code pre built ones. And I'm actually looking at it in behind here now, right now. So they essentially have, I Don't know, it's about like 15 pre built analytics or reporting that you could pull. When you go beyond that it starts to get very complex. And even using the query or the AI agent to write your queries, there's a really big breakdown in the type of reporting that you get back. And a lot of these like really big brands are using APIs with the querying and the reporting. So what I would recommend is to stick to the pre built ones that are in the console and then it's not pretty, but I think it doesn't have to be pretty for you to get insights from what that reporting is telling you. So just to kind of list off a few, you can actually look at how multiple campaigns work together to drive an increase in conversion. So you could take your campaigns from your ad suite and then it'll say this campaign, this campaign and this campaign, when used together it increases the conversion for that product by X. But when it's only this campaign and this campaign, you don't really see a conversion increase or in fact it's very low. That can kind of help you understand. Oh, okay. These are actually the integral campaigns that I need to either invest more into or make sure that we're running to ensure that we have that conversion for our customers. So it's good at giving that context and then you can maybe even overlay an audience or you know what to turn on and off from a campaign perspective based on that reporting.

B (9:46)

In talking with the Amazon team over the years, so much of our wins on the pilot house side come from being efficient using the tools that are available. I think so many advertisers maybe in the Amazon environment have a bit more of a set it and forget it, you know idea when it comes to toggling these things and overlaying them to find efficiencies. But it sounds like this gives us a lot more toggles. The one, the one thing that I don't this is what chat suggested is you can see what asins are awareness versus conversion or potentially even like you could probably look at which ones are driving like repeat, repeat buys versus the ones that are more one time purchases.

A (10:21)

Yeah, so you could definitely do that. And there's some pre built audiences that allow you to get to that answer somewhat. A lot of what I think Gemini or if you're using like ChatGPT to get this answer, they might be providing use cases from brands that have like advanced analytics or query writing. So some of it might not be available for everyone. But I would urge everyone to Just go into the AMC dashboard, you can find it in Measurement and Reporting and then click no code and analytics in the use case section and just review what type of reporting is available for them to fill out a template and run.

B (11:08)

And then just in terms of you says open to all brands. Now previously it's just been for larger brands and agencies. When should a brand be thinking about like diving into this stuff? Obviously go check it out, but when does it really become important with, with the size of a brand to. To, to start working on this stuff?

A (11:22)

I think that's a really good question. I would say from a reporting perspective that threshold is probably higher because there's more development hours. Let's say that you would need to put in to go and actually get the results from the reporting. So you probably want to make sure that you're spending over a certain amount of dollars, let's say at least like 20k and spend per month. I would say at like the bare minimum, before you want to start digging into the reporting element from an audience perspective, I would say anytime that you have an expensive campaign and you've done everything else so you've, you know, ensured that your content is telling the story that you want it to tell, that your title has the keywords that you expect and want to be in there, and you're at the price point that you think makes sense for the value of your product. After all that's done, then you can say, okay, this is still an expensive campaign. Maybe I do want to overlay an audience that makes sense onto this type of campaign. And I would say the threshold for that is small because if it works, you've now saved yourself money that you can then go reinvest elsewhere or continue reinvesting in that more efficient campaign.

B (12:41)

Now, on the pilot House team, are we diving into the SQL stuff? Do you write SQL queries?

A (12:46)

I don't personally. Like there was a time in my life where I played around with it, but I would say it's more like AI agents that we will use in prompt. We don't have any like expert SQL writer on our side.

B (13:04)

This is all very interesting. I'm just curious, in your day to day with with Amazon, what are some of the ways you're interacting with AI? Obviously the AI is in their tools, but are there other tools that you're using for other parts of the job, whether it's reporting or targeting or anything?

A (13:17)

I would say that it's really helpful in trying to, let's say, think through like the next generation of content that you need to write or create for let's say Rufus, you can kind of go back and forth with like Gemini to help you ideate and think through some of that stuff. I think we can use it for data sets as well, but I don't think it's like nearly as good as us just creating that template ourselves and then copy and pasting that data in for the report to show up the way that we want it to. I'm sure there are ways that we can prompt it and use it to get even more insights, but I wouldn't say it's being used for that aspect on a daily basis.

B (14:04)

All right, so listeners out there go take a look at some pre built audiences, test out a templated report. And I like your idea of just like after you've, you know, if you're, you're bidding on competitive terms, you've tried all the, all, you know, all the important T crossing and I dotting, then see if you can overlay an audience that puts people even more in market. Because there's a lot of audiences that are about like where people are in the funnel as well, right?

A (14:27)

Yeah, exactly. And, and that's actually one of the cool ways that you can use this because a lot of people will question their DSP spin. So demand side platform dsp. A lot of people question the incrementality that they get from those types of campaigns. And now you can actually run a pre built analytics report that says, hey, once they've seen your DSP ad, how do they actually respond to your sponsored brand ad? Or how do they respond to your sponsored product ad? And that's where you can start to overlay it and get some pretty cool insights on that front. That's actually what I would urge people to use it for. And that is a pre built audience. So you can just say, hey, how does my advertising campaign overlap and work with dsp? And it'll give you this report to show what your conversions are when certain DSP campaigns and certain ad campaigns work together.

B (15:26)

Great use case for the listeners to go check out. I always like to Google Amazon or Jeff Bezos and end our Amazon podcast with some topical news. I just saw that they laid off 16,000 people yesterday. I think I'd heard that. But is there anything you've noticing on the platform after the layoffs?

A (15:42)

No, nothing that, that we've seen or called out at this point. But yeah, I think last time they said when they laid people off that it was more of like a culture thing and it wasn't related to AI, but it's kind of in the ether and it's up for us to determine because I don't think they really actually tell us why they they've done it.

B (16:04)

The news today is that they're in talks to invest up to 50 billion in OpenAI. Because I think I actually just saw like the rate car around yesterday for the proposed rate cards for what OpenAI might be charging CPM wise for ads. And I can just imagine the connection between OpenAI and Amazon will be massive. Like just how just it's the easiest place to purchase things. So therefore if it's giving product recommendations, it's maybe most likely to do it in Amazon.

A (16:29)

So then you're saying that there's like a partnership between OpenAI and Amazon for OpenAI to recommend Amazon products.

B (16:38)

That's what I guess I'm just looking at headlines and they are, they're sort of in talks. The news today, as of like 24 minutes ago or something, was that they're going to invest up to 50 billion in OpenAI. So that, that to me is where my mind goes because the CPMs I saw yesterday were astronomical because it's not really, it's not social media, it's a, you know, a concierge that's taking you directly to what it thinks you should buy. So I'm sure the conversion rates are going to be quite a bit higher. But it's going to be really interesting to see the way all these platforms vie for attention in open AI. And it looks like Amazon's making quite a bid right now.

A (17:08)

Yeah, because before that I think they were giving favor to Etsy. Ah. But I mean, Etsy can only do so much versus Amazon. Like it sounds like OpenAI plus Amazon is the ultimate partnership.

B (17:21)

Yeah, you heard it here first, folks. Tyler, thanks for coming on the pod today.

A (17:25)

This was great. Thanks for having me.

B (17:33)

Thanks for listening to today's episode. If you're not getting the DTC newsletter, you can subscribe for free at directtoconsumer Co. And if you want to learn more about Pilothouse's all killer no filler services, take off to Pilothouse Co. I'm Eric Dick and this has been the D to C podcast. We'll see you next time.