#110 - Beginner’s Guide to Amazon Marketing Cloud AMC for Amazon Sellers

#110 - Beginner’s Guide to Amazon Marketing Cloud AMC for Amazon Sellers Speaker 1: AMC is the latest and greatest thing to enter the Amazon ecosystem. No, I'm not talking about the movie theater stock ticker. We're talking about Amazon Marketing Cloud. Speaker 2: This platform allows sellers to dive deep into their customer data and truly understand the path to purchase. But we keep hearing, I got AMC access, now what do I do? Speaker 1: If you're new to AMC, this is your beginner's guide showing you how you can leverage it to extract the most insightful data today on That Amazon Ads Podcast. Speaker 2: Alexa, play That Amazon Ads Podcast. Unknown Speaker: Which one would you like to hear? Speaker 2: The best one. Unknown Speaker: Okay, now playing That Amazon Ads Podcast. These gentlemen are completely changing the game. Speaker 1: After listening to That Amazon Ads Podcast, my ads are finally profitable. Unknown Speaker: I also heard they're pretty cute. Speaker 1: All right, Stephen, so AMC, super powerful tool within the Amazon ecosystem, relatively new. I actually did a poll the other day to see how many people were actually using AMC on a regular basis. And I was kind of surprised because it's a relatively complex platform, but 50% of people said that they were using this on a regular basis. So I wanted to do this episode to really just give people kind of a foundational understanding of the platform, if they do have access to it, how they can start to leverage it, even if they have no coding experience or SQL experience or anything like that. Have you personally used AMC much in your day-to-day? Speaker 2: I have not, and also 50% of respondents seems high. So we need to do another poll that says, how likely are you to lie on an anonymous LinkedIn poll? I bet over 50% of people in that poll were lying. About using AMC for whatever reason because is it not a requirement? Do you not need to have access to DSP in order to use the AMC? Speaker 1: Well, Amazon came out and they said that they were allowing people even if they didn't have a DSP account to start to onboard and have their own AMC instance set up. In my experience, Whenever I'm trying to do that with the Amazon rep that I'm working with for any given brand, they make it very difficult if we're not running DSP. And they a lot of times kind of beat around the bush. They never follow up on it. So it's been a little bit hard. So I think in general, they like it if you have a DSP seat and you're running DSP campaigns, but they Had announced previously that anybody could use it whether you had DSP or not. Speaker 2: Yeah, I think the only time I've ever used AMC before was when you and I were working on the Melissa and Doug account back at Edelman. I think enough time has passed that we can just say the brand names. It was like two years ago, so I don't think anyone cares. Speaker 1: Yeah, India is up. Speaker 2: Yeah, exactly. Yeah. I mean, like we were doing this back then. I don't even know if we had, did we have a ChatGPT at the time? Like 2021 ish? Speaker 1: I don't think so. And if we did, it was kind of towards the end. Speaker 2: Yeah. You were like, started taking SQL courses to like create audiences for us. Speaker 1: Yeah. I was trying to start to understand it. Unfortunately, I'm kind of an idiot when it comes to coding stuff and only got so far. I mean, I get, I get kind of what it's doing, but not to any sort of, uh, deep level. Speaker 2: Walk us through it, Andrew. How do we do it? What's the, how to make use of it? Speaker 1: Yeah, for sure. So just, just to give a little bit of background, you know, what, what AMC really is, is it's a robust analytics platform that allows you to pull custom queries on your Amazon shopper and ad performance data. So one of the cool things about it is that it has much longer look back windows. So while you're limited with your, Sponsored Product and those types of campaigns to just the last 60 days in the API or even if you have stored data, you know, it only goes back so far. This extends a lot further than that and it also gives you unique shopper data. So you can start to understand the trends of how customers actually interact. With your brand, through your ad campaigns, repeat purchase behavior, all types of stuff that you can really start to dig into on a per customer basis, which is super valuable in helping inform a lot of the strategies and kind of understand how your customer is finding you and what their purchase behavior is really like on the Amazon platform. So first things first, how do you get an instance? I think I kind of already alluded to it. You got to reach out to your Amazon rep and they'll help you set it up if you're I already have one. Great. This is going to be kind of an overview of how to start to leverage it with the tools inside of it. So we're going to jump into a screen share and start to dive into it. All right. So jumping into the screen share, we're looking at the Amazon Marketing Cloud dashboard. If you need to get here, you can just go to the measurement and reporting section and you should see it right over here on the left. Now, Amazon Marketing Cloud is a SQL-based platform, so it's actually slightly different than normal SQL. It's an AMC SQL, so it's kind of custom to Amazon, they have all kinds of documentation around what you can and can't do with the platform itself. But essentially, you write queries that reference different tables and different databases within Amazon system to pull different data points together. And it gives you complete customization capability. So you can We're gonna start to get really creative with this and blend certain data sets. So if you wanna look how things interact between sponsored products as well as DSP, you can blend those data sources together, something that you can't do in normal Amazon reporting. But this basically is what they call the query editor. So you'll see I've got a handful of tabs here that are pulling a bunch of different things. If you look over here on the left, you've got a list of some of the different tables that are available to you so you can reference and they basically just represent different data sets that you can pull from. By default, you're going to have access to the Amazon Ads data. So if you have a AMS account, DSP account, those are by default added to your AMC instance. But it does not include things like brand storefront data, and it doesn't include actual total sales data and total purchase behavior. So in order to get access to that, you need to enroll in some of the paid features. So you'll come to paid features. You'll see Amazon Insights, Amazon Retail Purchases, and Brand Store Insights here. So you can see I've subscribed to those. Speaker 2: Wait, scroll down to the bottom. Yep. Experian. Wow, that's really interesting. Speaker 1: Yeah. So, yeah, they got different data sets and stuff that you can connect as well. Haven't really messed with that. Speaker 2: So, yeah, if you sell, like, car parts, getting the Amazon Garage data and Experian car purchase data, I'm guessing, would be some interesting intelligence. Speaker 1: Yeah, for sure. The one that I think is most useful right now is just this Amazon Retail Purchases data. So this is going to give you access. Speaker 2: How much is that? Speaker 1: I believe it was like $200 or $300 a month, but they give you 60 days for free. Speaker 2: Okay. Speaker 1: And all of these, you get a free trial associated with them, 60-day free trial. So you can test it out and see how those inform different things. Speaker 2: Anyone else gives you a free trial? AdLabs. Make sure you like, subscribe and go sign up for your free trial at AdLabs. All right, go back to the query editor. Speaker 1: Yeah, for sure. Speaker 2: All right. What's under Advertiser Uploaded? Speaker 1: Yeah, so nothing there because this is where you can take your own data from your Shopify sources, all that type of stuff. You can actually upload it to Amazon Marketing Cloud to then create some like cross attribution or track purchasing behavior across different platforms. Not something I've done at this point, a little bit over my head. The subscriptions that I just alluded to are under that, obviously. You can show some of those different tables and all of these are just little fields that you would use within your queries to construct what it is you're trying to pull. So if you look over here at the query editor, it's basically just a combination of looking at different fields like first orders, user ID. We're referencing this Amazon attributed events by conversion time table. Which if you open this up, that's this table right here and it kind of starts to pull together some of those different data sets from those tables. Speaker 2: Can you mouse over the information icon on, it was like bid price on the last thing they had open there? Speaker 1: Sure. Speaker 2: Bid price. Okay. Sounds like it's based on what you're bidding, not like. And then also hit the drop down, close that one and hit the drop down under the sponsored ads traffic. Okay, pretty basic stuff. Speaker 1: Yep. Yep. Just kind of standard things that you would see within your account. Speaker 2: Basically, what this is, is that Amazon has a lot of data that is not in the ad console because it would clutter everything up with way too many columns, way too many things to manage, but they do have it. And so if you want these custom views, additional data points, you can just Create a query for it and pull that from Amazon's database and generate your own reports in the form of an Excel sheet. Speaker 1: Yep, 100%. And just to be completely transparent, I didn't write these queries. So let's get into how you might actually go about Constructing queries, these get pretty complex, especially if you're not, if you're an Amazon advertising, you're probably not also a SQL expert. And so either you're going to need to hire somebody who is and can learn this system very quickly and write these queries for you, or there's some other resources that you can leverage. So number one, within the use cases tab, there's going to be a list of a bunch of potential reports that you might want to pull or things that might be interesting. So like, Let's just say you're really curious about how long it takes for a customer to convert once you've displayed an ad to them. So you could look at the time to conversion report here. And they make it, you know, on some of these, they make it relatively easy for you to just, you know, input some information. And then some of them actually have query templates. And so you can come in here and, you know, utilize some of these pre-made templates. And like for this one, it's looking at the time to conversion over, you know, zero to seven day period. And, you know, this makes it a lot easier to have kind of a starting point for... Speaker 2: That's crazy that they're tracking the seconds. Speaker 1: Yeah, it gets pretty granular with that report for sure. Speaker 2: You would think that they would do like hours or something. Speaker 1: Yeah, well, some people purchasing pretty quick on Amazon. They've got the conversion down. Speaker 2: I want to do it now. Yeah, they have a less than one minute purchase. That's interesting. Speaker 1: Yeah, but these are really good for just giving you a starting point. Sometimes they're like good enough where you can just use them and get a decent output. Other times you got to tweak some things like sometimes you'll see within the query like specific campaign names where you can take your own campaign name, plug it in there. I find that that's a little bit difficult to do, especially when you're working, you're trying to get like aggregate data on a brand and you've got like thousands of campaigns. You know, it could be really daunting to try to input all of those into a query. And so, you know, sometimes I feel like you got to be a little bit more selective with what you're trying to do, or just hope that you don't have that many campaigns to actually plug in. These query templates are super helpful for just giving you a starting point. And along with that, I have personally found that ChatGPT is pretty good at writing these queries, especially if I have something custom that's not in this query library or template library. And it's something totally off the wall that I want to kind of look into. It's not perfect. Every time I use it, I have to go back and forth and back and forth and back and forth until ChatGPT actually gets it right. But there's a couple tips for how to use ChatGPT whenever you are trying to craft your own Custom queries. One of those is to... Speaker 2: Be mean to it. Speaker 1: Be very mean and very direct. No, one thing that I have found success with is if you go to... Some of these... Speaker 2: You upload like help docs or something? Speaker 1: Yes, so what I would do is figure out, you know, for the Amazon retail purchases, I just did this the other day, there's a list of fields that you could potentially use. So it's called a table schema. It's just like all the different possible variables and fields that you could input into your query and you give it that because then it kind of, then it'll know what can go where. I found that to be particularly helpful and then basically just some troubleshooting. So what I'll do just to show some examples here. Let's just scroll back up here. You can see. Where are we at? So yeah, I basically asked it to write a query to what we were trying to do with this was I have a client who has this theory that they see a bigger increase in new to brand customers in the first half of the year and it kind of tapers off and it's a lot more repeat purchases in the second half of the year. And they're, you know, a CPG like supplement style brand. And so they, you know, January is very big for them, get a lot of new customers. And so in order to validate it, I wanted to create a query where I could see new to brand customer acquisition versus repeat customer acquisition over time. And so I basically wrote a prompt. I can't really remember exactly what I said to it. But basically, just to get to the point, I go back and forth. So it writes me this query, I input it, I try to run it, and then it gives me an error. So I'm like, this gave me an error, I copy paste what the error is, and then it It rewrites the query. And then I just do that over and over again. Also failed. Speaker 2: You see at the bottom it says, ChatGPT can make mistakes. It's like, yeah, go figure. Speaker 1: Yeah, definitely does. Speaker 2: I've had such bad performance with ChatGPT lately where it's hallucinating all over the place and making up numbers that don't make any sense. So I have switched to Grok and Perpetua. Not Perpetua. Whoa. Perplexity. Perplexity, I will never use Perpetua. Speaker 1: Oh, good stuff. Yeah, I mean, it definitely messes up quite a bit. And there's some times where like, if you read what it's asking you, it'll kind of be helpful to go back and like this is I think this was before I uploaded like the table schema and stuff like I had to tell it what certain names or what certain fields were available. So I told it that, you know, just it's just a lot of back and forth trial and error until... Speaker 2: Is this within a project? Speaker 1: No, it's not. Speaker 2: That could also be helpful. Yeah, if you create a project, download the help doc information stuff from the AMC Help Center, and then it will just have all that information stored in that project. And then moving forward, every time you make a new prompt, it kind of has a baseline of knowledge. Speaker 1: Yeah, totally. That would be a good one. That'd be a good idea. You can see I finally got to a point where that data processed. I have a whole different query here now. Then I think I did one other thing, grouped the data by month to that query. Yeah, just kind of still some more troubleshooting every time it kind of messes some things up. But got something. Speaker 2: So yeah, how many rows do these Excel sheets end up being? And have you gotten to any length where Excel breaks because it was too many rows of data? Speaker 1: I haven't yet. I probably, I mean, I was working on one earlier that was, it gave me some kind of error within Excel and said like it couldn't open or something like that. But then I just did a different query, kind of told it, told ChatGPT what was happening. So sometimes whenever I'm running queries, I, it'll give you this little sample of the data that's actually going to output. And so with some of these queries, like you'll see, And I've seen that some of the rows are blanks, and I don't want blank rows. And so I'll go back to ChatGPT, kind of tell him what's going on. He'll rewrite the query. And then, yeah, that's generally what I see in terms of trying to limit some of those Excel issues. But it's not something I've run into a lot, actually, with doing this. And again, very green with all this type of stuff. So I'm not that experienced with it. I've done a lot of different samples and you can see some of the past queries that we've done. Really cool data around specific cities and postal codes like this brand is starting to... Speaker 2: You can get that for our sponsored ads? Or is that only DSP? Speaker 1: This was kind of like total level data, I believe. Speaker 2: Gotcha. Okay. Speaker 1: Yeah, so it gave us some insights. This brand's branching into retail and so I was trying to figure out like, you know, what specific postal codes, what specific cities were we selling the most products in, were we most popular in, to kind of help inform some of that. So there's a ton of things that you can do. I'll just show a couple examples and then we can start to wrap it up. But first things first, just to illustrate that example. So, new to brand customers. I mentioned before how this brand was saying they get a big influx of new customers in the beginning of the year, and they thought that it kind of tapered in the back half of the year. So we just use this to try to inform whether that was accurate and valid or not. And, you know, one of the biggest benefits to AMC is how far you can go back. So I think it's a five-year look back window with that total purchases table. So took this all the way back to 2022. And it just showed the new customer acquisition. So this is unique new customers entering to the brand purchasing for the first time over the course of several years. And we can see that, you know, that theory does kind of hold up. To some degree for 2024, where we do see things really start to taper off in especially Q4, but you know, August is pretty historically a big month for new customer acquisition. We also see that, you know, the first month of the year, relatively big for getting new customers this year, really starting to see things Pick up quite a bit. But yeah, super insightful for just helping determine the strategy for this brand. They were planning and are planning still currently to kind of cut spend, be a little bit, try to be more efficient in the second half of the year than we were in the first half where we were kind of pushing things and being a little more aggressive. So, you know, just kind of helps inform some of those strategic decisions. And then also, you know, super insightful to see the repeat customer trend over time to just see how How many different customers are second time purchasers in those same timeframe? So really starting to see things. Speaker 2: Pretty big sales growth in general for the last three months on both repeat purchases and new to brand purchases. Speaker 1: Yeah. Yeah. They're growing pretty rapidly, expanding the product line, all types of stuff kind of contributing to that. But yeah, super, super helpful for informing some of that strategy. And then just one other example here that I'll show is the time to conversion report. So this is measuring the how long between the customer's last interaction with an ad to actually purchasing. And, you know, this just helps inform how long your look back windows should be in any sort of retargeting efforts or just kind of gives you a sense of how long people are taken to purchase. As expected, a lot of the purchases are 50% happening within the first 30 minutes of the last ad interaction. But one insightful piece was that there's quite a few people and customers that are actually purchasing between one day and 30 days later. So we'd set our retargeting windows to kind of sit within that window. I'm making sure we're hitting those customers again if we can. And then once we get outside of that 30-day period, it's pretty much non-existent. Yeah, just interesting insights that you can gather with, you know, those longer dimensions. And I think, you know, in that time to conversion report, originally, it only goes from, you know, less than one minute to seven days. I use ChatGPT, just told her, hey, I actually want this to go all the way out to like 60 days, create some different cohorts and time buckets and all that type of stuff to analyze a little bit more deeply. So you can use those query templates and things like that as a base and then tell it what you want to do and how you want to change it. Go back and forth and pretty much figure it out from there. Speaker 2: But makes sense. Speaker 1: Yeah, man. Anything else to add here? Speaker 2: Got one final question. Maybe this is a topic for another time, but on that other Excel sheet, you're showing new-to-brand audiences. If the client's really trying to push there, I think you can create a new-to-brand AMC audience and then increase adjustments on those on the sponsored ad side. I have like a positive bid adjustment. Have you done that for that? Or you can only pick one audience per campaign. So is that the audience you choose? Speaker 1: Yes. Yeah. I pretty much just only tested new to brand at this point. Speaker 2: Is that why we've seen growth there in the last three months? Speaker 1: I don't think so. It's only been deployed within like the last month or so. So it's relatively recent. But yeah, that'd be a whole other topic. I mean, the audience piece. Speaker 2: Go back to the other chart. Yes, it failed. Speaker 1: It came down a little bit, but yeah, I just so with sponsored brands that new to brand audience is one of the defaults. So you don't even have to. have AMC in order to be able to add that to your sponsored brand campaign. So that's a nice thing. But with the sponsored products, I did have to go and actually create that new to brand audience within AMC. And you can do that under audiences. And there's actually a template. So I come in here. Once you're in the audiences tab, you can come to create an audience and you'll see that you can either write a custom query or you can use an instructional query. And it has some pre-built here for audiences. We're looking at, you know, no code stuff, so pretty easy to work with. But where is it? Audiences of high value, new to brand customers. I think this is it. But then you can input your advertiser, name it and set that date range look back window for the audience period. But yeah, then you'd create that audience there and then you have to give it a little bit of time, but then it starts to show up in your sponsored product ad account and you can add it to those sponsored product campaigns. Speaker 2: Well, Andrew, thanks for coming on and let people know where they can get in touch with you if they have any questions. Speaker 1: Yeah, make sure you like and subscribe to That Amazon Ads Podcast. I'm here every week. We'll see you next week. Speaker 2: And drop your comments below if there's anything that we didn't answer. If you want another episode covering more stuff or just a quick question that we can help answer in the comments, let us know and we'll see you next week on That Amazon Ads Podcast. Speaker 1: See ya. Unknown Speaker: Thank you.