Login
Take a tour
Book a demo

Book a demo

600+ data sources
"Democratizing data for marketers - A great low-no-code ETL tool

Allows my team of marketers who are not data engineers to extract and transform data without having to learn python."

Blog / Here's how Amazon's AMC Uncovers Patterns in Customer Behavior

Here's how Amazon's AMC Uncovers Patterns in Customer Behavior

In episode two of the Undiscovered Metric, we're joined by Marco Senftleben, Senior Solutions Consultant at Adverity, to talk about Amazon Marketing Cloud (AMC). Check out our video to find out how AMC can help marketers uncover patterns in customer behavior in much more granularity by combining data from different sources in a data clean room.

Find out how to connect your Amazon data in our video

Hi Marco, can you tell us a little bit about yourself?

Back in early 2000, I started in digital advertising with a big German publisher house. At the time, the most used search engine was called Fireball, and that’s where I was introduced to keyword advertising. A few years later, I jumped to the agency side, working on ad serving and reporting, before joining a tech provider called Eyeblaster. We were focusing on high-level, rich media ads, and if I remember correctly, we were the first company to serve a commercial full HD video on the internet. 

Over the years, the product from Eyeblaster was renamed many times and turned into Sizmek — a full ad-serving solution that implemented tracking of mobile and in-app ads, and dynamic creative optimization or for short DCO. In 2019, the business was sold to Amazon, and I joined Amazon Ads within the ad-serving business as a Director of Sales Engineers. Within the last few years of Amazon Ads, I was mainly pitching Amazon Marketing Cloud, or for short AMC. And last year I joined Adverity, where I work as a Senior Solutions Consultant.

Amazon Marketing Cloud describes itself as a clean room, could you explain a bit about what that means and why it’s so important right now?

Clean room solutions came into the game after most of the major advertising platforms made the decision to no longer provide or accept third-party cookies. Within a data clean room, all data providers bring their data pseudonymized, so nobody knows where it’s from. It can’t be connected to an individual across the different providers. 

So this is what the clean room is about. It’s a central place where everyone in the room knows their own data set and signals, but no one in the room knows the identifying details from the others data providers. And this way they can communicate about these pseudonymized signals.

What are some of the key data points that people are looking at when they're using Amazon Marketing Cloud?

I think I need to take a step back to maybe explain more about the media side of the Amazon business, which has grown to become a massive player in the advertising game over the last couple of decades. 

Within Amazon Ads, there is an advertising business for search with sponsored product, sponsored brands, and also sponsored display. But there is also the DSP, which is the demand side platform to programmatically buy media on and also off Amazon. So on Amazon Marketplace, but also on Amazon-owned media like Freevee, for example, or Twitch. 

But there is even more. There’s also Prime advertising, there’s Alexa advertising, and so on and so on. So within Amazon, from every angle, there is some kind of advertising. And without AMC, you don't have a chance to connect all these data sources or signals into a single source of truth and make a holistic view or holistic measurement out of it.

And this is where AMC comes into play because with AMC acting as a clean room, all of these signals from the different advertising spaces within Amazon come together with pseudonymized signals. This then allows you to get a better holistic measurement across your advertising on Amazon by analyzing all these signals together in the same place.

You can then also have cross-channel attribution with custom look-back window, which you can’t do just using the Amazon application (search, DSP, etc.). And you can bring your own first-party data or even activate or enable other third-party data. So, AMC is not about a single metric or a single use case. Nearly every kind of marketing media you can analyze is supported in AMC as long as you enable it in AMC.

And is Amazon trying to encourage more of that third-party data to come into Amazon Marketing Cloud?

Yeah, I would say so. Facebook, Instagram, other social channels, and programmatic buys run through different DSPs, CRM data, and newsletter information — all this has valued information and can be used within AMC. The question is with the customer if the decision will be made to bring that kind of data into AMC.

If you’re an advertiser on Amazon and you want to analyze signals at this kind of granularity, then you have to use the AMC.

If you're not bringing these data sources into AMC, you will lose all the signals or their connection within your Amazon advertising because you can't get these signals into any other data clean room. There is no way around it.

 

If you're not bringing these data sources into AMC, you will lose all the signals or their connection within your Amazon advertising because you can't get these signals into any other data clean room. 

Marco Senftleben, Senior Solutions Consultant, Adverity

Does it still require some data manipulation from the user to join these third-party sources?

It definitely needs some manual manipulation, because you can import the data from Google campaign manager, for example, into AMC. But then there is no transformation that happens. So it's just a matter of using the right programming language in the AMC interface to pull the kind of data you're actually looking for. If you know how to code SQL, then it should be pretty easy for you as a programmer to combine these kinds of tables. And then you’ll be able to get the cost data from Google Campaign Manager, once you import data from all your channels you can merge these in AMC to an overall cost.

What does someone starting out on Amazon Marketing Cloud need to know about?

First of all, I would not advise anyone who is fresh to the Amazon advertising business to use AMC, because that's the end point where you can really get lost in the nitty gritty. There are so many metrics and dimensions within AMC that you will get lost because you don't know which metric and which dimension to combine, which to analyze, and so on and so on.

But if you are already educated, you know how to run your business, and how to advertise on Amazon and also on other channels, then I think the first thing is to work on your SQL knowledge. The second piece of advice from my point of view would be to get someone who has development knowledge to connect to the AMC API, to make it more programmatic.

Are there any undiscovered or undervalued metrics that people overlook when they're trying to understand their data in Amazon Marketing Cloud?

If you don’t use AMC, all of the data is undiscovered. Without AMC, every advertising channel within Amazon exists in a data silo. There’s no overlap, there’s no connection between them.

Therefore, it’s impossible to make a reach or frequency report across search and DSP. These are the details you can only get via AMC — making reach and frequency reports across all your Amazon activities. 

 

If you don’t use AMC, all of the data is undiscovered. Without AMC, every advertising channel within Amazon exists in a data silo.

Marco Senftleben, Senior Solutions Consultant, Adverity

And other signals help marketers to make audience insights in terms of timezone information, geoinformation, and media mix analysis. They help determine which source the sale is coming from, which kind of device, which site was involved - and also I mentioned already the custom attribution part. This isn’t available within any of the Amazon advertising direct channels, because there are only fixed windows and fixed settings. But with AMC, you can define your own lookback window and your own attribution settings.

Are there any common mistakes that you see users making with Amazon Marketing Cloud?

I wouldn't say that there are any common mistakes because Amazon themselves made AMC easy to use. On the one hand, I mentioned that there is a developer UI where you can write your own SQL. That's true for sure that if you have that knowledge, you can write your own code to create your report. 

On the other hand, there are so-called Instructional queries (pre-written SQL queries), where Amazon has already prewritten SQL code which you can easily take and activate in your AMC account to get to the results.

So if, for example, you have a running search campaign and running DSP campaign, you can easily use the template from the instructional queries to get your reach or frequency reports across both channels using a template. This makes it pretty easy for marketers without much coding knowledge to use AMC.

What do you expect to catch fire within marketing and data over the next 18 months?

What comes to my mind is a Gartner study that says roughly 60% of all CMOs will cut their departments by half by this year because they failed in their promises to improve on data. And I think, besides cookie deprecation, besides the looming recession and so on, to be against these headwinds we need to focus on our baseline. When it comes down to your data, analyzing and linking it together, and then visualizing it, this is the fundamental part of working with data to improve your campaigns’ ROI. I think that's the main topic for this year, and maybe also next year.

What's some advice that you would give someone that's just starting out in this industry now?

Yeah, it's a pretty hard question for me, to be honest with you, because I grew up with all this tech. Data issues where when I started in the industry were very different than they are today. They were very, very small. But I mentioned it already, in my experience, it's about being able to program in a language — and that doesn't have to be SQL in this case. But just to have the full package of having some kind of experience with programming language, I think that would be a real asset for someone starting in the data business.


If you enjoyed the Undiscovered Metric, you can check out the previous episode here.

Mitesh Lakhani, Director of Solutions Consulting at Adverity explains how to read Facebook metrics, and why Facebook is a great place to start when building a marketing strategy.

Make insights-driven decisions faster and easier!

book-demo
Book a demo