The Amazon Analytics and Data Launchpad


Despite Amazon’s frequent role as part of a multichannel commerce strategy, building a single source of truth around Amazon data and running relevant analytics remain fundamental challenges for brands.

We’ve put together this authoritative guide to cover the major topic areas around Amazon Analytics, ranging from general Amazon level-setting to types of Amazon data to Amazon ads.

Not every topic may apply to you, but you can head to what's most relevant:

  • The nature of Amazon data
  • Types of Amazon data (metrics and reports)
  • How to actually get data out of Amazon
  • Goals of analyzing Amazon data
  • Challenges of analyzing Amazon + other commerce data
  • Amazon Ads
  • Leveraging an Amazon analytics platform

Level set: The nature of Amazon data (and why it’s so hard to work with)

Amazon data presents two base challenges for merchants trying to make sense of it, both of which revolve around ownership.

1. They’re Amazon’s customers… not your customers

Customers who purchase your products on Amazon belong to Amazon. 

At least, according to Amazon. 

The privilege of selling on Amazon is that you’re able to leverage a potentially extraordinary customer acquisition channel. But, the struggle of selling on Amazon is that the only rules are Amazon’s rules, and you must make concessions. 

  • Merchants cannot track particular customers, nor can they route customers to their owned/DTC sites from any of their Amazon pages.  
  • Amazon anonymizes all customers. Customer emails are hashed, so you cannot segment Amazon customers with your DTC customers. Nor should you expect to be able to granularly track order patterns between your DTC site and Amazon.

2. There’s no such thing as an “Amazon customer,” at least when it comes to the data you can see

Amazon doesn’t make it simple to piece together a clear view of your business.

In general eCommerce, you can access and track four pillars of data: customers, products, orders, and fulfillments. The majority of your analytics efforts fall under or involve these categories, right?

For instance, in Shopify, BigCommerce, and other DTC platforms, your data is fairly complete: you can view and download your customer data (i.e., any PII), order data (what customers are buying), product data (e.g., SKUs, costs), and fulfillment data (all components of the supply chain).

However: Amazon, the sneaky bunch, only surfaces products and orders data–even though internally it collects and hoards immense amounts of customer and fulfillment data. 

After all, these are not your customers.

Amazon eliminates the idea of a customer from a data perspective so that merchants can’t “steal” customers from Amazon and only sell to them DTC.

What does this mean for your analytics?

In short, it makes Amazon analytics hard. In order to run useful analytics on Amazon data, you need to create the concept of customers via your orders and fulfillment data.

It’s even harder because although Amazon provides a lot of data, it isn’t the easiest to work with, and it’s even harder to automatically extract. We’re covering these points in the next couple sections.

Types of Amazon data

An extensive article could be written about all the Amazon data that you need to collect, track and analyze… Which is why we wrote one here: Amazon metrics and reports

Think of this section, though, as a TLDR of that article.

There are five major categories of data within Amazon analytics, each with relevant metrics and reports from Amazon itself:

  1. Amazon site performance
  2. Fees
  3. Customer experience
  4. Inventory
  5. Shipping

All these categories of data boil down to one purpose: 

You, as the merchant, will have all the data you need to help Amazon provide the best customer experience possible. 

Amazon will punish you financially if you don’t do this successfully. 

The most important Amazon metrics

Buy Box %

Just because you sell your products on Amazon doesn’t necessarily mean that, in practice, you will be selling your products on Amazon. 

You are competing against other brands who may be selling your products, which is why you need to “win the Buy Box:” that is, where customers purchase products and add to cart in the Amazon UI.

“Winning” the buy box means that Amazon shows you as the first seller of a particular product, which in practice means that the customer will purchase that product from you

Buy Box %, then, is how often in a particular time frame you are winning the Buy Box. It serves as an overall pulse-check to keep tabs on and analyze as long as you are selling on Amazon.

If the % stays high, it indicates that you are doing everything Amazon wants to keep customers happy.

Example of Winning the Buy Box

Consider the PDP below:

The Buy Box is highlighted in purple. Note that KOS (the original manufacturer) is selling the product, and they’re leveraging Amazon’s fulfillment network (i.e, the product is Fulfilled by Amazon, or FBA).

However, they have competition. 

In scrolling down the page, the shopper can see that other companies are selling this same protein powder:

In all likelihood, though, the shopper will opt to have the product fulfilled by KOS itself.

  • KOS won the buy box and appears at the top of the PDP.
  • KOS is selling the product at a lower price point than selling competitors.

Order Defect Rate (ODR)

To Amazon, an order “defect” is defined as anything besides an optimal customer experience. In practice, Amazon considers an order defective if:

  • A customer makes a chargeback
  • An A-to-z Guarantee claim is made
  • A customer writes a 1- or 2-star review

If a customer sees numerous 1- and 2-star reviews for your products, then they’re less likely to purchase. To Amazon, though, a high ODR signals that a brand should be suspended from selling, as it isn’t meeting their quality standards.

Amazon requires you to have an order defect rate of less than 1%. With that in mind, monitoring your ODR is vital for ensuring customers are happy and ensuring you’re in good standing with Amazon as a seller.

Order defect rate is calculated monthly so that customer claims can be processed and to give you a chance to rectify negative reviews.

If you hit the 1% benchmark, you have a brief window of opportunity to improve your ODR for future orders.

The easiest way to ensure your order defect rate is continuously below 1% is to make sure you provide outstanding customer service, ship on time, and deliver quality goods.

Unit Session Percentage

Unit session percentage is sort of Amazon’s version of conversion rate or view-to-buy rate.

Amazon defines unit session percentage as: "The number of units purchased relative to the number of customers who viewed the products. Calculated by dividing the number of units by the number of sessions for a selected time period, and then expressed as a percentage."

Unit session percentage shows how effective your PDPs are at driving unit sales. A high unit session percentage means you have products that are in high demand. A low percentage means you need to focus on getting better reviews, new product images (perhaps upscaling images for better quality), and exploring different pricing options.

While a lower unit session percentage doesn’t impact your standing with Amazon, it does play a role in Amazon’s algorithm, which means your products may rank lower when potential customers are searching for relevant keywords.

Getting data out of Amazon

Fundamental to the quest of running useful analytics on your Amazon data is to get data out of Amazon.

It’s impossible to uncover meaningful insights aside from the highest-level sales information via visualizations like this:

Instead, you will need to build visualizations and run analyses yourself. 

TLDR: There are two ways to get data out of Amazon—manually, by downloading spreadsheets, and programmatically.

That said, easily and reliably extracting your Amazon data is often a major pain point for brands—we’ve seen this as a fairly universal challenge, whether they’re doing a couple million or tens of millions of dollars on Amazon every year. 

To get data out of Amazon, you have two choices:

  • Downloading relevant reports from Amazon Seller Central or Amazon Vendor Central and combining those into spreadsheets 
  • Automating your Amazon data extraction via the SP-API, likely from an ELT pipeline

For smaller brands, it can be manageable yet annoying for an analyst or designated “Amazon data person” to build Amazon performance reports and presentations.

But for enterprise brands with larger and international data sets, it can be incredibly time consuming (sometimes 20+ hours/week) to simply organize Amazon data when taking the manual approach.

  • Spreadsheets of a particular report only cover regional data. A brand selling in 9 regions and downloading a particular sales report will have to download 9 sales reports. 
  • For certain reports, such as the Settlement Report, which is a brand’s profitability report, Amazon decides when to send these to you. This means that in addition to the analytical challenges the settlement report poses due to complexity, you don’t have the ability to zoom into individual marketplaces to get the granularity that you might hope for.

Without automated Amazon data extraction and reporting, larger brands will hit a ceiling in what they can feasibly report on, which means that at a certain point, they will end up flying blind in the Amazonian jungle. 

Goals of analyzing Amazon data

The only way to successfully grow your brand via Amazon is to keep Amazon happy, and the truth is, the goal of any Amazon analytics is to ensure that you do keep Amazon happy.

This means keeping on top of the Amazon-specific reports and data points, ensuring that customers are happy, and that you’re not in violation of any Amazon policies. 

Beyond that, though, in order to achieve practical business goals on Amazon, such as expanding into different international regions, you’ll need to:

  • Track your profitability by region
  • Track and report on your FBA inventory, also by region
  • Run optimized Amazon ads (which we’ll cover in a subsequent section)

Overall Profitability

Some brands make the mistake of exclusively tracking topline Amazon revenue rather than factoring in all their costs. 

Amazon’s Settlement Report is fundamental to accurate profitability reporting, and we highly recommend pulling all settlement reports into a BI/visualization tool either manually or programmatically to have a much easier time tracking costs: 

A couple examples of Settlement Report visualization/reporting from the Daasity app.

The Settlement Report includes an array of costs including:

  • Chargebacks
  • Warehouse holding costs
  • Shipping costs
  • Shipping taxes
  • Other taxes
  • Advertising costs
  • Promotions offered via Amazon

However, it’s necessary to layer in all your other typical product costs that factor into your margin, leaving you with your Amazon net profit. 

We highly recommend taking this a step further and breaking down your profitability by region:

FBA Inventory

Your operations/supply chain team will need weekly reporting around items dedicated to all Amazon regions that you sell in: 

Multiregion FBA inventory report via Daasity

Key to maintaining appropriate inventory levels (and ensuring your team avoids headaches and inventory panics) will be tracking overall quantities, by SKU, by region as well as sell-through rates.

Challenges of analyzing Amazon + other commerce data

Being able to analyze Amazon data alongside other commerce data is fundamental to continued success, growth, and profitability. 

Screenshot from the Daasity app showing Amazon data, transactional eCommerce data, and subscription eCommerce data

Especially as you scale, it will be imperative to have a crystal-clear understanding of, particularly, where and how your inventory will be distributed based on DTC and Amazon demands. 

However, it is easier said than done.

Joining eComm/DTC + Amazon data

Whether you’re selling via Salesforce Commerce Cloud, Shopify Plus, BigCommerce, or another platform, you ultimately face two options when you’re looking to combine data:

  1. Simplify DTC data
  2. Massage Amazon data to fit DTC data

Although neither option sounds great, option 2 is a better choice because nobody wants to simplify their eCommerce data and lose the granularity and actionability it brings. That said, you’re now left with some heavy data lifting to do. 

Even if you do pay for an Amazon-specialized BI/visualization tool, you might have nice-looking Amazon data, but you’re still left with two data silos that don’t overlap.The fundamental problem of Amazon data + eCommerce data remains.

In order to combine the two data sources, you have to create the two parts of the eCommerce data model that Amazon is missing (Customer and Fulfillment) out of your order data. 

Then, you need to map out the constructed Amazon data onto the eCommerce data using a number of the Amazon data reports using logic you create (which gets even more complicated, because not every report has all of the information you need, so you have to combine the data from different reports, which you have to download every time you want refreshed information) and combine those with your eCommerce data in separate spreadsheets. 

Amazon Ads

Amazon and eCommerce expert Dave O’Brien, who runs OTS Group Inc, provided an enormous amount of expertise and content for our comprehensive guide to Amazon advertising. We’ll summarize some key points in this section.

If you’re selling on Amazon, advertising on Amazon isn’t an option; it’s a must.

In doing so, you will not only bring in revenue from the ads themselves, but you will help your advertised products organically rank higher. Products with the best sales numbers, the highest conversion rates, the most reviews, and the highest ratings are the ones you’ll see at the top of Amazon’s Search Engine Results Pages (SERPs).

  • To ensure that your ads rank highly and convert well, you need to do serious research on your product and product category in order to find relevant, specific, and high-intent keywords. To facilitate this process, we recommend platforms such as Sellozo and Heilium10 for keyword research and ad budget optimization.

Your north star metric in the jungle (measuring advertising success)

On Amazon, TACoS should be your north star metric—and it is pronounced like the food.

TACoS stands for total ad cost of sale, and it measures how your ad spend impacts your total revenue, not only the revenue attributed to ad spend—this would be ACoS (ad cost of sale), which is the inverse of ROAS.

The reason we recommend TACoS for Amazon is because it includes your organic sales performance on Amazon, too. Given that successful advertising helps organic ranking on Amazon, it’s important to factor in everything to indicate your overall performance. 

To calculate your TACoS, divide your total ad spend by your total revenue, and multiply the result by 100:

Other Amazon Ads reporting and analysis

Beyond TACoS, however, we recommend tracking the typical advertising performance data, from ROAS to attributed ad revenue:

And, again, to gain a more complete view of your overall business performance, we recommend tracking .com advertising channels (Google, Facebook, Tiktok, etc.) and Amazon Ads:

The main reason being that if you’re tracking your multi-channel ad spend and resulting revenue, you can run tests to see the impact of DTC channels on Amazon and vice versa.

For example, the marketing team at shoe care brand Reshoevn8r found that it could turn Amazon PPC spending down and still enjoy an incremental lift in Amazon sales from their Facebook and Google budgets.

In short, they were able to track their marketing “halo effect.”

Amazon analytics purpose-built for consumer brands

Daasity is the only data and analytics platform purpose-built for consumer brands selling via eCommerce, Amazon, retail, and wholesale. 

Daasity was founded by data experts and analysts from enterprise consumer brands, and we deeply understand the challenges of gaining a holistic view of business performance by combining Amazon + other data. 

Daasity enables brands to automate their Amazon data extraction (even for the challenging data, such as the Settlement Report), centralize Amazon + other commerce data, and granularly analyze performance. 

To learn more about Daasity + Amazon, read on here.

May interest you