.One of the best reports Amazon provides Sellers for understanding their business performance comes in the ‘Detail Page Sales and Traffic by Child Item’ report.

It is not without its limitations, namely that you cannot view your ASIN level data in time series format, but considering it includes fields such as Ordered Product Sales (OPS), Conversion Rates (Unit Session Percentage and Order Item Session Percentage), Buy Box Percentage and Sessions… there is no better report for diagnosing what is driving changes in your business performance. 

In fact, this is the report we use in our ‘Bridge Analysis’ to calculate the ‘drivers of change’, which help us zero-in on growth opportunities for Brands and separate the ‘signal’ from the ‘noise’.

But one thing I could never reconcile was how Amazon was calculating OPS for this report.  If you  have ever tried to compare the total OPS for the ‘by ASIN’ Business Reports against the ‘by Date’ Business Reports, you may have seen the issue yourself.  By Amazon’s definition, Ordered Product Sales is “calculated by multiplying the price of order items and the number of units sold for the selected time period.” But if you compare the ‘OPS per Unit’ against your Amazon listing price, more often than not you’ll see that those two don’t tie either (for example, OPS per Unit equals $25.55 but your Amazon listing price is $24.99, what gives?).

Having spent years working with Amazon data, their documentation, and the various APIs, I have come to expect a certain amount of squishiness and lack of precision from them as they attempt to accommodate for rapidly changing and evolving data points and definitions.  When things don’t make sense, or the documentation seems out of date, my conciliatory refrain is “It’s Amazon.”

These last few months we have been integrating the Marketplace Web Services (MWS) API data for Seller Central accounts into our DATA STUDIO solution, and through the tedious process of data validation… I have finally gotten to the bottom of the question, “How do I reconcile Amazon Ordered Product Sales across the Business Reports”. 
​
First let’s level set:
​

You can find the ‘Detail Page Sales and Traffic by Child Item’ report by clicking “Reports” and then “Business Reports” from the main navigation bar at the top of the Seller Central dashboard. The report “Detail Page Sales and Traffic by Child Item” can then be found under the ‘By ASIN’ section of the left navigation menu of the Business Reports module.

Within this report, you can select from a suite of metrics including B2B specific (not shown here) and Amazon will deliver reporting metrics for your ASINs that had performance data over the time period selected. 

As OPS = SUM(Item Price * Units Ordered), if you take the ‘Ordered Product Sales’ for an ASIN and divide it by the ‘Units Ordered’, you’ll get OPS per Unit, which by definition should equate to ‘Item Price’.  But checking ‘Item Price’ against your actual Amazon listing price, you’ll see that OPS per Unit (or ‘Item Price’) can be overstated, usually in the range of ~2-4 percent. Per the example mentioned above, your OPS per Unit will be something like $25.55 while your Amazon listing price is $24.99, a difference of $0.56 or a 2.2% difference.

So, either we have a data disconnect between Amazon’s reporting of ‘Units Ordered’ and ‘Ordered Product Sales’, or Amazon has incorrectly defined or labeled the field ‘Item Price’. Teaser, it’s the latter; Amazon has incorrectly defined ‘Item Price’ for this report. 
​

The best way to get a ‘true’ accounting of Ordered Product Sales, in total or by ASIN, is by leveraging MWS API data. It is also possible to perform a similar analysis by exporting the raw orders data out of Seller Central, but not one I’d recommend as you have to manually pull 30 days’ worth of data at a time...we’re all for streamlining that process, which is why we are working on providing MWS data via a database for our customers, no manual report pulling required.

For some time now we have been delivering streamlined access to Amazon Advertising Data via stand-alone databases.  Leveraging the Advertising API data has brought many benefits including time saved pulling manual reports, easy connection to most reporting platforms, and scaling Amazon reporting for internal teams and external clients. Now that we have begun incorporating MWS data, we are seeing similar benefits… and new possibilities.

Among many other data elements, MWS delivers Amazon order information at an ASIN and day level.  Each ASIN ordered is tied to an Amazon Order ID, with all of the customer fees broken out. Included in those fees are ‘Item Price’, which indeed matches the Amazon list price, but also included are: ‘Item Tax’, ‘Shipping Price’, ‘Shipping Tax’, ‘Gift Wrap Price’, ‘Gift Wrap Tax’, ‘Item Promotion Discount’, and ‘Ship Promotion Discount’.

As part of our data validation we mapped MWS data to the Business Reports to ensure we could replicate and automate these reports going forward.  Through that process we discovered that ‘Item Price’ in the ‘Detail Page Sales and Traffic by Child Item’ report is actually calculated as ‘Item Price’ + ’Item Tax’, whereas the ‘Detail Page Sales and Traffic’ report calculates OPS using only ‘Item Price’. The ‘why’ on this I’ll explain as ‘It’s Amazon’…  but it is probably a legacy definition from before they were collecting sales tax.
​
If you’ve ever been called upon to explain this discrepancy, you now have the most important disconnect. (I’ll save the other edge cases for another time.) But beyond this one use-case, there are other practical applications for tracking your MWS orders data in conjunction with your Business Reports. First and foremost, Promotional discounts are not being netted out of your OPS in the ‘Detail Page Sales and Traffic’ Business Reports. This can lead to surprises in your Amazon payments, as well as distort your conversion rates as price moves and price elasticity aren’t fully transparent in the business reports.  
​

​The Business Reports are still a valuable resource for your business and contain data currently unavailable from any other source; namely Sessions, Buy Box Percentage, and Conversion Rates. These reports do however have their limitations, and supplementing them with MWS data can help enrich your understanding of your business.   

Beyond what has been noted above, with MWS data you can calculate the Lifetime Value of your customers across both Amazon fulfilled and Merchant fulfilled orders.  This is not possible through the manual reports.  You can also measure customer repurchase cycles and gauge how well you are selling across your line.  You open the door to Geo-mapping your sales, which can inform your DSP targeting (Amazon’s programmatic advertising platform or Demand Side Platform) and help in evaluating other off-platform initiatives.  And I’d be remiss if I didn’t mention, with MWS data you can efficiently calculate ‘Total ACoS’ or ‘TACoS’, and more importantly, through all the above, establish your ACoS threshold for customer acquisition initiatives.
​
All in, there is much to be gained by taking control of your MWS data, well beyond answering the irksome question of data consistency across reports. 

​A lot of people want to know, “How does Amazon work?”. Warm up the Wayback Machine, we are going back to the early days of Amazon to trace the lineage of what I call ‘The Amazon Game’.
​

• Amazon has evolved, but the origins of the platform are relevant today
• Advertising plays a different role on Amazon than it does on other search platforms
• Amazon PPC Advertising is not optional so plan for it in your financial models

​

The origins of Amazon Search began in the early 2000’s as Amazon expanded far beyond books and it became imperative to improve “how to serve customers with the most relevant products”. At first read you may think, “Just like Google, right?”, but not so fast. 

Google search was designed to facilitate customer research while Amazon was designed to drive a sale. This may seem like a distinction without a difference, and maybe in the early days it was, but as the platform has evolved it’s become very important indeed.

As Amazon ‘learns’ what products are purchased for specific customer queries, it increases the product’s organic search relevancy and ranking for that query.  And while the algorithm’s inputs and relative weighting are a well-kept secret, a Brand that focuses on the factors that support conversions will win in the end.
​
The 4P’s of Marketing (Product, Price, Place, Promotion) are a good place to start, but they need to be translated to the idiosyncrasies of Amazon: 

• Product = Product Reviews, Seller Feedback, Images, Infographics, Titles, Bullets, Enhanced Brand Content
• Price = Buy Box, Off-Platform Price Competitiveness, Free Shipping
• Place = Search Ranking, Relevant Meta-Data and Proper Indexing, Prime Eligibility
• Promotion = Sample Packs, Bundles, Coupons, Deal Placements… and ADVERTISING?

​

Fast forward to 2012 when Amazon enters the Pay-Per-Click Advertising game. 

The discussion for launching Amazon Advertising, specifically Pay-Per-Click (PPC) Advertising, provided a great philosophical discussion internally. 

On one hand, the ‘Sanctity of Search’ and ‘Preserving Customer Trust’ were foundational principles for Amazon.  But on the other, there were valid use-cases for which advertising would improve the customer experience.

One example arises with the question of what to do when a new model of a top-selling product launches for the upcoming season. Do you surface the old top-selling television, or its new and improved replacement model? It is in the ‘Customer’s best interest’ to be served the newest model efficiently and advertising placements were a valid means for doing so.
​
And while it did lead to some bad behavior by Vendors, namely forgoing Co-op funding or cost reductions on unprofitable products to fund their advertising efforts, these were eventually addressed over time.  But regardless, Amazon had opened Pandora’s box, added a powerful new weapon to a Merchant’s arsenal in the process, and increased the complexity of how Amazon works.
​

With the launch of Amazon PPC Advertising, Amazon changed the game and the distinction from Google Search becomes clear.

With Google there is a separation between Church and State; advertising conversions do not influence organic search results.  For Amazon, if your ads convert, not only will Amazon serve your ad more often for that targeted query, they will also boost the ranking of your organic placement for that same query. 
​
This new reality fundamentally changed how the Amazon Game is played and made advertising a critical component of a brand’s success.  And not only did it add complexity to an already convoluted platform, but also ushered in the era of ‘Operational Marketing’… which we will get to in a separate post.

If you are on Amazon, Advertising is not optional. Even if Amazon is not a priority for your business, a brand’s terms need to be defended to avoid losing customers to your competitors… fortunately, this can generally be done cost effectively.  And if you are looking to grow your Amazon business, a sophisticated (objective-based, operationally-minded) customer acquisition strategy is imperative for success; the marketplace is an arms race and you cannot leave your best weapon on the sidelines.  

It has become a universal truth that Amazon has turned the world of Retail on its head. Having spent the past decade working for and with Amazon, I have had a front row seat to the disruption. And despite its importance as both a retail and advertising channel, it is still often misunderstood.
​
It is easy to be distracted by the constant stream of policy changes, the launch of new ad units, or shiny ‘Beta’ launches, but it is the fundamentals of Amazon that determine whether or not your brand succeeds or fails… both on the platform and off. And as true as that statement was in the early days, it has only become more so as Amazon has grown in importance.
So, to get us back to the basics, let’s start with a simple question, “What exactly is Amazon?”
​

• If you can not win the Buy Box, it all falls apart
• Optimize your detail pages so customers can find you in Amazon’s Search Results
• Use your detail page content to educate and motivate customers to convert
• Amazon affects sales in other channels, and other channels affect sales on Amazon 

​

​Generally speaking, anyone with access to your product can list it for sale on Amazon.  Amazon allows these sellers, authorized or not, to compete for the ‘Buy Box’, the mechanism through which the vast majority of sales are routed (as opposed to ‘More Buying Choices’.) 

​The primary determinations for the Buy Box winner are price and inventory availability. This is great for customers, who value selection, price, and convenience, and serves as the foundation of the famous Amazon Flywheel. For Brands however, it can mean price degradation, inconsistent signals for Forecasting and Inventory planning, and an inability to promote your products through advertising.
In short, if you cannot consistently win the Buy Box, you will not succeed on the platform.

​With millions of products available for sale, Amazon serves as a Search Engine to deliver customers with the most relevant products based on the query they have entered. In order to do this successfully, Amazon uses the meta-data and detail page content to ‘understand’ what the product is and how best to surface it in search results. 

You may have heard the term ‘Detail Page Optimization’, the purpose of which is to ‘optimize’ your pages for Amazon’s Search algorithms. The trick here is to balance content written for algorithms, with content written for customers… my next point.
​

Once a customer lands on your detail page, it comes down to the quality of the content to determine whether or not you achieve a sale. I like to think of this in old-school retail terms, namely the challenge of the ‘Unassisted Sales Floor’. But instead of retail displays, ‘Point-of-Purchase’ (POP) materials, and informative packaging, it is images, videos, infographics, titles, bullets, ‘Enhanced Brand Content’, and reviews that do the selling.

There are plenty of best practices to find out there, but my best recommendation is to shop your pages like a customer. Do you give them all the information they need to make a purchase decision? And just as importantly, if they determine that this product is not for them, are you empowering them to shop across your line to keep them within your Brand?
​

This final point speaks to the importance of Amazon as more than just a sales channel.  Amazon acts as a multiplier of all the good and bad of your business. Have strong reviews of your product on Amazon? It will help your Brick & Mortar conversions, and may even help you win floor spots in the first place. 

But the inverse is also true. Gone are the days of dumping overstock in some far-away region, or leaving it to unchecked distributors to get your products out in the market. Any missteps your business makes has the potential to disrupt, if not destroy, your brand. 
Amazon has a famous saying: “Start with the customer and work backward”, which is a great mental exercise regardless of your business. But for Brands today, I would also consider: “Start with Amazon and work backwards.” You will be happy you did. The health of your Brand depends on it.
​

Customers are on Amazon, and whether you like it or not, your products likely are as well. Rather than cede control to some third-party without your same standards, best to tackle the channel head on and establish control over how Amazon represents you to the world. And if it becomes a great sales channel for your Brand as well, bonus!

The three questions we get asked by agencies, brands and/or retailers the most often are:

1. how can we more easily access our data? 
2. is there a simple way for us to identify what we need to focus on to improve our ad performance?
3. how do we know how much to spend on inventory? 

​Getting your data out of Amazon and maintaining it is very difficult, and Amazon isn’t in a hurry to help you.  For that reason, we have built DATA STUDIO.  Firstly, it is a simple tool that helps you connect to your account(s).  Once connected, your data is securely stored so you can review it using our views.  Secondly, because we are maintaining and backing up your data in your own dedicated data warehouse, we provide easy access to it so you can link it to your own BI tool for custom reporting.

Amazon doesn’t tell you where you could better allocate your advertising spend based on the success or failures of your current campaigns and keywords.  Instead, you need to have some type of analytics tool to highlight where you should focus your attention so you can take immediately action.  Our ADVERTISING STUDIO is setup for you to review, drill-down and action your amazon campaigns and keywords to improve your advertising performance.

Everyone listed on Amazon knows that not having inventory can impact your search placement.  So, it’s important to manage your cash flow so you don’t run out.  Our INVENTORY PLANNING software walks you through the process of managing your open-to-buy spend so you can better allocate your ordering budget based on inventory goals, sales targets, and planned receipts.  After all, you don’t want a lack of inventory to negate all the work you’ve done to optimize your advertising performance. 

Want to learn more...check out our video.

Being a planner or buyer is challenging and knowing what state your inventory is in for any item at any given time is crucial to your success.  If you over order, you may run into excess and obsolete issues.  If you do not order enough, you may not be able to fulfill as many customer orders as you need to.  In other words, the challenge for planners and buyers is to balance your ordering which you can only do if you know and understand your inventory state and position for every SKU.

There is very little room for a planner to make mistakes. A simple breakdown of inventory into one of the three states is an easy way for a planner to identify imbalance so they can take corrective action.  Every stocked item can be positioned as either overstocked, understocked, or balanced.  In an ideal world, 60-80% of you inventory is in a balanced state.  The remainder of which would be classified as overstocked or understocked.  Inventory in these two states should drive your workflow by calling out exceptions so you can take corrective action.

Your traditional inventory "saw-tooth" diagram best illustrates the three inventory states, as seen above.  And, there are two key inventory parameters that define the saw-tooth diagram - safety stock and cycle stock.  The simplest inventory state to identify and the one that draws the most attention is the understocked position.  It is when an item is below the safety stock level or has no inventory, and is therefore in a (potential) stock out position.  Next, an item where the stock on hand is greater than the maximum inventory level (where maximum inventory = safety stock + [2 x cycle stock]) is considered to be in an overstocked position.  Lastly, an item where the on hand value falls within the cycle stock range is considered to be in balance.

Knowing these three states and being able to report your inventory in one of these three states is the first step towards building an actionable, exception-based workflow. There are varying degrees of urgency associated with each state depending upon the circumstance and tolerance.  Understocked items are the most important to action which requires you to place an order and, possibly, expedite it. Balanced items fall under typical ordering schedules where you are buying under normal circumstances taking into consideration forecast accuracy, lead times, PO minimums, containers, etc.. Overstocked items can drive exceptions requiring you to cancel or defer orders, and, in severe cases, throw promotions.

Most planners have too many items to manage. You want to, therefore, work and manage the exceptions.  Learning and applying these inventory states to drive your exceptions and workflow will not only help you reduce workload, but also focus your attention on items that need handling and review.  After all, there are only 18 action exceptions that can be derived from categorizing your inventory into these three states...can you figure them out?

There are numerous ways to calculate safety and/or cycle stock.  When optimizing your safety stock based on service level objectives, not only is the forecast and forecast error considered, but so should the lead time and the lead-time error.  More simply put, the longer the lead time and the likelihood your supplier won't perform to that lead time, the more safety stock you should carry to maintain your service, and vice versa. 

When placing orders, if you are trying to minimize your total annual costs by using the economic order quantity (EOQ) to define your cycle stock (:= 1/2 EOQ), the EOQ is directly proportional to the forecast, which as already discussed, could be unreliable as the forecast selection is dependent upon the projected forecast error over lead time.  ​