Extracting Data from Retailer POS Systems is Hard

Creating Value From the Data is even more Challenging

Consumer brands face an uphill battle when it comes to acquiring point-of-sale data from retailer portals, API’s, or syndicated feeds. The account management team lacks the engineering acumen required to create their own data pipelines, let alone have access to proper analytics tools. Fortunately, the SaaS community plugged this hole by standing up automated connectors between the two parties – retailers and brands. As a result of this trend, figuring out complex data extraction methods is no longer a competitive advantage for brands. The goalposts have shifted to creating value after the data lands in a database. 

This article will explore three ways analysts can reallocate their time away from data gathering to higher ROI analytical tasks.

      • AI: Modern BI tools enable the exploration of new modeling capabilities.
      • First principles thinking: Compared to reasoning by analogy, brands can break problems down to much deeper levels given the vast array of data points available. 
      • Scenario planning: The business world isn’t black and white. Analysts can use “what if” planning tools to help describe the different commercial scenarios under various input values.




The bar of entry into the advanced analytics world is lower than ever. Pick a business intelligence tool at random, and it will likely include advanced modeling features. You don’t need a Ph.D. in Mathematics to gain access to this world anymore. Since each brand will have a forecast analyst close to the data, the natural project to attack first is time series modeling over the P.O.S. data. Once a time series demand model is generated using the automated P.O.S. pulls from the SaaS partner, a number of capabilities are unlocked.  

      • Predicted weeks of stock (WOS): Instead of relying on a naive, backward-looking sales volume rate, brands can reorient the WOS calculation to a predicted future volume and spot out-of-stock items before they happen. Using predictive time series models, brands can stay one step ahead of seasonality curves and avoid excess/shortage inventory positions. 
      • Marketing lift: A secondary benefit of generating a predictive time series model lies in the marketing world. For example, the marketing team can gauge the incremental volume lift of a coupon by comparing the actuals against the “pre-coupon recognition” predictive model generated before the event. 


First principles thinking  


The explosion of data made available by retailers and the SaaS community unlocks the ability to use “first principles thinking” instead of reasoning by analogy. Let’s first define what these concepts mean. 

First principles thinking is a problem-solving approach that involves breaking down complex issues into fundamental truths or basic principles and then building up logical reasoning from there. Instead of relying on assumptions, analogies, or traditional thinking patterns, first principles thinking encourages individuals to question and analyze the core elements of a problem or situation. Examining the fundamental principles and reasoning from scratch allows for innovative and creative solutions that may not have been evident using conventional thinking methods. 

This approach to problem-solving fits the e-commerce world well, given the wide availability of data. Let’s take a simple question like, “Why is product XYZ sales down YoY on Amazon” and analyze this from a first principles (abbreviated) approach. To begin, the brand must lay out the various reasons a product could start to lose traction on Amazon.   


Data Location

Lost buy box ownership

Keepa API

Inventory OOS issues

Selling Partner API

Large change in ad spend


Huge increase in competitors YoY

Market share vendors


While the above table does not cover all the possible reasons for an item to decline in sales YoY, it does highlight the idea of breaking the business problem down into small, data-backed concepts. First principles thinking can provide an objective root cause analysis from which every brand could benefit. 


Scenario planning 


In e-commerce, a business can pull many input levers to increase sales. You can run a coupon using a percent discount or a consistent dollar amount. You can boost product visibility on the SERP (search engine result page) by investing in pay-per-click advertising. These levers have trade-offs that need to be considered using scenario planning. 

      • Coupons – Brands can look back on historical data as a guide to understand the relationship between conversion rate and the level of discount given. Use scenario planning to help the sales team find the “sweet spot” for the next coupon. 
      • Sponsored ads – There is a relationship between paid ad sales and organic sales. Brands can look back at their advertising data in relation to SERP penetration data to better understand how paid ads eventually create much more profitable, organic sales. Scenario planning can help the marketing team understand what the total value of paid ads is under different budget levels. 

With the rise of SaaS companies between retailers and brands, the focus has shifted to creating value from the data once it is in a database. This article explored three ways analysts can reallocate their time to higher ROI analytical work: leveraging AI and modern BI tools for advanced modeling, applying first principles thinking to break down complex issues into fundamental truths, and using scenario planning to evaluate the impact of different input values on sales and marketing strategies. Analysts can unlock valuable insights and drive strategic decision-making in the competitive retail landscape by evolving how consumer brands look at data.