The Amazon Repricing Problem
Automation is the use of technology to optimize productivity in the delivery of services or products. If correctly executed, automation can offer a business a distinct competitive advantage but if poorly executed, it can damage a businesses ability to output a consistent level of service. In the case of Amazon repricing technology, precise execution is key.
Amazon repricing software automates a process that would be virtually impossible to execute with human labor – but are all Amazon repricing tools offering the same level of execution?
In 2003, I was invited by Amazon to be one of the first retailers to list sporting goods as a third-party seller on their marketplace platform. As our Amazon sales grew steadily to over several million dollars per year, our team used many third-party repricing tools to attempt to optimize and improve our sales. Naturally, we Googled the terms ‘Amazon Repricing Tools’ and ‘Amazon Repricing Software’ and tested all of the solutions available.
What our team noticed was that even repricing solutions provided by ‘leading companies’ charging significant fees did not always reprice our products correctly. Our Amazon repricing data frequently showed missing data, delayed repricing runs, missing/incorrect shipping prices, or in some cases, our products were even changed to the wrong prices.
Differences in Amazon Repricing Technology
Why aren’t all Amazon repricing tools able to provide accurate results?
The simple answer is that is the task of consistently capturing data from Amazon is not trivial and presents a significant challenge for automated software. Amazon repricing software can only be as accurate as the data it is able to capture and here lies the issue. Unfortunately, Amazon does not make the task of sampling pricing information, competitive merchant’s prices, and other inputs required to reprice accurately, an easy task. This has been the source of frustration for many third-party tools including those providing Amazon repricing services.
Amazon repricing services uses different techniques to capture pricing information and is a reason why performance can vary between providers. The techniques used to sample Amazon data include:
- Screen-Scraping or Web-Scraping Techniques
- Using variants of the Amazon APIs
- Using a combination of the above
Through extensive research, our development team learned techniques using Screen-Scraping have serious limitations and importantly is prohibited by Amazon’s strict terms of service. We were surprised to learn that some of the largest providers of third-party seller tools were relying on data from Screen-Scraping to capture prices. The inaccuracy issues that we had experienced as a seller and the inconsistencies between repricing services made complete sense after learning that Screen-Scraping is used by a large number of Amazon repricing services.
Screen-Scraping, like any automated process, experiences an error-rate which in a high volume retail scenario can have critical negative consequences. The Screen-Scraping process relies on the ability of the process to capture data in a fashion that mimics thousands of individual web browser requests. If for some reason a request fails and pricing data is not correctly captured, there is no way for the repricing tool to be able make any changes.
In recent months our research and development team has identified changes by Amazon that will render all Screen-Scraping techniques obsolete. If you’re a seller using a Screen-Scraping based Amazon repricing techniques you are almost certainly experiencing major issues since these changes have been put in place.
Amazon’s own API, called Amazon MWS, also has limitations. Due to the size of Amazon’s vast database which handles petabytes of data across the globe, services querying product data using the Amazon API also experience an error rate. What do these error rates mean for Amazon repricing services? In short, if there is no process or error-correction detection in place to handle data sampling errors they will not reprice products correctly. For high volume sellers, these errors and product miss-pricing adjustments can result in thousands of dollars of lost revenue.
Teikametrics’ Data-Driven Approach
Our goal at Teikametrics is to provide the best Amazon repricing service and our development team has put in thousands of hours to test, re-test, build, and release multiple iterations of our software. Through this data-driven process, clients using Teikametrics are able to experience an Amazon repricing service with a level of accuracy that is superior in terms of accuracy and performance to other solutions available.
Our Amazon repricing software dynamically identifies Amazon API inconsistencies, reacts, and adjusts to optimize the adjustment of all products. In addition, we’ve developed proprietary adjustment techniques to dynamically search for missing data to provide the most complete data set to execute price updates.
The resulting performance is why some of the largest retailers in the Amazon marketplace choose Teikametrics technology to accurately manage and reprice their products.