In this special guest feature, Eyal Lanxner, CTO and Co-Founder at Feedvisor, discusses the motivations behind algorithmic repricers based on AI and machine learning to find the optimal strategy for securing Amazon’s Buy Box, which is an essential means for increasing exposure and sales in this marketplace. Prior to Feedvisor, Eyal managed the research and analytics groups at VeriSign and was CTO at Zoomix Data Mastering (acquired by Microsoft). Eyal holds a B.Sc. and M.Sc. in computer science from Bar-Ilan University, Israel.
Amazon’s third-party marketplace has grown dramatically since its inception, starting with five percent of product units sold by third-party sellers in 2001 and increasing to over 40 percent of products units sold today. Now, more than two million third-party sellers compete for the attention and wallets of Amazon consumers. To cope with this increasing competition, many sellers turn to automatic repricing tools for assistance, the best of which employ artificial intelligence and machine learning algorithms. Such technologies probably provide the best chance for Amazon sellers to succeed on this business platform.
One of the goals of advanced technologies is to find the optimal strategy for securing Amazon’s Buy Box, which is an essential means for increasing exposure and sales in this marketplace. The Buy Box is located on the most visible and accessible location on Amazon’s product pages, and represents Amazon’s recommended seller out of all those selling that product. The seller that appears in the Buy Box is determined by Amazon’s proprietary ranking algorithm, which considers multiple factors associated with the sellers, including fulfillment method, shipping time, seller rating and, of course, price. The weights of these and other factors can change from product to product, and from time to time. The algorithm can actively change the seller in the Buy Box in a matter of minutes, adjusting itself to the changing competition and pricing in the arena. This is because of the increasing number of price changes occurring in the marketplace, including Amazon itself generating over 2.5 million price adjustments each day so as to remain competitive with other marketplaces and its own third-party sellers.
Here’s why winning the Buy Box is so important: It’s directly correlated to the levels of revenue, as the Buy Box is believed to secure over 80 percent of product page conversions. Hence sellers are required to continuously monitor the competitive landscape around their products, identify price changes within the competition and adjust their prices accordingly. However, because of the rapid price changes in the arena, it’s nearly impossible for sellers to manually monitor the entire competitive landscape across their product lines. This is where repricing tools come into play.
Algorithmic vs Rule-Based Repricing
Repricing technologies can be roughly broken down to two main categories: algorithmic repricers and rule-based repricers. The general objectives of both are winning the Buy Box, increasing sales volume and improving profit margins.
While algorithmic repricers exploit data and science to fulfill these objectives, rule-based repricers often fall short. Algorithmic repricers will usually assess the entire competitive landscape over time, consider the many factors comprising it and adjust themselves to changing market conditions. Furthermore, such algorithms will typically strive to predict how the arena will evolve in the future and make a more informed decision based on this prediction. By contrast, rule-based repricers tend to look at a fraction of the data available, employ simplistic business logic and, most importantly, don’t learn and adjust themselves over time.
The Benefits of Algorithmic Repricing
Research shows that sellers using algorithmic repricing win the Buy Box more frequently than sellers who price manually. A study found that only up to 10 percent of Amazon sellers use algorithmic repricers, and those that do account for approximately one third of the best-selling products. The same study found that 60 percent of sellers using algorithmic repricers on Amazon publish prices that are higher than the lowest price for a given product.
Algorithmic repricers free sellers from analyzing data by themselves and determining the business logic to apply over each of their products, effectively eliminating human bias and error. This doesn’t only optimize business performance in the short term, but can also free up time for handling more strategic activities, such as product sourcing, customer service and engagement in additional categories or channels. Only by shifting focus to these business-critical activities can online sellers maintain consistent growth over time.
As Amazon continues to grow, running a business on this platform is only getting more nuanced and complex, requiring the assistance of advanced technology. As of today, we see more and more online sellers drawing this conclusion, realizing it as a major key for success and positioning it in their higher priorities. These are the sellers who probably have the best chance to make it through this hyper-competitive arena.
Sign up for the free insideAI News newsletter.
