BY CHRIS TAYLOR, TRANSACT
We are starting to see more evidence of Amazon’s push to leverage generative AI in helping to drive an easier, more streamlined, and more effective shopping experience for their customers. AI-generated product review summaries is the latest example. Let’s take a close look at the reasoning behind it and the impact this might have on the consumer buying decision process and, ultimately, sales.
The primary purpose for providing review summaries is to make it easier and faster for shoppers to make purchasing decisions. In theory, If shoppers can get a general sense for how people have liked the product without having to comb through hundreds of reviews, it will enable them to decide and click the buy box more quickly and sales should ultimately increase. But is it really this simple?
AI-generated reviews don’t come without some potential problems, however. Is there important context lost in these summaries? If you read the example summary below, there are statements like “these headphones have received positive feedback from customers in terms of sound quality, performance, value”, but it doesn’t give any context around how many customers rated each benefit highly. Another example: “customers have praised the noise cancelling…”, but how many praised this benefit, 1? 50? 1,000? 25%? Depending on the parameters of the algorithm Amazon builds, these statements can mean different things and carry different weight to different people. Assuming Amazon does keep the individual reviews and the star ratings for the top benefits available, shoppers will still be able find this information, but what if they like the ease and speed of the summaries so much that they stop scrolling down for context? If this happens to a large degree, and I believe it will, buying decisions will be misguided to a meaningful degree.
Perhaps a much bigger concern lies around inaccuracy and inauthenticity in these reviews summaries for products that are loaded with “fake” reviews. Most of us industry insiders know inauthentic, “bought” reviews continue to be a growing problem, even on Amazon. If a product is loaded with fake reviews, the problem is exacerbated when the reviews are lumped into a summary that shoppers rely on in guiding their buying decision.
We believe in the premise that a feature that helps consumers in their buying decisions will provide a better customer experience and ultimately lead to greater sales, both generally good things, but not when it leads to misguided or misinformed decisions. If a reviews summary is to be implemented, we would recommend Amazon provide more relative data in the summaries to provide more context, for example: “92% of customers rated sound quality at 4.5 stars or above.” We also believe that this new AI-generated summary feature will make the problem of fake reviews even more harmful in misguiding customer buying decisions and that Amazon must do a better job of policing.
We advise larger brands who strive to protect brand integrity by refusing to buy fake reviews to put consistent pressure on Amazon to invest more in alleviating this problem that not only compromises customer experience in misguiding purchasing decisions but also makes the Amazon marketplace playing field completely unfair.
As Amazon and other top eRetailers continue to roll out new features and programs, we at TRANSACT help brands understand the impact to their business and how to action to capitalize.
Contributed by Chris Taylor, Vice President eCommerce Strategy at Transact Omnicom’s practice dedicated to providing connected commerce consulting and eRetail services.