One of the top worries at startups has always been how to iterate on products. They used to have to rely exclusively on client feedback and instincts, but now they have a wealth of data to consider. This dynamic opens up new possibilities, but it also necessitates a new type of arbitrage. During a panel discussion with Jean-Denis Grèze of Plaid, Stephanie Mencarelli of InVision, and Pete Thompson of eBay at TechCrunch Disrupt 2021, we examined this new context. We looked at how to be data-driven, how to segment your user base, iteration speeds, and more.
Our discussion began with a fascinating question: Can a startup or a digital firm be too data-driven? The answer, according to Thompson, is “absolutely not” — but Mencarelli was not so sure.
Thompson:
It all comes down to how you use the data and how you balance it with other types of input. However, in my opinion, it is becoming increasingly vital to use data to find things within the company that cannot do any other way. It finds items that manual processing or human curation might miss.
Mencarelli:
I will somewhat differ and say that there is a point at which you may become too data-driven and miss what I refer to as the “edges of the innovation bell curve.” As a result, it is critical to monitor what smaller cohorts are doing, as they may be early adopters of new behavior.
“You need to be able to think about other things that you’re not doing or features that you could build that the data won’t actually tell you,” Thompson noted earlier. At eBay, he provided one example:
We have recently launched a feature called Image Search, which allows you to search for photos rather than text. In addition, this is a perfect illustration of how our analytics would never have told us that on our site, for example, for fashion, younger audiences simply want to be able to browse and see other items that seem like the same image, rather than relying on a text-based search. Moreover, you have to learn these things through various types of feedback loops.
