Despite advances in technology and mass adoption of eCommerce over the last several years, the basic way of searching for products online hasn’t changed all that much — search results are often listed “in order of relevance” in neat rows on a webpage.
Eli Finkelshteyn, founder and CEO of Constructor, said one of the main issues is that engineers evaluate eCommerce search and discovery on relevance and speed whereas retailers evaluate success based on whether or not a sale was made.
Constructor uses artificial intelligence (AI) to provide search, autosuggest, recommendation capabilities to eCommerce merchants, using machine learning to improve results as more consumers use the tools. Importantly, Finkelshteyn said the AI software provides merchants with evidence for the recommendations and optimizations it makes in order to give retailers better insight — and prove that AI isn’t just a marketing gambit.
“When they’re making decisions, they can understand what they’re disagreeing with,” Finkelshteyn said. “If the AI is showing you its work, then you can say ‘OK, I get why you were doing that original thing, and I’ll disagree with it with knowledge,’ instead of just ‘I don’t know why it’s showing this thing, it looks crazy.’ Maybe there was a really good reason, but it just didn’t tell you.”
Constructor last week announced a $55 million Series A funding round, which Finkelshteyn said will be used to expand the platform and give smaller retailers a leg up. Though some investors wanted Constructor to expand outside of eCommerce to broaden its potential market, Finkelshteyn said he’s not trying to be an all-encompassing search product.
“We’re trying to build a big company, but we’re not trying to do it by selling a one-size-fits-all product to everybody,” Finkelshteyn said. “We’re very focused on our particular niche, which is just eCommerce. We want to make the very best product discovery platform for eCommerce.”
What it boils down to, Finkelshteyn said, is making the search and discovery process personal for both the retailer and its customer base using data in a similar way to how Netflix and Spotify create recommendations: with every search, the searches get better and the AI learns how different types of customers shop. Finkelshteyn said he wants to create as good a discovery experience online as exists at brick-and-mortar stores.
“Somebody’s going to figure that out, and as they figure it out, it’s going to give a much better experience for all retailers and for all users, and we’d like to be at the forefront of that,” he said.
Solving Search Abandonment
Constructor, of course, is not the only company to see room for improvement in the way eCommerce search and discovery is done. Earlier this year, Google Cloud launched a new Retail Search product, built on Google’s decades of search experience in an effort to reduce retailers’ estimated $300 billion search abandonment problem. Part of Google Cloud’s existing suite of tools for retailers, Google Cloud Retail Search allows retailers to enhance consumer experiences with personalized results and relevant promotions.
Read more: Google Cloud Retail Search Aims To Solve $300B Abandonment Problem
Finkelshteyn told PYMNTS he didn’t want to take shots at Google but noted that the company previously had a Site Search product that it shut down in 2017. “So, power to them for trying it again,” he said.
It’s up to retailers, however, to decide how much to trust that Google will stay with the Retail Search initiative, Finkelshteyn said. Other high-profile products, including Google Wave, Hire by Google and Google+, have been shut down by the tech giant in the past, though certain features of each have been integrated into products that are still supported.
Instead of focusing on beating the competition, Finkelshteyn said Constructor approaches the search abandonment problem from the perspective of consumer experience.
“We want to give them the most attractive products and we want to give it in the most attractive way that gives them the overall best experience, and have those products be personalized to them,” he said. “And that’s how we prefer to solve the problem. If somebody is abandoning, it’s probably because you didn’t give them as good an experience or a product as you could have.”