AI-powered ecommerce recommendation engine Constructor nabs $55M

Constructor, a San Francisco, California-based ecommerce personalization startup, today announced that it raised $55 million in a series A round led by Silversmith Capital Partners. The funding, which brings the company’s total raised to $61.5 million, will be put toward product development, hiring, and go-to-market efforts, according to CEO Eli Finkelshteyn.

Online commerce conversions are well behind in-store — the average online shop sees less than 3% in conversions. But even though $4.2 trillion was spent on ecommerce platforms in 2020 alone, few ecommerce retailers have invested in a digital merchandising strategy. It’s estimated that $35.3 billion is lost every year by U.S. businesses in customer churn caused by avoidable customer experience issues, such as fair treatment and responsive service.

Constructor’s AI-powered software aims to address the challenges with discovery tools including search, autosuggest, browse, recommendations, and collections. Each of the company’s tools collects data, learning from queries and users to optimize for customers hundreds of millions of times every day.

Finkelshteyn founded Constructor in 2015 with former Shutterstock CTO Dan McCormick. An ex-data scientist at Shutterstock, Finkelshteyn believed that most things about online discovery were rooted in assumptions engineers made in the ’90s and that they missed opportunities in the realms of personalization, AI, and user interface — inspiring him to create Constructor.

“McCormick and I started geeking out together on the topic of search, and soon realized that it was an area that not only could be better from a user experience standpoint, but also had tons of untapped revenue for Shutterstock, which we proved by A/B testing various improvements,” Finkelshteyn told VentureBeat via email. “We went our separate ways for a few years, but never forgot the possibilities in search, and when we found ourselves free at the same time, we started Constructor to see if we could help other ecommerce companies better compete with Amazon through improved product discovery.”

AI-driven personalization

Constructor uses clickstream data — the links users follow on a website — to find rankings for products across search, browse, recommendations, and more. Clicks, additions to carts, and purchases inform which products the platform recommends to users. On the backend, Constructor shows how algorithms are ranking products and offers control over what users see for any query.

Finkelshteyn notes that the set of possible behaviors with ecommerce are fairly predictable. Whereas it can be difficult to figure out what a “successful” search or browse experience looks like between disparate sites, ecommerce sites are helpfully uniform. Users can search or browse to see a list of products, add them to cart, or “wishlist” them before checking out, providing a rich dataset to plug into a machine learning algorithm.

“Focusing on ecommerce and building our system from the ground-up to do so meant it took us longer to get to market than our competitors,” Finkelshteyn said. “This also means  that our customer base will never be as wide, but that for our target audience — ecommerce companies — we do a much, much better job than every other product discovery company, and we can prove it with A/B tests.”

On the explainability side, Constructor shows how and why its AI arrives at the results it shows, highlighting the factors that led to its decisions in employee dashboards. This provides a way for users who disagree with the default algorithm to push a promotion or change a result with an explanation of what they’re disagreeing with, according to Finkelshteyn.

“At this point, people are tired of seeing every company under the sun claiming to do AI, especially because most of them are not actually doing any AI at all,” Finkelshteyn said. “Constructor differentiates itself from its competitors because it was built from the ground up as a machine learning solution that harnesses and learns from the unique clickstream available within ecommerce to provide a provably better experience that drives users to find and purchase more of what they want.”

Finkelshteyn sees Constructor competing mostly with older installs built in-house as well as startups like Vue.ai, Syte, and Intellimize. The company’s customers include Backcountry, Bonobos, Sephora, American Eagle, Target, and Kmart Australia, which have driven recurring revenue to grow 233% year-over-year growth and revenue to double over the past year.

“[The health crisis] helped many companies realize how critical having the best ecommerce experience is to their business, but it did not fundamentally change the course of the business,” Finkelshteyn said. “More commerce has moved online every year since the invention of the internet, and even more commerce takes on a hybrid model that melds traditional brick-and-mortar commerce with digitization. Constructor sees this process continuing until all commerce has a digital component, and its mission is to help retailers outcompete Amazon in that new digital world while making shopping a delight for its users.”

Constructor expects to increase the size of its workforce from 60 employees to 80 by the end of the year.