This US-based tech company is improving product search and discovery for retailers around the world
By Fiona Briggs

When shopping online, have you ever had a really bad product search experience? Maybe you were inundated by pages and pages of loosely relevant results — who has time to scroll through all that?! Or the popular item you searched for had a whopping zero results; how could that be? Did you have to rephrase your query so the search engine could get the gist?
Poor experiences like these aren’t just annoying to shoppers. They’re costly to retailers too, who leave money on the table when shoppers give up in frustration and take their wallets elsewhere. So, when a retailer has just what its customers want and need, it behoves the brand to make it easy for shoppers to find that (and to do it quickly).
Just ask Eli Finkelshteyn, co-founder and CEO of Constructor, which was founded to address these very issues. Constructor’s artificial intelligence (AI)-based platform is the “secret sauce” behind successful product search and discovery at some of the world’s biggest ecommerce brands: Very, Birkenstock, Sephora, Target Australia, home24, Bonobos and many major grocery chains, to name a few.
They’re using Constructor to drive highly personalised experiences across channels that not only improve the customer experience but also boost results. Sephora, for instance, saw a $40 million (USD) revenue lift through Constructor, and Target Australia posted a 14% lift in search purchase rates.
“Retailers today are seeking to connect with and serve their customers in a way that truly adds value for both parties,” Finkelshteyn said. “We help them serve up results and product recommendations that are both attractive to shoppers and benefit the bottom line: a win-win.”
How and why it works
Finkelshteyn explained that Constructor’s platform is constantly learning from customer behaviours onsite and optimising experiences — hundreds of millions of times per day. “We use AI and machine learning to learn from every search query on our customers’ sites, and from how shoppers interact with, and give feedback to, our platform’s recommendations and results,” he says.
So how does it all work? Constructor combines clickstream-based AI across a retailer’s digital properties (what did a given shopper click on? spend the most active time viewing? add to cart? purchase? and so on) with data from brief “Quizzes” that retailers give their shoppers, along with retailers’ other data streams, including loyalty program information. The result: in-the-moment personalisation in both online search results and product recommendations that appear across all of a retailer’s website pages.
Relevance is dead?
So, given the company’s technology for personalisation, you might, at first, be taken aback to hear Finkelshteyn declare that in ecommerce, relevance is broken and dead. But he’s not talking about the hyper-personalised experiences that Constructor delivers; he means what often passes for relevance in the industry… and the bar, unfortunately, can be low.
“If someone searches for shoes, and the retailer shows them shoes, is that really a win?” he mused. “Are these shoes popular in the customer’s geography? Do these shoes come in a size the customer can wear? Has the customer demonstrated interest in this style of shoe? You can show me high heels, and technically they’re relevant to the query, but I’m never going to buy them. Someone else might, but I won’t. What’s attractive to me isn’t what’s attractive to them. Retailers need to put themselves, figuratively, in their customers’ shoes: What would make them happy, while also serving the retailer’s goals?”
With Constructor, even a general search for footwear becomes more meaningful to the shopper. Its technology can rank and prioritise search results so they show items in the shopper’s preferred brands, styles, colours, price points and so on. And with natural language processing (NLP), Constructor incorporates typo tolerance and detects synonyms: understanding in the footwear example that a British shopper looking for “trainers” and an American one looking for “sneakers” want similar items shown. Constructor also optimises results and recommendations for each retailer’s unique key performance indicators (KPIs), such as revenue and high-margin products.

The AI in action
Headquartered in San Francisco, California, USA, Constructor has a presence worldwide, and over the last year, significantly increased its business in Europe.
The company recently signed UK-based digital retailer Very, which is implementing Constructor’s search, browse and autosuggest tools across its website and mobile app to give shoppers faster, more personalised experiences. With a large product catalogue — offering fashion, tech, homeware, toys, sports equipment and more from 2,000+ big-name brands — Very recognises the importance of helping each shopper quickly find what they want, so they aren’t overwhelmed and have a useful, cohesive experience.
“Finding the right products quickly is a vital part of the overall digital customer experience,” said Paul Hornby, digital customer experience director at The Very Group. “We’re excited to be taking product discovery to the next level by partnering with Constructor.”
As part of its multi-phased implementation, Very also plans to roll out Constructor’s personalised quiz functionality — where shoppers answer brief questions and, based on their responses, immediately get tailored experiences and recommendations.
Another European retailer, Germany-based home24, also recently drove impressive results from its Constructor deployment, with double-digit growth in search conversion. The home and living retailer offers hundreds of thousands of items but had faced constraints with its previous online search solution, which had limited fields for searchability and left shoppers unable to filter results by key attributes, such as material, colour and price.
Constructor addressed these areas and more, with its personalised search technology.
“Thanks to our work with Constructor, it’s even easier for shoppers to search for — and find — the best items for them,” said Gianluca Randisi, VP of product, home24. “Aside from the positive return on investment, we love the ease of use and hands-on support that come from working with Constructor.”
Eyes on the future
Even amidst tight economic times and uncertainty that many retailers are facing today, they’re doubling down on turning search results, product recommendations and superior customer experiences into revenue — and that means doubling down on Constructor.
Constructor itself has boasted a 100% client retention record over the last few years, underscoring the key and strategic role its technology plays, and doubled its revenue in FY22.
And as the industry continues to evolve, Finkelshteyn is committed to supporting clients and their evolving needs and trends. Nowadays, one of those trends is use cases for ChatGPT.
“ChatGPT is getting people excited about applications for AI in retail,” Finkelshteyn said. “We’ve been testing ways to incorporate our clickstream-based AI with ChatGPT’s AI, and we’re excited about the potential. It would allow shoppers to type in longer-form queries, even phrasing their needs in the form of an involved question — very different from the terse search terms we’ve come to expect. ChatGPT helps us understand what the customer is asking for, and our clickstream-based AI then helps us understand what products the retailer we’re working with carries that match the customer’s needs, and that the results we return to the customer are attractive to them and personalised to them.”
With ChatGPT and other emerging technologies, though, Finkelshteyn advocates for a measured and well-thought-out approach. “We’re still in the early innings of what’s possible with generative AI, and we’re excited about the experiments we’re planning in the space, and that some of the retailers we work with are willing to partner with us on those experiments,” he said. “We always ask ourselves as we create anything new, ‘Is what I’m doing truly useful and beneficial for the shopper and the retailer?’ This commitment — not to the ‘coolness factor’ but to utility — is at the heart of what Constructor does: using technology to truly create the best possible experiences for retailers and their customers.”