Eli Finkelshteyn, Founder and CEO at Constructor, shares more on Constructor’s recent funding and the latest in GenAI developments in this chat with MarTechSeries:

_______

Hi Eli — tell us about yourself and more about what inspired Constructor.

Happy to, and thanks for having me. My background is in AI and computational linguistics; I used to work as a search engineer and data scientist at Shutterstock, back when it was a startup. As I tackled the problem of ecommerce search there, I got excited by how even the small changes we implemented really impacted the revenue the company made.

At the same time, I recognized that I was part of a small team, and the impact we could make with so few people had its limits. So, even as we celebrated our accomplishments, I was well aware that compared to Amazon, we had five people working on the problem and they had close to 1,000. Essentially we were building a kayak when they already had a battleship.

With that in mind, my cofounder Dan and I founded Constructor so that we could create the best commerce search engine available anywhere, and so that other ecommerce companies could deliver superior search and product discovery experiences — ones that rivaled what the biggest companies in commerce were building in-house. The goal was to build a flexible platform for search and product discovery — to do it once and do it well — so that other ecommerce businesses would no longer need to build from scratch, and so the technological advantages the biggest companies in commerce had would be a thing of the past.

I’m incredibly proud of what we’ve accomplished to date: helping ecommerce companies around the world better cater to their shoppers, measurably improve the customer experience, and meet their business goals. Our technology is incredibly pervasive: In the last six months alone, Constructor powered more than 100 billion shopper interactions, helping retailers, brands, manufacturers, and distributors deliver hundreds of millions of personalized experiences every day. We recently secured $25 million in Series B funding too, which we’ll use to continue to better serve our customers.

Take us through your recent funding round. What key funding journey learnings and takeaways would you share with our audience?

The most important thing I’ve learned about fundraising is that it’s far easier to get if you don’t need it. We were in a tough position five years ago when we were not yet in-market with our search engine and were dependent on venture capital. I remember spending months trying to fundraise. One of our investors at the time said, “If you can’t fundraise, you could always just be profitable.” He meant it as a joke, but once we started really thinking about it, we realized that we actually did have a path to profitability if our search engine worked as planned. It would require tightening our belts in the short term, but if we really put our minds to it, we could do it. And if we did, we’d have far more freedom to build the company we wanted to build.

Since that time, we’ve always either been profitable, or kept a line-of-sight to profitability, so we could switch back to it if needed. Our last two fundraising rounds were by far the easiest we’ve ever raised. Both came inbound and preemptively while we weren’t fundraising, and we didn’t actually need either one of them.

If I could go back in time and give myself advice, it would be to focus on building the company our customers want — not the company investors want. If you build a company and product customers really want, you may not even need investors, and investors will eventually notice and want to speak with you either way. Always put the customer first. You can have a business without a lot of things, but you can’t have a business without customers.

How are you seeing the latest AI and GenAI developments change the retail workflow and processes today? What key innovations have piqued your interest from around the world?

The thing I love about being in the ecommerce space today is that technological developments are constantly expanding what’s possible, and it feels like there’s something new every year. Of course, there’s the really shiny stuff like ChatGPT, but there are also tons of other important advancements happening simultaneously.

The things that really pique my interest are technologies where we can prove an ROI for our customers. Some of them, like our AI Shopping Assistant, are customer-facing and feel like magic compared to what was possible two to three years ago. But others, like our Attribute Enrichment functionality, work on the backend of ecommerce product data, generating new product attributes and fixing existing ones so that products are easier for shoppers to find. The vast majority of customers won’t even know it’s happening. But, it has led to a proven increase in purchases everywhere we’ve tested it, and it does lead to a better user experience, whether shoppers realize it’s happening or not.

Long story short, we want to focus not on what’s shiny or demos the best. We want to focus on what drives value for our customers and for their shoppers.

Marketing Technology News: MarTech Interview with Roman Gun, Vice President, Product @ Zeta Global

We’d like to know more about how leading brands like Sephora use AI to drive their ecommerce customer life cycles and journeys.

We help companies like Sephora, Petco, Under Armour and many more deliver personalization across the buyer journey: in search results, browse experiences, product recommendations, landing pages, product-finder quizzes and more. They use our platform to do that, while simultaneously optimizing experiences for ecommerce metrics that matter to them: such as revenue, conversions, profit, inventory balancing and so on.

Constructor pioneered, and continues to innovate on, clickstream-based AI. That means, on a given ecommerce site, we look at what their shoppers click on, add to cart, purchase, search for, spend the most active time viewing, etc. to determine the best items to show each individual.

Most other search engines match results just based on keywords typed. So, for example, someone who searched for “green apple” (wanting Granny Smiths) might also see green apple sports drinks and green apple-scented shampoo in their search results, but may not see Granny Smith apples because they aren’t tagged as “green.” But our AI accurately gauges shopper intent to power results and experiences that are attractive and hyper-personalized, and lead to more conversions. That’s why Sephora experienced a $40 million revenue lift when they first tried Constructor, and other customers have reported results including a 16.5% increase in average order value, a 247% increase in revenue per visit, a 92% increase in recommendations conversions, 20X ROI on dollars they invested in Constructor and more.

Our products use other forms of AI, too — including machine learning, large language models (LLMs), transformers, natural language processing (NLP) and generative AI — and we’re committed to exploring and implementing AI in innovative ways that drive value. Some of our recent GenAI-powered solutions include Attribute Enrichment, which I mentioned earlier, addresses a common and costly pain point for ecommerce merchandisers: missing or mislabeled product catalog data. When shoppers can’t find what they’re looking for due to bad data, companies often miss out on a sale. Attribute Enrichment, though, automatically generates and enhances product data to fill in the gaps and expose shoppers to a broader range of products.

We’re especially excited about our AI Shopping Assistant (ASA), too — a conversational product discovery tool that blends generative AI with our personalization technology, helping brands keep shoppers on-site for their research. Unlike typical search technology, ASA supports long-form queries like, “I’m going camping with my kids in the mountains for the first time. What do we need?” Results make sense contextually, are personalized to the shopper, and reflect real-time inventory.

What types of opportunities would you say companies are losing out on if they aren’t deploying AI today?

Given what we do, my answer may surprise you, but I don’t think businesses should feel compelled to use AI — and certainly not if they’re only considering it because it’s the trendy thing to do or to “keep up with the Joneses.” AI can bring a lot of benefits, but you want to harness it for the right use cases and in the right way, such as to improve the customer experience. In short, don’t use it just because you can or because it looks cool. Use it only if you can make it actually useful to your customers.

So, when considering an AI deployment, businesses should look at their goals and what they’re trying to do. Is it best accomplished with generative AI? A different form of AI? Or some other technology entirely?

On the vendor side, we’re in an AI bubble and “hype cycle” right now. Too many tech vendors are bolting AI onto their products in an effort to proclaim “Us too!” But, again, if you’re just putting AI in your product for the sake of AI, you’re likely going to create something no one ultimately wants.

So, when developing or implementing AI products, I recommend two things:

• Prioritizing value. The solution doesn’t need to look flashy or even be perfect. It has to be useful, though: helping better serve end-users, driving internal efficiencies, etc. — something where ideally the value is measurable.
• Looking toward the future. The AI-based products that will survive don’t have to be immediately adopted by all people. They just need to be adopted by some people who love them and are willing to tell their friends. It’s much better to make something that’s beloved by a small but growing segment of customers, versus something that’s tried by all your customers but beloved by none. Internet usage took time to get to what it is today. Adoption of the new AI-based products that survive will likely look the same. What you build doesn’t have to be adopted by everyone in the world immediately. It just needs to be really useful to a growing cohort of your customers.

Marketing Technology News: Harnessing the Power of SMS Marketing Automation in your MarTech