When it comes to online personalization, businesses are hitting a wall. They’re navigating a new Wild West of rising customer expectations, highly saturated markets and the decline of third-party cookies. But the stakes are high. Across industries — retail, streaming, even events — consumers are getting tired of subpar online experiences. A mere highlight reel of what’s trending doesn’t cut it. Busy people demand tailored personalization to help discover their next favorite thing.
The outlook might seem bleak. But there is hope. For ecommerce companies in particular, that hope lies in the rethinking of how they sell their products. It all has to do with “DNA” (no, not the kinds that make up genes — just the specific data that expresses unique qualities), both of products and consumers. By understanding, then pairing, these two adjacent strands of DNA, companies can achieve the “double helix” of the perfect recommendation — thereby increasing conversions, loyalty and long-term success.
Digging Deeper than Metadata
Most ecommerce companies are familiar with metadata, which provides basic, surface-level information about other data. But only a handful of those companies understand the value of taking it a level further and looking at “DNA.”
What does this look like in practice? It looks like analyzing all of the pieces of data that are used to mathematically represent a product. That’s not limited to standard tags like “description” or “category,” but also the more granular details, like the inputs of the photo or cover art used to represent the item, or even user-generated content related to the item, like reviews.
Then, AI-powered deep content extraction can interpret each of those granular elements to form a product’s unique DNA strand. The result is a far more holistic interpretation of a product’s personality. A trench coat and a leather jacket, for instance, will both show up when a customer filters by “outerwear,” but they’re remarkably different options.
The same strategy holds true for understanding and applying customer DNA. In just a few clicks on a website, a consumer provides huge insight into what they’re likely to want to see — and purchase. Even without personal identifiable information (PII), the right algorithms can reveal how a certain user tends to feel about certain kinds of products, how they tend to interact with them, and what that implies for their probable next actions on the site.
This kind of behavior-based approach (in contrast to demographic-based tracking methods such as third-party cookies) yields incredibly tailored, in-the-moment clues about how to effectively personalize their experience. As such, a customer’s DNA strand essentially breaks down to what makes them uniquely interested in a given item at a given time based on their real-time behavior — allowing a retailer to know for certain whether they’ll prefer the trench coat or the leather jacket.
Forming the Double Helix
For most ecommerce companies, the key to success is determining how to surface the right item, to the right person, at the right time. For decades, business leaders have bought into the myth that third-party cookies are the way to achieve this goal. But the truth is it’s never been the best method, and there’s a more accurate and efficient way to achieve this success. They’ve got to take the disparate strands of product and customer DNA, and combine them to form the “double helix” of the perfect recommendation. In other words, the right algorithms can intelligently match who’s most likely to be interested in what, purely based on their real-time behavioral data points.
By weighting both strands equally, businesses can take a far more holistic approach to personalization. The approach pays dividends: 91% of consumers are more likely to do business with brands that remember them and provide relevant recommendations and product offers.
On the flip side, customers say that they typically take personalization for granted — but if a retailer gets it wrong, they may depart for a competitor. The stakes are high, but if businesses can learn how to leverage the real-time data that they already have at their disposal — both on their products and on their customers — they can form that double helix and start pulling ahead of the pack.
Solving the Cold Start Problem
One crucial benefit of centering product and customer DNA is solving the cold start problem — the two-pronged phenomenon in which businesses don’t have the right technology to provide accurate recommendations to new and anonymous users, or to recommend new products that haven’t yet been interacted with onsite.
The cold start problem has plagued ecommerce providers for years. It’s only been exacerbated with third-party cookies fading fast, and many companies are scrambling to figure out how, exactly, to target new and anonymous customers. It’s a major dilemma, and one that requires quick thinking, given that up to 68% of people visiting a website are new customers and 98% are anonymous.
The DNA approach allows businesses to confront this problem head-on. Deep content extraction technology can analyze what makes an item unique, even without anyone having interacted with it yet. It can simultaneously identify which user it should surface the item to, based on their demonstrated behavior-based preferences. The result? A brand-new item could be added to a catalog and feasibly recommended to a user who’s likely to be interested in it within mere seconds. It’s a win-win: the consumer gets to immediately buy something they’re excited about, and businesses boost their conversion rate (and number of loyal customers to boot).
The DNA approach is a completely new paradigm for ecommerce personalization. Every business has the data at their disposal to leverage this method — most just aren’t aware of how to unlock the power that data holds. But if they deploy the right AI, they’ll supercharge their approach to personalization, help customers discover more of what they love and build better long-term relationships along the way.
Alexandre Robicquet is a Stanford AI scholar and current CEO (and Co-founder) of Crossing Minds, an AI-powered recommendation platform for ecommerce and content. He holds three Master’s degrees in mathematics, machine learning and artificial intelligence.