Shoppers buying gifts for others are haunted later by these searches when shopping for themselves.
That neatly explains why intent is the next frontier of eCommerce recommendations. Speaking with PYMNTS’ Karen Webster for the JPMC Merchant Series: Global Innovators in Payments, Alexandre Robicquet, CEO and co-founder of recommendation engine Crossing Minds, said: “When you ask for a recommendation, you don’t ask a complete stranger. You ask someone that knows you and has an intimate understanding of what motivates and drives you.”
He joked that you don’t need a Harvard MBA to understand the logic. When someone knows you, their recommendations are informed by the movies and books they know you like, “what things you don’t like, what music you listen to. All those little bits of information are extremely valuable to provide context,” including if you’re shopping for yourself or someone else.
It’s the same with online recommendation algorithms that are over reliant on basic browsing and demographics. Far more important is “the mindset and the mood that you are in. There’s a very contextual principle tied up with that,” he said. Offer a retailer a choice between basic browsing data or deep insights into a person’s preferences at different times, and they’ll take the deep insights.
Understanding intent is an upgrade from the demographics retailers and brands have long used, and AI-powered recommendation engines like Crossing Minds are tapping into intent and mindset to discern what consumers really want, be it for themselves or someone else.
Robicquet said the power of AI recommendation comes from a consumer’s immediate context.
“First, only focus on the data that is explicitly shared and that happens in real-time. First-party data,” he said. “Second, try to do that with a very behavior-based approach.”
It’s a departure from the way recommendations have traditionally been done in eCommerce, where using basic browsing data doesn’t give retailers the context to offer relevant results for different shopping visits, whether it’s a birthday present for your partner or an item for yourself.
See also: Next 3 Years: Converting a Stranger in 3 Clicks With AI Recommendation-as-a-Service
Sellers Schooled By Streaming
Noting that Netflix reportedly used viewer data to inform the storylines of hit streaming series like ‘House of Cards,’ he said the commerce applications can be even more impactful.
“That approach has been seen in eCommerce and some other areas with this concept of clickbait versus the long-term conversion,” he said. “What images really impact buying? What prices impact in terms of converting? And then how do you translate that on the longer term into ‘did they return it?’ and did that actually create long-term value?”
In commerce, it’s important to understand objectives and timelines, since building a recommendation system or understanding personalization for in-the-moment engagement requires different approaches.
“Some impacts are measured after a year, and some impacts are measured after two seconds,” Robicquet said. “That’s very indicative of understanding the audience and how to position your brands and the AI you want to deploy to support the goals of your brand.”
Sharing the story of work Crossing Minds did for a book publisher, with different summaries conveying the story, and others focusing on the cover imagery, he said: “You would be shocked to see how much the cover [image] outperformed” story summaries.
This underscores the critical responses consumers have to different content in selling environments, which is why Crossing Minds focuses on clients extracting all information and allowing the AI and algorithms to choose what is most relevant to show different consumers.
In an eCommerce world driven by consumer reviews and star ratings, the evolving concept of product recommendations embedded in the site or marketplace stands to cause a meaningful change in how product sales are influenced by content and presentation.
Robicquet believes that what is already the default in search — people rarely going beyond the first page of results — applies to eCommerce too, as surfing Amazon and typing in a description will bring up a handful of thousands of SKUs that best reflect nuances in search terms.
On a three-year timeline, recommendations as we know it don’t vanish, but rather becomes “innate” to platforms, where the decision then becomes whether to show an ad for a quick sale or show matching recommendations to create lasting customer relationships.
More like this: Shopify Invests in Recommendation Platform Crossing Minds
The New Personalization
Asked about best practices in recommendations moving forward, Robicquet noted that brands don’t always appreciate the wealth of first-party intent data they already possess.
“Google Analytics is a perfect example of that,” he said. “You have a plethora of clicks, scrolls, moments they spend that can be so rich if leveraged properly to provide recommendations.”
Not that merchants and brands shouldn’t push something they know customers might like to see. Sellers can push items consumers didn’t specifically search for “if it is something that you define based on your customer behavior,” he said, as long as it’s “respectful and reciprocal.”
What’s important is to keep the “creepiness” factor low. That’s what killed third-party web cookies and what has created a need for intelligent recommendation engines. “Then trust is being built. It doesn’t come out of nowhere,” he added.
This kind of one-to-one session-based personalization overcomes the creepiness factors and takes recommendations into a real-time transaction that enables consumers to make a faster decision on the product they not only want in-the-moment but will want to keep.
Robicquet said this gets into creating different recommendations at different points in the journey, making a one-click checkout with adjacent upselling opportunities.
As for how this drives Crossing Minds forward, he said the focus is on making the recommendation engine easier to deploy and showing hesitant eCommerce merchants the selling power of recommendations-based AI tools that pick up on vital customer clues.
We’re always on the lookout for opportunities to partner with innovators and disruptors.
Learn More