Platform Taps Human Psychology to Give Shoppers Suitable Product Selections

Anabel Maldonado’s media site, The Psychology of Fashion, serves as a platform that merges two of her strongest interests: fashion journalism and psychology. Maldonado is now evolving that interest with a soon-to-launch, e-commerce and AI-powered platform called Psykhe.

Here, the founder and journalist discusses the new platform, how it works and how it differs from other personalization technologies in the market.

WWD: What is Psykhe and how does it work?

Anabel Maldonado: Psykhe is an e-commerce platform that uses AI and psychology to make product recommendations based on the user’s personality profile.

As a b-to-c aggregator, the platform pools inventory from leading retailers. Psykhe’s algorithm, powered by machine learning and personality-trait science, tailors product recommendations to users based on their Big 5 personality trait scores, which are captured through a psychological test taken on sign-up.

Psykhe is launching its recommendation engine technology within the fashion e-commerce space, but has b-to-b plans to expand to other consumer verticals in the near future.

WWD:How does it differ from other platforms in the market?

A.M.: Other platforms that employ machine learning don’t use psychological data.

View Gallery

Related Gallery

New York City Hotspots Reopen

Psykhe’s algorithms rely on a psychological framework and personality-trait science, not just surface product details and past purchase history. Our use of psychographic data, the added layer of psychology, understanding why each consumer responds to certain aesthetic details, renders Psykhe’s algorithm more powerful and predictive than current models, which look for patterns around basic product attributes with less context.

Other platforms that employ machine learning without user psychology, rely on the assumption that if you show the user enough products, that they will click around enough for it to establish a pattern. This way, initial recommendations aren’t really personalized, and could feel random to the user. Psykhe’s recommendations are very tailored right from the start, thanks to the personality test results.

On a product and taxonomy level, what we do differently is not only look at surface attributes, such as a hem length or print type, but how all these product details amalgamate into an overall underlying quality that corresponds to a good match for a particular user psychological profile.

Our model assesses all product details to assign psychological attributes to over 1 million products from our commerce partners, enabling more emotionally attuned matches for the user. Psykhe’s personal understanding of the consumer captures stronger patterns around what exact combinations of detailing evoke a gut-level reaction. Our model seeks to mimic the way that humans can intuitively look at a product and say “that’s me” or “that’s not me.” This level of alignment is what we are measuring and systemizing using AI and psychology.

Anabel Maldonado 

WWD: What is the user experience like?

A.M.: The user takes a three-minute personality test (the shortened version of the official, “Big Five Inventory-2/BFI-2” test) and is led to their profile page, where they can read the explanation of their test results in an editorial format, and upload a profile picture. From there, they can also access the storefront, where they see their top product matches across all categories. As an aggregator, Psykhe pulls products from its affiliate partners such as and Moda Operandi. If a user would like to buy a product, they click-through from Psykhe, as checkout and fulfillment happens on the partner site.

Users can filter their results by Product Type (clothing, bags, shoes, accessories and essentials, which includes blankets and candles); Occasions (workout, chill, everyday, business, evening, holiday); Mood (calm, happy, adventurous, confident, romantic); as well as other filters such as Sustainable Credentials, always returned in a personalized order based on user personality scores.

If a user dislikes a product, they can “zap” away products they don’t like by clicking on an “x” below each product, and the product is instantly replaced with another, creating an element of variance and surprise. Users can also save products to specific lists, such as for events or holidays, unlike other platforms that tend to have one generic wishlist. (All these tuning actions provide data for our global and individual-user models.)

WWD: What was the impetus behind creating this recommendation engine?

A.M.: Ultimately, the impetus behind Psykhe was to enable technology to really know the consumer, promote self-knowledge among consumers themselves, create better connection between person and product, and in turn, inspire good consumerism by changing how we think about clothes.

During my time in the industry, I was increasingly frustrated with the narratives that made fashion seem more frivolous than it really is. I thought, “this is a trillion-dollar industry, we all have such emotional, visceral reactions to aesthetic based on who we are, but all we keep talking about is the ‘hot new color for fall’ and so-called blanket ‘must-haves.’” There was no inquiry into individual differences. The fashion world was really missing the level of depth it deserved, and I sought to demonstrate that we’re all sensitive to aesthetic in one way or another and that our clothing choices have a great psychological pay-off effect on us, based on our unique internal needs.

I sought to change how we think about the purpose of fashion, instead of putting consumers through an endless cycle of trends and social validation-seeking, creating waste and confusion. I decided to devote my time to the burgeoning field of fashion psychology — why we wear what we wear. During this time, I looked increasingly to the Big 5 model of personality, which has known, proven correlations with fashion and other aesthetic preference, behavior and life outcomes.

While I knew that a fashion psychology framework was valuable, even if I wasn’t yet sure about the ultimate application, it soon became apparent with the growing need for personalization, that AI was the third missing piece of the puzzle. I realized that our framework lends itself to create the first AI and psychology-based recommendation engine, so that people can find things in alignment with who they are more easily.

WWD: What do you think is driving the demand for greater personalization?

A.M.: The need for personalization simply stems from our need for a more “human” experience across the endless and increasing digital channels and platforms that we encounter. Personalization is just humanization: reducing choice for the end-user in a considered, sensical way. In real life, humans offer options to other humans in this way. Say you have a friend over for lunch. You might offer them a soup, sandwich or a salad. If they’re a good friend, you’ll know their dietary preferences, and let them know that you have gluten-free bread. You wouldn’t offer them 30 random things, the way an unpersonalized platform would.

The effects of COVID-19 gave a major boost to global e-commerce. Cracking personalization, in the fashion space and otherwise, is key to unlocking the new era, the third decade of e-commerce, where we go from manually trolling, filtering and searching, to encountering intuitive machines that have learned our preferences and habits. Psychology is necessary for these machines to become smart and know the user. With all the technology we’re using, data and machine learning make it possible to personalize and humanize.

As AI evolves toward the Internet of Things, allowing devices to analyze data and make decisions without human involvement, psychographic data — a consumer’s Big 5 profile, and their mood fluctuation patterns — will be crucial for accurate decision-making.

Source: Read Full Article