Looking past the hype of real-time personalization
Updated: Mar 10
Every year the consulting firm Gartner publishes a highly interesting study. The "Gartner Hype Cycle for Digital Marketing" is an index of current buzz terms and marketing trends. At the lowest point of the latest hype curve, the "Trough of Disillusionment", we currently see the keyword: personalized advertising.
At first glance, this seems surprising. Personalization and targeting have long been indisputable components of any contemporary marketing creed. But the personalization hype of recent years has given way to a certain disillusionment. The great vision of targeting could not be fulfilled: to reach people with content that really interests them, directly and exclusively, without losing anybody along the way.
This also applies to online retailing - and it's not just due to the death of third-party cookies. The problems are much more fundamental.
In e-commerce, it is all about personalized real-time product recommendations. Anyone searching for a new winter jacket at an online fashion retailer will immediately receive suggestions for other products that might be of interest to them. Such "recommendations" are generated by algorithms, artificial intelligence and, above all, data. Lots of data. And this is where the problem begins.
Only big players like Amazon or Zalando manage to feed their systems with enough data to make really meaningful product recommendations to their customers. Most online stores, however, are overwhelmed as only a few of them possess the immense amount of scatter-free data needed for an entire product range. And even if they did have these data pools, they would hardly be able to process and clean them in a meaningful way.
The result is usually the exact opposite of an individual product recommendation. Based on their narrow data base, most online stores merely suggest products that many people have bought anyway. All of them are stuck in the bestseller bubble.
With the imminent end of cookies, the problem is getting worse. In 2022, Google will ban all types of advertising cookies from its Chrome browser. Apple and Firefox have already done so. The cookie concept has been crumbling for some time. Two-thirds of all customers are annoyed by the query layers with a quarter of online shoppers altogether rejecting website cookies. Personalization in its current form has already lost a large proportion of potential customers. Real-time personalized advertising, as we have wanted it up to now, does not and will not exist.
What can online stores other than Amazon or Zalando do? On the Gartner cycle we find a solution to the problem - still at the very beginning, pre-hype, so to speak. It is called visual intelligence. A solution that does not require any cookies at all.
Searches and recommendations based on visual intelligence are no longer science fiction. Via Google's Lens app, visual search has arrived in the everyday lives of many people. Major retailers like H&M and platforms like Ebay are already using this technology.
The advantage of image-based product search and recommendations: they create a shopping experience that is much more natural and intuitive than anything historically collected user data can offer.
Two examples: A customer uses her cell phone camera or uploads a picture from Instagram in the online store to find a specific product or similar product suggestions (Visual Search). A customer browses the product page of an online store and, based on her search, receives style-appropriate product recommendations that support her current purchase decision (Visual Recommendation).
Within milliseconds, more than 1,000 features of the image are analyzed, and stylistically suitable products are suggested. The AI reacts live and almost as intuitively as the human gaze. Prior knowledge about customers is not necessary. The technology works regardless of whether the person has visited or bought something in the store before. At the moment of their purchase intention, the AI reads the customer's wishes from their eyes.
As a result, search hits and suggestions reflect the image customers have in their minds. Product suggestions based on visual intelligence have been proven to lead to higher sales figures. In tests we have conducted with customers, the conversion rate to recommendations has increased by 258 percent.
Visual intelligence is not yet a hype. But that will change very soon. The experts at Gartner predict a fast career for it as many online stores will offer image search and visual product recommendations within the next two years. By comparison, personalization, which we have been talking about for many years and which has already passed its first hype phase, could still take another ten years.
Even beyond large retailers, visual intelligence already today is improving product searches and recommendations for medium-sized companies in e-commerce. And not just in industries such as fashion or furniture, from which one might expect the use of such image-based tools: Food or spare parts can also be sold more successfully with them. The e-commerce industry should not leave this topic to Google and Co. Because the dream of real-time personalization for all customers is over.
Philipp Derksen, 43, is founder and product owner of Vviinn, an AI-based software solution for visual product search and product recommendations for online stores. The Berlin-based entrepreneur and graduate engineer has been working in online business for over 20 years and in e-commerce for 15 years. His company Mediaopt has been developing software solutions for the e-commerce industry since 2009 and advises companies on their digital sales strategy.