It was called a planogram. 2011, beauty retail: a diagram of which cream stands where, what goes at eye level, what goes near the floor. I executed that algorithm by hand — jar by jar. If you think machines started deciding for shoppers recently: no. The machine used to be me. I was in my final year at university then — marketing and advertising. The shelf taught faster.
Then I spent years traveling store to store as a field auditor of shelf placement across a national FMCG market. The standard on paper is one thing; the shelf is another. The difference between them is measurable — and you can hold someone to it. That thought never let go of me.
There were also years at a charitable foundation in Dnipro — marketing, fundraising; that is where I first saw how a nonprofit works — from the inside — and where my design practice began.
After the foundation I went out on my own — brand packaging: identities, social media, print. I was getting other people's products ready to meet the shopper.
Then came my own product photography studio in Dnipro, Ukraine. I opened it knowing exactly what I wanted to shoot: products, not people. The work gave my enthusiasm and imagination room to run — every type of product was its own task, with its own props and its own rules, and I wanted to know all of it, end to end. We went through a pile of failures and far more successes; I remember those years with love and gratitude. The camera also taught me what spreadsheets cannot: the image decides before the words do. The tone a product is shot in, whether there is a hand in the frame, whose hand it is — a shopper reads all of it in a fraction of a second.
It was there, at the studio, that the idea itself was born: the next step was meant to be a research practice built on it — a place where shelves and storefronts are not only photographed, but studied. The war in Ukraine cut across those plans and set them back for years. But the idea survived — I carried it with me all that time.
Meanwhile, retail kept moving online, and an algorithm now arranged the shelf. Eye level became the top row of recommendations. Facing counts became the order of product cards.
Everything I once did by hand is now done by code — except nobody comes to check the code.
The link between an image and a purchase decision was never a theory to me — it was right there in client feedback after every shoot: this started to sell, this did not. I wanted to scale that thought — to make the problem visible to those who do not notice it. I dreamed of making a contribution to this industry, however small. I kept studying the field the whole way — and in 2026 I understood: I had the knowledge and the resources for a step this serious. It was time.
That is how AVBR Lab came to be — an independent nonprofit: we audit e-commerce recommendation algorithms from the outside, the way a shopper meets the shelf. No agreements with platforms. No access to the internals. Only an open protocol any researcher can rerun and get the same numbers. The first wave covered Shopify; the BigCommerce census is complete; the program is scheduled through 2027.
For fifteen years I stood on both sides of the shelf. Now I stand on the third side — the checking side. The view is best from there.