My team and I redesigned the new product submission process for Zdrowe Zakupy, a barcode scanner app that shows a health score based on ingredients, additives, and nutrition values.
My contributions included researching how the old process worked (or rather, what didn't), conducting user interviews, analysing the data and qualitative feedback, hashing out the user experience, and prototyping the final solution to speed up new product submissions.
With grocery stores constantly rotating and expanding their shelves, users were hitting a blank "new product" page on roughly every fifth scan.
The existing flow asked them to type out the full ingredient list by hand. For a product with 20+ ingredients, standing in an aisle, holding a basket, squinting at tiny print, this took several minutes and required some commitment. Sometimes mistakes were made, injecting bad data that frustrated everyone downstream.
What's more - missing products were, unsurprisingly, the most common complaint in app store reviews.
How might we ensure users can reliably find newly launched products in the app despite frequent changes in store inventory?
How might we streamline or automate product submission and review to minimize delays and errors in adding new products?
We looked into the qualitative data and conducted a few interviews with power users to get a deeper understanding of the problem. It became clear that users needed the value instantly, instead of getting a "new product" screen. They were fine with occasionally adding a new product themselves — but they also needed to receive the score information back while they were still at the shop.
Any screen requiring more than one line of instruction was skipped. The flow had to be self-explaining through visuals alone.
Without prompting, every interviewee said "can't I just take a photo?" — it's how they already save information on their phones.
After submitting, there was no feedback. "Did it go through?" — the silence felt like failure, so people stopped trying.
If it took longer, users put their phone away. The in-store context is unforgiving.
The redesigned flow replaces all manual entry with three photographs: product front, ingredient list, nutrition label. A backend OCR system extracts the data. Users contribute without typing a single character.
The user is asked to add 3 photos to add the product to the database.
Product front, ingredient list, nutrition label. Visual overlays show exactly what to frame.
Photos go to the backend. OCR extracts everything. Under 30 seconds total.
When the score is ready, the user gets notified. Their scan now works.
It was about 5% of no-product-view from that month. So the no product view dropped from 20% to 15%.
We were scared that users would send inappropriate photos. But it didn't happen.
Even if 3 photos sounds like a lot, it's still smoother than the previous flow.
We did everything we could, but some users complained that they had to wait several days for the approval.
Some users said the process was still long and they'd like to make it even faster — we decided to eliminate some of the steps.
About 30% of users didn't notice that this feature had changed, so we should think about an onboarding process.