From one store to an omni-service
Turned one of Russia’s grocery chains’ app from a hypermarket delivery service into an omni-app covering five store formats. A/B tested to 30% of traffic: format switching 0 → 1.62%, loyalty engagement +25–30%, offline ARPU +1.99% — against a 1.36% dip in order conversion.
Role
Product Designer
Industry
Retail
Year
2025
Platform
iOS & Android
Context
Five formats. One app that only knew one.
Over five years the group added store formats — a pet market, a discounter, a wine shop, and express delivery in some cities. The app did not follow. It was still built around the hypermarket: one store, one basket, one way in. People had to figure out what the formats were, which one had the thing they wanted, and how to move between them. Most did not. They stayed in the format they knew, or left to find the item in another service.
The business needed the opposite. Each format carries its own range, promos, delivery terms and pricing —and none of that survives inside a 35,000-item hypermarket catalogue, where there is no room to put a premium pet food line or a discounter promo in front of anyone.
I worked as one of several product designers on the squad, alongside a product manager, a researcher, analyst and the iOS & Android developers. I owned the format-selection page, the loyalty and offline blocks on it, and the validation round — the interview brief, the sessions, and turning what came back into a fix list before development started.
Challenge
Constraints
— The architecture assumed a single store. Navigation was hierarchical inside one format, so a page sitting above all formats had to be grafted onto a model that did not expect it.
— Order conversion could not be sacrificed. Any movement had to be explainable.
— Three audiences with almost nothing in common shared one screen: offline card users, online shoppers,and people who had never ordered.
— The loyalty card had to be one tap away without taking the page over.
— Jumping straight to a product without entering a store stayed out of reach — the navigation stack did not support it.
— The platform teams were mid-modularisation and running parallel start-up experiments, so everything newlanded in a moving codebase.
Discovery
Three audiences behind one entry point
Ten brainstorms turned the business question into user tasks — from a two-hour delivery to showing a barcode at the till. Analytics gave us the split, and the split decided what the page had to carry: 40% of users open the app only inside a store, 30% order online, 30% downloaded it and never ordered at all.
The number that framed the whole project was the share of users who had ever changed store format: zero. Not low — zero. Format choice was buried in the delivery-method selector, and nobody read it as a choice.
We priced the work honestly before starting: 10 brainstorms, 500 mock-ups, 7 A/B tests, 1,640 hours.
40%
use the app only for the loyalty card, discounts and the bonus game
30%
order online for delivery or pickup
30%
installed the app recently and never placed an order
0%
ever switched store format before the redesign
Delivery & category selection format in the old version
UX workshop
Core concept
Make the choice a page, not a setting
The core hypothesis was blunt: a dedicated page would do what a hidden selector could not — say what the formats are, what each is for, and how fast each one reaches you. I built it around the 3 segments rather than around the catalogue, so each audience finds its own block within a screen of arriving.
Want to spend less? Econom. Premium cat food? Zoomarket. Forgot the yeast for a pizza? Express delivery. Each format card carries a delivery estimate, a minimum order and a one-line reason to pick it — the copy does the sorting the navigation had been failing at.
01 - Format Switcher
Every format in one block, each with its delivery window, minimum order and reason to exist.
02 - Loyalty Card First
Barcode and cashback categories one tap from the top, sized so they never take the page over.
03 - Stores Nearby
Which formats exist in your city, how far they are, and whether the item is in-stock before you leave the house.
04 - First-Order Offers
Delivery and pickup discounts aimed at the 30% who installed the app and never used it.
Nearby Stores help users explore available store formats, compare travel distance, and check product availability before visiting
Color was the fastest way to tell formats apart
Interviews were unambiguous on one point: color and large, literal icons separated the formats faster than any label we wrote. So each format got a color, a mark and a fixed card anatomy — name, range size, delivery estimate, minimum order. The same card works in the switcher, in the nearby-stores list and in the bottom sheet, so a format reads the same everywhere it appears.
That turned the page into a filter rather than a menu. Post-release analytics on format switching and catalogue entry confirm people use it that way: pick a format, then start shopping.
Updated store homepage design with color-coding to distinguish between categories
What shipped is a single entry point stacked by audience: the loyalty card and cashback categories at the top for the offline half, the format switcher below it, nearby stores under that, and first-order offers for people who have not bought yet. Previously you had to be inside a specific store before you could see what it sold. Now the choice happens before the catalogue instead of inside it.
Validation
Nine interviews before any code
I wrote the brief and ran depth interviews on concept prototypes with 9 participants, then ranked the concepts on what came back. Two hypotheses were on the table: that showing every store and every order option in one place would make people order more often, lifting ARPU and LTV; and that putting the loyalty card and bonus game on the same page would raise offline conversion and bring people back more often.
What the interviews revealed
— Some users believed the app was for delivery only, and did not know other collection methods existed at all.
— Format differences did not read from names. Colour and large icons did.
— Card-only users needed the barcode immediately and would not tolerate it competing for space.
— The bonus game, shown in full, overloaded the page.
What changed before development
The loyalty block was cut back to the barcode and cashback categories, with the bonus game reduced to its core action. Format cards were rebuilt around color and icon instead of name. Collection methods were spelled out on the card itself rather than one level down. The concept went into build ranked against alternatives, not chosen on preference.
9
depth interviews with clients and non-clients
2
hypotheses tested before a line of code
4
findings that became the pre-development fix list
7
A/B tests run across the redesign
Final concept
Impact
A deliberate trade: online margin for offline reach
The final design was released as an A/B test to 10% of the audience first, to keep any conversion movement contained, then widened to 30%. The test ran for a month.
Format switching went from 0% to 1.62% — people started choosing a shopping scenario instead of defaulting into the hypermarket. Loyalty-card displays and cashback-category selection rose 25–30%. Catalogue entry rose 1.6%: small, but stable and statistically significant. Offline ARPU rose 1.99%, pickup conversion 0.04%.
It cost something. Order conversion fell 1.36%, online ARPU 1.22%, online LTV 0.87%, delivery conversion 0.2%. Support contacts originating from the new page appeared at 0.12% — low enough to say the page itself was not confusing anyone.
Adding a choice slows the fastest path. Users who already knew what they wanted now pass a screen they did not need, and that shows up in online conversion. What it bought was the thing the app had never had — a measurable share of people deliberately picking a format, and a much stronger offline loop. Recovering the online path is the next iteration job, not evidence the direction was wrong.
Other projects
2x conversion on the mobile homepage
Redesigned a top-10 Russian bank's mobile homepage from discovery to design language. Validated with eye-tracking and a staged A/B rollout: product-showcase conversion 4.2% → 8.14%, with no metric regression.
TikTok Creators: Safety Tools
I turned enforcement-quality data into structured product recommendations, built the policy-to-product layer underneath safety surfaces, and designed the internal tooling and measurement moderation operations ran on. The direction that evidence argued for — prevention at the point of creation rather than review after it — is the one TikTok’s 2025 creator-safety tools follow
From LMS to streaming service
Rebuilt a Russian cultural-education platform — 157,000 users, 100+ courses — from an ageing LMS into a streaming product, then shipped iOS and Android off the same system in ten days with no additional design work. Purchase conversion +32% on day one, D7 retention 23% → 30%, authorisation 65% → 91%