Tricolor was losing active users. Two releases in a year turned the decline into growth
Second Screen is a mobile TV companion for subscribers of Tricolor, the largest satellite operator in Russia. I designed the product from scratch and took it through two releases: the TV schedule, a link to the set-top box, TV streaming to the phone, voice control, and offline viewing.
- 236K+
- active
users - 4 mo.
- Android MVP
in the store - TV
- mobile became the base
of the TV app - 68%
- say they watch
TV better

Problem
An outdated set-top box could not go online — and the subscriber left
Hundreds of thousands of Tricolor subscribers used outdated set-top boxes with no internet access. Because of this, the company was losing active users: by the time I started, their number had fallen from 250,000 to 198,000 in a year.
Goal. Retain customers. Metric: a steady, lasting decline in churn. Hypothesis: new value that raises loyalty will keep them.
Tricolor shared the data of its own research: what subscribers do while watching TV.
63% use a smartphone while sitting in front of the screen
48% check the TV schedule in apps
31% look up forgotten movie titles and actor names
My contribution — three decisions over a year of work
Decision 1. MVP
What to give up to ship in 4 months
Context. The deadline for the first release was set by the client — 4 months to the store. We knew what hurt from Tricolor’s data. How people behave in the interface — we did not, and there was nowhere to get that data: the audience is closed, only current subscribers. These are people aged 35–55 outside big cities, where homes are not always covered even by 3G.
Decision. We cut the scope to fit the deadline, and I was responsible for what went in. I used one selection principle: the MVP includes what works right after signing in with Tricolor ID, without linking any equipment. The MVP was a tool for testing, not a smaller copy of the product. I made up for the lack of behavior data with what I had: competitor analysis of the NTV-Plus and MTS apps and public reviews of similar products.
ResultThe Android MVP reached Google Play in 4 months, with iOS right after. The cost of the principle showed up at once: subscribers expected to control the TV, did not find it, and went to support — the load on support grew. That set the scope of the second release.


Decision 2. Redesign
How to roll out a redesign I argued against
Context. After the first release, the client set a task: a redesign in neomorphism instead of the flat design of the MVP. At the same time, the second release also brought new features — everything had to be reworked at once.
Decision. I argued against it: for our audience, contrast is critical, and the style did not fit the Tricolor ecosystem. The client did not change the decision. So I stopped arguing about whether to do it and focused on how to do it without breaking the release: I split the transition into five stages by layers — typography and spacing first, then shadows and depth, then color. Development received changes in portions, and new sections went through the same stages.
ResultThe second release shipped on time: all sections were reworked, and there were twice as many of them — on two platforms. Shadows were harder on Android than on iOS and slowed the team down. Contrast suffered — exactly what I had warned about.
A conscious trade-off. I did not change the decision — I changed the way it was rolled out, and that saved the team’s deadlines. I knew the price in advance and accepted it on purpose: an argument about taste is won not by arguments but by data, and we had none.



Decision 3. Second screen features
Linking the phone with the set-top box
Context. The second release had to make the phone a second screen — what subscribers did not find in the MVP. Asked what the TV app was missing, 51% named a smarter remote, and 44% named streaming to a smartphone. This had to be done at the same time as the redesign.
Decision. I linked the phone with the set-top box: I made connecting to the receiver a separate section — it opened the second half of the product. I designed the player and offline viewing from scratch — for homes with poor mobile coverage. I did not split control — remote and voice — into separate navigation items and combined it in “Services”, so that navigation would not grow with every new service.
ResultThe phone became what people expected — a remote and a screen.



My role: Senior Product Designer
- There were two designers on the product: my area was all of iOS and the dark theme, and the flows were designed to be shared by both platforms.
- At every stage, I worked on scenarios, flows and navigation architecture, and chose the color palette, shadows, corner radii and element shapes at the same time. This helped to agree on a complete product with the client and not go back at late stages.
- During the redesign, I led the neomorphism: I refined the style and rolled it out across all iOS sections.
- I took a junior designer under my wing — Android screens and the Android UI Kit. For the first months we worked side by side, and by the end of the third month she closed tasks on her own.

Result and reflection
The decline stopped in the first year after the second release
Two months after the second release, Tricolor surveyed its subscribers:
- 68%
- said that watching
TV changed for the better - 54%
- open the app
three times a week or more - 50%
- called the app
useful for them
I joined on a falling curve. A year after the second release, churn slowed for the first time in several years, and a year after that it turned into growth.
- 198K
- at the start
of my work - 205K
- a year later,
the decline stopped - 236K
- a year
after that
I led the product up to the second release: the numbers of the following years were shaped by more than my decisions. But the turning point came in the first year after them.
First release: Android and iOS, light and dark themes. Sign-in with Tricolor ID: TV guide, favorites, a personal “For you” feed, entity cards, onboarding, profile.
Second release: all sections reworked, plus new features — connecting the receiver, a player with TV streaming to the phone, recording shows, a remote, voice control, offline viewing, templates for stories and articles.
The design of a TV app based on the mobile one began. For me this is a confirmation: the product met a real need and earned trust for the next step.
What I would do differently
- Product format. Some features could have been built into an existing Tricolor product, but the question never came up for discussion. Now I raise it myself, before design: a separate app or a feature in an existing product is a product decision, not a technical choice.
- Behavior data. We knew what hurt, but not how people behave in the interface — this affected every argument, not only the one about neomorphism. Now I build in at least one round of checks with users, even into a tight timeline.
- The speed of release cost us the quality of the technical implementation. Complaints about connecting to the receiver were predictable, but nobody raised this risk early. Now I would insist on a technical buffer before the release.
Contact me