From 126% to 12.7%.
Four years promoting the Chitai-Gorod mobile app at MediaNation: from a loss-making start at 126% DRR to a 12.7% peak. I brought the client into the agency — and I still own this work.
Below is my take on the case. The full write-up with stage-by-stage numbers is on the MediaNation site: Chitai-Gorod case ↗.
Why this one is personal
I brought Chitai-Gorod into MediaNation years ago. We have been running the mobile app ever since, and I am still in touch with the client team. So this is not a distant agency case — it is work I own as co-owner and CEO.
Start: 126% DRR and choosing Android
The app launched in April 2020; paid acquisition started in May 2022 on Android and iOS. The first months looked bad: DRR 126%, ROAS 79%. On iOS there were only three purchases in month one at a CPO that did not pay back. We turned iOS off and put all effort into Android — that is where people actually bought. Everything below is an Android story.
Analytics first, then unified performance campaigns
While AppMetrica only sent installs into Yandex Direct, the algorithm optimised past the economics. We wired postbacks for content views, add-to-cart and orders — including revenue. After enough data accumulated, campaigns moved to cart, then purchase optimisation.
We started in Campaign Wizard: first purchases and a training sample. When volume grew, we moved to Unified Performance Campaigns — Search, the Yandex Advertising Network and smart banners under one strategy. UPCs worked because real economics were already flowing into the account, not just installs.
Four years of movement
By December 2023 DRR was down to 28.4%, CPO down 76%, and monthly purchases were up by orders of magnitude. A separate postback for first purchases trained campaigns on new buyers in parallel with warmer audiences and the existing app base. In January–March 2024 first purchases rose 69% with CPO down 41% and average monthly revenue up 45%.
March–August 2026 brought the category’s seasonal dip. Purchase volume and revenue cooled versus the 2025 peak, but profitability held: DRR stayed in the 27–40% band, 27.4% in August — 2.4× better than start on DRR and 4.6× better on CPO. First-purchase CPA rose only 8% over the half-year while budget was tighter — UPC algorithms were still finding new buyers.
Takeaways
- Without postbacks, optimisation is blind. If Direct only learns from installs, ads buy cheap clicks, not buyers.
- The tool follows the data. Campaign Wizard got us started; UPCs scaled. Both work when the account sees revenue and the funnel.
- Profitability holds for years if you run the channel like a product. 12.7% was not a one-off miracle — it was four years of sequential decisions.
Thanks to the Chitai-Gorod team and our MediaNation team — especially everyone who stayed with the Android channel for four years and pushed analytics through to a first-purchase postback.