Running contextual ads in the fashion segment is always a walk through a minefield. An overheated auction, marketplaces buying up top search slots, and temperamental automated bidding strategies that burn through budgets at the slightest mistake. Practicing marketer Sergey Ignatyev from the agency Hope Group shared a detailed case study on relaunching campaigns for a Russian women's clothing brand. The starting point was grim: a "fading" Yandex Direct account, inflated cost per lead, and an ad spend-to-revenue ratio (ДРР) exceeding 53%.
The goal was uncompromising: rein in the algorithms' appetite, bring ДРР down to the 15–20% range, and take the store to a steady 2.5 million rubles in monthly revenue. Let's break down step by step how the team pulled the project out of a financial hole.
Cold Start and the Off-Peak Season Pit
The first and most insidious trap business owners fall into when relaunching contextual ads is expecting instant sales in week one. When a campaign starts from scratch or after a long pause, the account needs time to calibrate. In this project, the algorithms trained for nearly a month. In the first week, exactly one lead came in; in the second, five. A 500% increase looked nice on paper, but in revenue terms it was a drop in the bucket.
Seasonality made things worse: the relaunch took place in May and June. During that period, the brand was selling light summer items with a low average order value. With high auction bids, ДРР stayed at a critical 53%. Many owners panic at this stage and shut off ads, locking in losses. Ignatyev's team, however, planned on a six-month horizon from the outset and got to work on routine account "hygiene."
They methodically purged non-target traffic:
- Disabled junk placements in YAN (mobile clickers, game showcases, and spam sites);
- Added hundreds of negative keywords for near-target and informational queries;
- Separated the automated Campaign Master from the classic product gallery so the algorithms wouldn't cannibalize their own traffic;
- Set up strict retargeting for abandoned carts.
In similar situations involving reducing an online store's ad spend, this surgical cleanup of junk placements becomes the only way to stop uncontrolled budget drain.
Relaunching Creatives and Reaching Target Payback
The second critical lever was visual content. Instead of generic stock templates, the brand's founders began shooting live content in natural urban settings, highlighting seam quality and fabric fit. Authentic photos with a clear value proposition boosted ad click-through rate (CTR), which lowered the effective cost per click in the auction.
The turning point came in August with the launch of the fall-winter collection (coats, knitwear, insulated outerwear). Here, a classic unit-economics combination kicked in:
- Average order value more than doubled.
- By that point, the Direct account had accumulated a dense conversion history, and automated strategies began pinpointing paying audiences accurately.
As a result, ДРР first dropped to 30% and, by peak season, settled at the target 15%. Cost per qualified inquiry (CPL) came in at 1,500 rubles. At outerwear order values, that lead cost delivers strong return on ad spend (ROAS) and more than recoups the investment of the early ramp-up months.
The main takeaway from this story is simple: contextual advertising in e-commerce is not a lottery with a magic button, but sequential math. If you understand your seasonality and are willing to patiently tune analytics during a cold start, Direct turns into a predictable revenue generator.


