High ad cost share (ACOS) on Ozon is rarely fixed by lowering bids or restarting campaigns. In most cases the fix is order inside the campaigns and inside the store's product cards. Below is what we did on real stores and the numbers we recorded: from 23% to 4%, from 14% to 7%, and below 1% in auto parts.
What sellers usually assume about ACOS
I talk to Ozon sellers regularly, and the complaint is always the same: "ads stopped working" and "ACOS went up". Once we dig in, 9 out of 10 sellers share one belief: ACOS goes down either by cutting the bid or by relaunching the ads.
In reality it is more complex. Ozon has around 35 metrics, and each one affects both search positions and ad ACOS. That is why most of your time should go into product card and store metrics, not into the ads themselves. The exact split matters less than the principle: account settings are the smaller part of the result.
Case 1: from 23.1% to 4.09% by analysing reports
We fed a partner store's ad reports to an AI model. It flagged loss-making campaigns, proposed a new testing strategy and pointed to specific changes in ad format settings. ACOS fell from 23.1% to 4.09%.
Generating hypotheses like these used to take us several days of spreadsheets and endless numbers. Now it takes about five minutes and produces a ready action plan. The AI was not the only reason for the result: a partner handled logistics and supply end to end, so stock never ran to zero and warehouses across clusters stayed filled.
Case 2: ACOS halved while the budget doubled
At a new partner store we cut ACOS from 7.74% to 4.17%. At the same time we doubled ad spend, from about $1,400 to $2,630 per week.
One step only: we optimised the campaigns. We removed outdated ones, ran tests and found the bids that actually worked. The partner had been overpaying badly and was not using potential the store already had.
Case 3: from 14% to 7.2% at a new client
At this client ACOS had sat above 14-16% for months. In the first third of March we brought it down to 7.2%, roughly halving the cost of acquiring a buyer. What we did:
- switched off tools that were draining budget without results;
- cleaned up the campaigns: the client had been spawning dozens of new campaigns, which caused overspend and needlessly pushed up the auction;
- introduced more effective combinations and rebuilt how the system worked.
Case 4: 0.86% ACOS in auto parts
First we brought a partner's ACOS down from 7-10% to 1.5-2% over a holiday period, then went further and broke below 1%. This is not a narrow niche: auto parts is a competitive category. We fully reworked the store, refreshed SEO for Ozon's updated algorithm and for traffic coming from AI search, and cut ad ACOS while spending about $750 per week on traffic. Margin grew, and the partner and we started earning more net profit than before.
Automatic strategies or manual management
At one client store we switched off automatic strategies in promotion a little over a month ago. Units sold grew almost 3x, while ACOS rose only 1.4 points, from 1.2% to 2.6%. Three times the sales for 1.4% more cost share looks like a good trade. My conclusion: if the goal is sales growth, use manual campaigns with controllable bids. In our tests automatic campaigns performed noticeably worse.
If you are afraid to touch the ads
The opposite situation happens too: a seller is afraid of advertising and relies on organic only. A partner came to us having never run any promotion. We launched ads alongside the rest of the store work and in under a month added close to $16,700 in sales on top of the existing volume, at 2.3% ACOS.
Where to start
- Do not touch bids blindly. Look at product card and store metrics first.
- Remove outdated and duplicate campaigns.
- Switch off tools that spend budget without sales.
- To grow, move to manual bids and test.
- Watch your stock: advertising a product that is out of stock is money wasted.
If you want an outside view of your store's ads and product cards, we audit and manage Ozon stores.
