All case studies

How we got a Serbian tools ecommerce store to a steady 30+ ROAS.

Client
Confidential
Industry
Tools & hardware ecommerce
Channels
Meta & Google Ads
Market
Serbia
Engagement
Ongoing partnership

This client, a large ecommerce retailer selling tools and hardware in Serbia, had no real account structure and no creative testing behind its ads. We rebuilt the account, introduced net new creative concepts, and scaled spend gradually rather than all at once.

30+
ROASconsistent, month after month
8,000+
SKUs in catalogkept live across every channel
Rebuilt
Account structurefrom no real structure
Meta Ads Manager, 1 April to 31 July 2026, campaign names blurred, showing a results ROAS between 40 and 54 on each active campaign
Meta Ads, April to July 2026. Campaign names hidden at the client's request.

The starting point

This client is a large ecommerce retailer selling tools and hardware in Serbia, with a catalog of more than 8,000 SKUs. When we took over, there was no real account structure and no creative testing: ads were running, but nobody was learning anything from them.

Serbia's ad market has comparatively cheap CPMs, and this store's average order value sits on the higher side for the category. Both of those work in an advertiser's favour, but neither one matters if the account underneath isn't built to take advantage of it, and this one wasn't.

What we did

  1. Rebuilt the account structure from scratch

    There was no structure to speak of, so we set one up from the ground up, organised so results could actually be tested and measured instead of guessed at.

  2. Introduced net new creative concepts

    Instead of running variations of the same handful of ads, we introduced genuinely new creative concepts to test, which is what surfaced the angles that ended up working.

  3. Worked closely with the client's development team on the catalog

    With more than 8,000 SKUs feeding the account, catalog accuracy is its own job. We worked directly with their dev team to keep product data, pricing, and stock current across every channel, so ads were never running against a catalog that had already gone stale.

  4. Scaled spend slowly, not all at once

    Once winning combinations were clear, we increased budget gradually rather than jumping straight to a higher spend level, so performance had room to prove itself at each step before moving to the next.

The results

ROAS has held above 30 consistently, not as a single strong month but as the ongoing norm for the account. Serbia's cheap CPMs and the store's higher-than-average order value both help, but a well-tested account is what turns favourable conditions into a number this high instead of an average one.

Not every account can promise a number like this. Serbia's market economics happen to make it possible here, and the account is now built to actually capture that instead of leaving it on the table.

The lesson
A 30+ ROAS invites the assumption that the media buying did all of it. Some of it came from the market: cheap CPMs and a high AOV were true before we touched the account. What wasn't there yet was a structure built to actually use those conditions, and that's the part that turned a favourable market into a number this consistent.

Want a write-up like this with your name on it?

Send us the spend and the number it has to return. What comes back is the work we would run and what it costs.