From 2.67× to 13.70× ROAS — without blowing the budget
We took over a Google Ads account generating €23K in revenue on €8.7K spend and turned it into €129K on €9.5K. Same product catalogue. Same market. Better decisions.
A working account that could work much harder
Prodavnica Alata is a Serbian e-commerce retailer selling tools and hardware across multiple categories. Their Google Ads account was live and generating sales — but the economics were mediocre.
When we took over the account, ROAS sat at 2.67×. That’s not a disaster, but for an e-commerce business with healthy margins and a proven product range, it left a significant amount of revenue on the table. The account had grown organically over the years — campaigns added when needed, never properly restructured.
“The campaigns were alive, but not optimised. We saw immediately that budget was flowing to the wrong places.”
Inherited complexity, diluted budget
The account had accumulated legacy campaigns running broad match keywords against categories that hadn’t been profitable in over a year. Budget was spread thin across too many ad groups, making it impossible for Google’s machine learning to gather enough signal on any single campaign.
Conversion tracking was partially broken — firing on some pages but not others. Without clean data, Smart Bidding couldn’t do its job. The Performance Max campaigns that had been set up were under-fed with assets and running generic copy that didn’t reflect the product range.
Eight focused interventions
No magic. Just methodical fixes applied in the right order.
Rebuilt tracking from scratch via Google Tag Manager. Verified every conversion action was firing correctly before touching bids.
Merged 14 underperforming campaigns into 4 consolidated structures. Less noise, more data per campaign.
Shifted budget from legacy Search campaigns to a properly-built Performance Max campaign with full asset coverage.
Rebuilt asset groups around product categories with tailored headlines, descriptions, and audience signals.
Fed first-party customer lists and website visitor segments as audience signals to accelerate PMax learning.
Implemented a weekly negative keyword process to eliminate irrelevant traffic draining budget.
Set ROAS targets informed by actual margin data rather than arbitrary goals. Let the algorithm run with confidence.
Cleaned the Merchant Center product feed — corrected titles, categories, and pricing data to improve Shopping coverage.
Same budget. 5× the return.
Comparing a 90-day period before our engagement against the equivalent period after implementation.
We deliberately cut campaigns targeting low-value micro-conversions (newsletter signups, wishlist adds) that were inflating the conversion count without contributing to revenue.
The remaining conversions are purchase events only. Fewer conversions, dramatically higher average order value — exactly the outcome we optimised for.
Real data, real account
The numbers come directly from Google Ads account exports. No modelling, no projections.
Read the Total: account row — €8,713 → €9,466 spend (+8.6%), €23,240 → €129,661 conversion value (+458%), 13.70× ROAS. It counts all account spend across both periods; the per-campaign rows above compare only currently-active campaigns. Campaign names blurred for privacy; date ranges left visible.
The account didn’t need more budget. It needed better decisions.
A +8.6% increase in spend produced a +458% increase in revenue. That’s not a rounding error — it’s what happens when tracking is clean, budget flows to the right campaigns, and the algorithm has the data it needs to do its job.
Prodavnica Alata now has a Google Ads account that compounds. Performance Max continues to learn, ROAS has stayed above 12× through the following quarter, and the business has headroom to scale spend further without diminishing returns.
How we know the numbers are real
Your ad account has more in it.
If you’re running Google Ads and suspect your budget isn’t working as hard as it should, we can find out why. We take over accounts, fix what’s broken, and scale what works.
Services that made it happen
More work like this
Browse showcases
Share
with your friends