ROAS above 9 on cold campaigns for a Serbian omnichannel pet-supplies retailer

Challenge

The retailer was spending on Meta without trusting the numbers coming back. Add-to-carts and purchases were being missed on the website, and the pixel reported revenue in the wrong currency, so nobody could say which campaigns were paying for themselves. Cold campaigns took the hit: with no reliable ROAS the team kept budget on warm audiences and left the top of the funnel alone. They wanted measurement they could scale on, and someone to run the buying against it.

• Add-to-cart and purchase events captured incompletely on the website
• A currency mismatch with the Meta pixel (RSD against EUR) made reported revenue wrong
• Little visibility into campaign performance, cold campaigns worst of all
• No confidence to scale ad spend, because the ROAS could not be trusted

‍

Solution

We fixed the measurement first. We installed a server-side tracking gateway on the retailer’s store, so conversions are sent from their own server instead of being left to the browser, and we wrote the currency logic the platform could not handle on its own. With the data straight, we took over the daily buying on Meta and optimised against what the store actually banked.

What we set up and what we run:

• Server-side tracking, installed by us: every add-to-cart, purchase and key conversion event captured and sent server-side. No work for the client’s developers.
• RSD-to-EUR conversion logic: revenue reported in one currency, so the platform reporting and the retailer’s own accounts finally agree.
• End-to-end verification: we test the data flow from the store through to the platform, so every action is attributed to the campaign that earned it.
• Daily optimisation: we monitor ad performance daily, weekends included, and move budget onto what is converting.
• Cold-traffic scaling: with a ROAS the team can rely on, prospecting budget goes up in measured steps instead of guesses.
• Reporting they read: a weekly summary of spend, CPM, CPC, purchases, cost per purchase and ROAS, next to what we changed and why.

‍

Results

Within the first month of the new tracking going live the retailer had accurate reporting for the first time, and a clear read on which campaigns deserved more budget.

• 25% more purchases matched to campaigns, because conversions the browser had been dropping are now captured server-side.
• 22% lower cost per purchase on the same campaigns, once buying decisions ran on the corrected data.
• ROAS above 9 on cold campaigns, which turned prospecting from a risk into where the growth comes from.
• One team instead of two: tracking and media buying handled by the same people, so the separate tracking contractor came off the budget — several thousand euros a year.