Between December 2025 and March 2026, a leading Gulf electronics retailer cut Meta cost per purchase 61% and grew ROAS 182%. GA4 attributed purchases rose 40% over the same period.

Then the context that makes those numbers more significant: spend was down 50% year over year.

This is not a story about pouring more budget into a working system. It is about what happens when a large-scale Meta advertiser fixes the signal layer and finds out how much budget was being wasted on users who were never going to buy.

The situation

At this scale, the easy explanation is usually the wrong one. When efficiency starts to slip on Meta, teams often look first at creative, bidding, or media strategy. Here, the real constraint sat lower in the stack.

This leading Gulf electronics retailer was already spending heavily on Meta. The challenge was not reach. It was maintaining efficiency while conversion volume and account complexity kept rising. Event tracking and attribution were not accurate enough for the amount of purchase activity happening across the account.

That created two problems at once. Meta was not getting a complete picture of which users were actually converting, and the account was not feeding the platform the full depth of first party conversion data available. The retailer and its agency partner, Keyade Middle East, identified the issue correctly: this was a data framework problem, not a media strategy problem.

What changed

Journify fixed the signal layer on Meta by activating server side conversion flows for the retailer's online events. Purchase signals started reaching Meta with much stronger coverage, better identity matching, and the technical conditions Meta needs to trust and use the data properly.

The implementation reached an Event Match Quality Score of 9.3/10, with best practices met across event coverage, deduplication, and data freshness. Purchase event coverage reached 97%, well above Meta's 75% benchmark.

That matters because Meta does not optimize around what happened in the business. It optimizes around what it can see and match back to real users. Once purchase events were flowing back with near complete coverage and high match quality, Meta had a much cleaner training set. Instead of learning from partial conversion data, it could learn from almost the full picture of who was buying and when.

What moved

Reach improved first. Cost per session dropped 55%, even though spend was down 50% year over year. That is not just cheaper traffic. It is a sign that Meta became more selective once it had better conversion feedback. The platform no longer had to cast as wide a net to find users likely to purchase.

Then intent tightened. With 97% purchase event coverage flowing back into Meta, the platform had far less reason to spend on low intent users who looked promising on the surface but rarely converted. Cleaner signal changed who entered the funnel in the first place.

That is why the efficiency gains were so large. Cost per purchase fell 61% while GA4 attributed purchases still rose 40%. The difference in magnitude between those numbers is expected. GA4 is counting conversion volume, while ROAS reflects both stronger conversion outcomes and lower costs. Both tell the same story from different angles.

Value moved last. ROAS rose 182% because Meta was making better decisions at the top of the funnel and carrying that improvement all the way through to purchase. The punchline is that this happened on half the budget. Once the signal was clean, the spend that remained went further than the full budget had before.

This was measured year over year, and there is no visibility into seasonality, creative changes, pricing, or product mix between the two periods. Even with that context, the mechanism is clear: cleaner purchase signal changed platform behavior, and the account stopped paying for users who were unlikely to buy.

The results

Measured December 10 to January 24 vs January 25 to March 1. The baseline period reflects performance under a previous CAPI provider. The comparison period reflects performance with Journify.

ROAS: +182%

Cost per purchase: -61%

GA4 attributed purchases: +40%

Cost per session: -55%

Spend: -50%

Event Match Quality: 9.3/10

Purchase event coverage: 97% vs Meta benchmark of 75%

Meta was not underperforming because the media strategy was weak. It was underperforming because it was learning from incomplete purchase signal.

At scale, better signal does not just improve reporting. It changes who the platform chooses to spend on.