Performance marketing, measured for real.

Why do ad platforms over-report results?

Each ad platform counts the conversions it touched, under its own rules. When a customer sees an Instagram ad, clicks a Google ad and then buys, both can claim the sale. Add the dashboards together and you often get more sales than you made. A campaign reporting 6× might be closer to 2× in sales it actually caused.

Quick test: if the conversions your platforms report add up to more than about 15% above your real orders, you’re double counting.

How do you measure what ads really add?

Branded search is the usual first test because the evidence is strong: eBay’s experiments found brand-keyword ads had no measurable short-term benefit, published as NBER Working Paper 20171.

“…brand-keyword ads have no measurable short-term benefits.”

Blake, Nosko & Tadelis, NBER Working Paper 20171 (eBay field experiments)

1. Dashboard reconciliation: platform conversions against back-end orders or CRM revenue, every month.
2. Branded search pause: brand-name ads switched off for two weeks to see how much revenue they really add.
3. Geo holdouts: paid media reduced or paused in one comparable region for a few weeks, comparing total revenue. See incrementality testing.

What replaces broken ROAS

• Server-side tracking: Meta Conversions API and Google Enhanced Conversions, deduplicated and set up under consent mode.
• First-party data: GA4 configured properly, and a CRM that links leads to revenue.
• Media mix modelling for accounts with enough spend and history, typically from around AED 100,000 a month.

What happens in the first 90 days?

• Days 1–30, audit and baseline: tracking validated, dashboards reconciled, spend audited, branded search isolated, incremental baseline documented.
• Days 30–60, restructure: non-incremental spend cut, audiences rebuilt, server-side tracking live, creative tests designed to reach significance.
• Days 60–90, scale what’s proven: validated channels scaled, first holdout run, reporting cadence and decision log in place.

What changes by budget size?

Reach is rarely the constraint in the UAE: Facebook’s ad reach was 9.70 million people (85.0% of the population) and YouTube’s 8.37 million in late 2025, according to DataReportal. Measurement is.

• Under AED 50,000 a month: tracking, focus on one or two channels, reconciliation and branded-search tests.
• AED 50,000–500,000: geo holdouts become practical and server-side tracking pays for itself.
• Above AED 500,000: media mix modelling with regular recalibration and several tests running in parallel.

Previous-employer results, labelled as such: in-house at a Dubai curtains, blinds and interiors company, the paid account reached a 5.2× ROAS with acquisition cost down 30%, judged on booked measurements and orders. Read the ROAS that lies for the method.