You do not need to run PPC campaigns to hold them accountable. You just need to know which numbers to ask for.
That list changed in 2026. Google’s AI now decides where much of your budget goes, privacy rules hide a big chunk of your conversions, and AI-generated search results are reshaping how customers find you. The old dashboard favourites, clicks and cost per click, no longer tell you whether paid ads are actually growing the business.
Here are the five numbers that do, what each one tells you, and the one question to ask your team about it.
1. Profit-Based ROAS, Not Revenue ROAS
What it tells you: whether your ads make money after product costs, shipping, returns, and platform fees, not just whether they generate sales.
Why it matters now: a campaign can report a healthy 5x return and still lose money once real costs come out. Worse, Google and Meta’s automated bidding will happily scale your least profitable products if you only feed them revenue data.
Ask your team: “Is our ROAS calculated on revenue or on margin?”
2. Incrementality: Would We Have Made These Sales Anyway?
What it tells you: how many conversions your ads actually caused, versus sales that would have happened without them. Ads shown to people already searching for your brand are the classic budget leak.
Why it matters now: this became much easier to answer in 2026. Google now offers built-in incrementality experiments, and the industry standard is to treat these tests as the ground truth, with platform dashboards as a rough signal only.
Ask your team: “When did we last run an incrementality test, and what did it change?”
3. Cost per Qualified Lead, Not Cost per Lead
What it tells you: what you pay for a lead your sales team actually wants, not just a form fill.
Why it matters now: cheap leads that go nowhere are not savings, they are waste. And because Google’s AI learns from the conversions you report back, feeding it unqualified leads trains it to find more of them. Connecting your CRM back to the ad platforms fixes both problems at once.
Ask your team: “Do sales and marketing agree on what a qualified lead is, and does that data flow back into our ad accounts?”
4. How Much of Our Data Can We Actually See?
What it tells you: what share of your conversions are directly measured versus estimated. Privacy rules and cookie loss now hide roughly a quarter to a third of conversions, and Google’s mid-2026 analytics changes made consent settings decisive for what gets counted at all.
Why it matters now: two campaigns can show identical results on paper while performing very differently in reality, simply because one is set up to capture data properly and the other is not. In the Gulf, where customers move between languages, devices, and borders, this gap is even wider.
Ask your team: “What percentage of our conversions are measured versus modeled, and is that improving?”
5. Is Google’s AI Helping Us or Spending for Us?
What it tells you: how the new AI features, which now write ad copy and choose which searches your ads appear on, are performing against the campaigns you control directly.
Why it matters now: Google reports these tools deliver around 7 percent more conversions on average, but that is Google grading its own homework. Left unsupervised, the AI can pay a premium for customers who were already yours, or generate off-brand copy. For bilingual Gulf accounts, auto-generated Arabic and English ads need a human eye on quality.
Ask your team: “Are we separating AI-driven results from our core campaigns, and who is reviewing the AI-generated ads?”
The Bottom Line
Clicks, CTR, and cost per click are diagnostics. They help your team fix things, but they should never headline a report to you.
The report you want answers one question: is paid media driving profitable growth we would not have had otherwise? If your current reporting cannot answer that, the five questions above are where to start.
Sandstorm Digital is a Dubai-based performance marketing agency helping brands across the Gulf turn AI-era search into measurable growth. Talk to our PPC team about what your reports should really be telling you. [email protected]
