Meta Ads provides a large amount of performance data. The challenge is knowing which numbers actually matter, and what to do when they're telling you something. This guide focuses on the metrics that genuinely indicate whether your campaigns are working, with guidance on what good looks like for Saudi businesses.
Define Your Goal First
The metrics that matter depend on your campaign objective. An awareness campaign should be measured by reach and video views, not conversions. A lead generation campaign should be measured by cost per lead, not impressions. Starting with clarity about what you're trying to achieve prevents you from optimizing for the wrong things.
Key Metrics by Goal
For Awareness Campaigns
- Reach: How many unique people saw your ad
- Impressions: Total number of times your ad was shown
- Frequency: Average times each person saw your ad (keep below 3–4)
- Video views and video view rate: What percentage watched past 3 seconds, 25%, 50%, 75%
For Traffic Campaigns
- Link clicks: Total clicks to your website or WhatsApp
- CTR (link click-through rate): Percentage of impressions that resulted in a click, above 1% is generally good for cold audiences in Saudi Arabia
- CPC (cost per click): How much each click costs
- Landing page views: How many people actually loaded your destination page
For Conversion Campaigns
- Conversions: Number of target actions completed (WhatsApp clicks, form fills, purchases)
- Cost per conversion (CPA): The most important metric, how much does each lead or sale cost? Compare this to your average customer value to determine profitability
- ROAS (Return on Ad Spend): For e-commerce, revenue generated per riyal spent on ads
- Conversion rate: What percentage of ad clicks lead to a conversion
The Relevance Score Equivalent
Meta provides ad relevance diagnostics, three ratings that assess quality ranking, engagement rate ranking, and conversion rate ranking compared to similar ads targeting similar audiences. Low ratings on any of these signal specific problems: poor creative quality, audience mismatch, or weak landing page experience.
How to Use Data to Optimize
Review campaign data at minimum weekly. When a metric is underperforming:
- Low CTR → Test new creative or improve your offer messaging
- High CTR but low conversions → Fix the landing page or WhatsApp follow-up
- High CPA → Test new audiences, improve creative, or review offer competitiveness
- High frequency → Refresh creative or expand the audience
Data-driven optimization separates consistently profitable Meta campaigns from those that burn budget without results. For the full Meta Ads overview, see How Facebook and Instagram Ads Work.
The Metrics That Decide Anything
Cost per result and total results are the two numbers that determine whether a campaign continues. Reach, impressions and engagement describe delivery and rarely change a decision.
Beneath those, conversion rate tells you where a problem sits. Healthy clicks with poor conversion is a page or offer problem; poor clicks with healthy delivery is a creative problem. Each sends you somewhere different.
Frequency belongs in the same weekly view as an early warning. It rises before performance falls, which makes it the most actionable number in the report.
Why Platform Attribution Overstates Results
Meta counts a conversion when someone saw or clicked an ad within its attribution window and later converted, whether or not the ad caused it. For retargeting especially, many of those people were returning anyway.
This is not dishonesty so much as a structural bias. Every advertising platform reports on the assumption that it was responsible, and running several platforms means the same conversion gets claimed more than once.
The correction is to compare platform-reported results against total business enquiries. If the platform reports forty leads and the business received twenty-five enquiries in total, the reported figure is describing something other than new customers.
The Holdout Test
The most reliable way to know whether a campaign produces incremental business is to stop it for a fortnight and watch total enquiries. Nothing inside the dashboard answers this question.
Run the test in a normal period rather than during Ramadan or a holiday, and change nothing else while it runs. A single confounded test is worse than none, because it produces a confident wrong answer.
Most advertisers find their campaigns are genuinely productive but less so than reported. Knowing the real figure prevents the common error of cutting acquisition to fund retargeting that was harvesting its results.
Building a Report Worth Reading
One page, four numbers: total enquiries, cost per enquiry, enquiries that became customers, and cost per customer. Everything else is diagnostic detail that belongs underneath.
Track them monthly against the same months last year rather than against the previous month, so seasonality does not masquerade as performance.
Include what changed and when. A report showing results without decisions makes it impossible to learn anything from a good month.
Tracking Enquiries Through to Revenue
The gap most businesses never close is between a lead and a customer. Campaigns get judged on cost per lead, and two campaigns producing leads at the same cost can differ enormously in what those leads are worth.
Closing it requires nothing sophisticated: record where each enquiry came from and whether it became a customer. A spreadsheet updated weekly is enough for most small businesses.
Within two or three months that record usually reorders the account. Campaigns that looked expensive on cost per lead frequently turn out to be the cheapest sources of actual revenue.
What to Do With a Campaign That Is Not Working
Diagnose before rebuilding. Poor delivery, poor click-through, poor conversion and poor lead quality are four different failures with four different fixes, and the reports distinguish them clearly.
Change one thing and give it a week. The instinct when results are bad is to change everything at once, which produces a new result and no understanding of what caused it.
Know when to stop. A campaign that has been genuinely tested across creative, audience and destination without producing viable results is telling you something about the offer or the market, and continuing to spend rarely changes the answer.
Reporting to Someone Who Is Not You
A report for an owner or a board needs three things: what it cost, what it produced, and what happens next. Screenshots of platform dashboards answer none of them and invite questions about metrics that do not matter.
Translate everything into business language. Not cost per result but cost per customer; not reach but how many people made contact; not frequency but whether the audience is being over-exposed.
Include the uncertainty honestly. Saying that platform attribution likely overstates results by some margin builds more confidence than a precise figure that later fails a holdout test.
Comparing Meta Against Your Other Channels
Meta reports on its own terms, as does Google, and both will claim the same customer. Comparing them on their own numbers produces a total larger than the business received.
Compare on cost per customer using your own records rather than platform reports. It is the only comparison that survives different attribution models.
Where the numbers disagree sharply, the business record is the one to trust. Platforms measure their contribution; only you can see the whole picture.
