The hardest question in digital advertising isn’t “how do we spend?” but “was what we spent worth it?” Ironically, digital ads produce a huge amount of numbers: impressions, reach, clicks, views, conversions and returns. But more numbers don’t mean a clearer answer. One platform says the campaign delivered an excellent return, another says the same about the very same orders, and the bank account says something else again.
This article offers an organised way to measure: what to measure, how to set up tracking, how to read the numbers without being fooled, and how to build a report that supports decisions.
Start from the business question, not the platform
Before choosing any metric, define what the business wants from the campaign. The right metric comes from the goal:
| Business goal | Primary metric | Supporting metrics |
|---|---|---|
| Store sales | Return on ad spend and cost per order | Average order value, conversion rate, share of new customers |
| Service enquiries | Cost per qualified lead | Share of serious leads, share that become contracts |
| Bookings | Cost per confirmed booking | Actual attendance rate |
| Awareness for a new brand | Reach within the target audience, and frequency | Completed video views, branded search |
A recurring mistake is judging a sales campaign by its clicks, or an awareness campaign by its direct sales. Every campaign should be measured by the metric it was designed for.
The core metrics and what they mean
- Impressions and reach: impressions are how many times the ad was shown; reach is how many different people saw it.
- Frequency: the average number of times a person saw the ad. Rising frequency with weak results is a sign of ad fatigue.
- Cost per thousand impressions (CPM): the price of reaching your audience. It rises in competitive seasons such as Ramadan and White Friday.
- Click-through rate (CTR): the share of people who clicked out of those who saw the ad — an indicator of the strength of the ad and its message.
- Cost per click (CPC): useful for comparison, but not a measure of success in itself.
- Conversion rate: the share of visitors who took the desired action — it reflects the quality of the page and offer more than the ad.
- Cost per acquisition (CPA): spend divided by the number of results, such as cost per order or per lead.
- Return on ad spend (ROAS): revenue from ads divided by spend. A ROAS of 4 means 4 riyals of revenue for every riyal of ad spend.
The return that actually matters: profit, not revenue
ROAS is calculated on revenue, but revenue isn’t profit. A store with a 20% margin needs a completely different return from a store with a 60% margin to be profitable. So calculate your break-even ROAS: the minimum at which you don’t lose money, after accounting for product cost, shipping, payment fees, tax and returns.
A simple example: if the average order is 300 riyals and the cost of the product, shipping and fees is 180 riyals, 120 riyals remain. That’s the most you can spend on advertising to win an order without losing money — a break-even ROAS of about 2.5. Anything below that is a loss, even if the number looks good in the platform dashboard.
Think about customer lifetime value too: if customers come back and buy several times, it may be acceptable for the first order to make no profit — provided you have real data on repeat purchase rates.
Setting up tracking: no accurate data, no measurement
Pixels and server-side APIs
Every ad platform needs its pixel on your website: Meta, TikTok, Snapchat and Google. With tracking restrictions in browsers and mobile operating systems, server-side APIs (such as Meta’s Conversions API and TikTok’s Events API) have become essential to reduce lost events. Check that each event arrives once, with the right value and currency. Meta specifics are in our Meta ads guide.
Google Analytics 4
A relatively neutral tool across platforms that brings together visits from every source. Set up your important events in it as conversions (key events): purchases, form submissions, WhatsApp clicks and calls.
UTM links
Add UTM parameters to every ad link to identify the source, medium and campaign, such as:
?utm_source=snapchat&utm_medium=paid&utm_campaign=ramadan-sale
Agree on a single lowercase naming convention with your team, or your reports will show “Snapchat”, “snapchat” and “snap” as different sources.
Tracking WhatsApp and calls
In Saudi Arabia, a large share of conversions happen on WhatsApp or by phone, not on the website. Track clicks on WhatsApp and call buttons as events, use a different opening message for each campaign or a dedicated number where possible, and ask your team to record each customer’s source in the sales system.
Why the numbers don’t match
You’ll notice that the total sales across the Meta, TikTok, Google and Snapchat dashboards is higher than your store’s actual sales. That’s normal, and here’s why:
- Every platform credits itself: someone saw a Snapchat ad, then clicked a Google ad, then bought. Both platforms may count the sale.
- Attribution windows: each platform counts a sale if it happens within a set period after a click or view, and the default settings differ.
- View-through attribution: some platforms count sales from people who saw the ad without clicking.
- Modelling: when a platform loses some data, it estimates some conversions statistically.
The practical solution: have a single source of truth for revenue — your store or sales system. Use platform dashboards to compare within each platform (which campaign or ad is better), use Google Analytics to compare across channels, and use store numbers for the final judgment on profitability.
One metric that brings the picture together: blended return
One simple and effective method: calculate total revenue (from all sources) divided by total ad spend (across all platforms) for each period. This metric, sometimes called MER, doesn’t tell you which platform is better, but it does tell you whether advertising as a whole is working — and it isn’t affected by overlapping attribution. Track it monthly and compare its movement with changes in spend.
Testing real impact
Sometimes the question isn’t “how many sales did the platform credit to itself?” but “how many sales wouldn’t have happened without the ad?” A few ways to measure that:
- Lift tests: available on Meta and others for some accounts, splitting the audience into a group that sees the ad and a group that doesn’t.
- Geographic tests: running the campaign in some cities and pausing it in similar ones, then comparing sales.
- Temporary pauses: stopping a channel for a period and watching the effect on total sales, while accounting for seasonality.
A monthly report that supports decisions
A good report is short and answers specific questions:
- The overall picture: spend, revenue or results, blended return, compared with last month and the same period last year where available.
- By channel: spend and cost per result for each platform.
- Best and worst ads: which angles and designs worked, and which to stop.
- What we learned: the results of the tests.
- Decisions for next month: where we increase, where we cut back, and what we test.
A report that shows dozens of numbers with no conclusions or decisions is no different from a screenshot of the platform dashboard.
Common measurement mistakes
- Judging a campaign after two days, before learning completes and delayed conversions arrive.
- Optimising for cheap clicks instead of the results that matter to the business.
- Adding up platform numbers and treating the total as real revenue.
- Ignoring returns and cancelled orders.
- Not counting WhatsApp and phone conversions, so the campaign looks weaker than it is.
- Comparing Ramadan with a normal month without accounting for seasonality.
Frequently asked questions
What’s a good return on ad spend?
There’s no single good number for every business. A good return is one that exceeds your break-even ROAS after accounting for product cost, shipping and fees. Calculate your own number first, then judge campaigns against it.
Which tool should I rely on: the platform dashboard or Google Analytics?
Use each for what it’s good at: the platform dashboard for optimising campaigns within it, Google Analytics for comparing channels, and your store or sales numbers for the final judgment.
How long should I wait before judging a campaign?
Leave the campaign until it exits the learning phase and gathers a reasonable number of results — usually at least a week for campaigns with enough volume. And remember that some conversions happen days after someone sees an ad.
How do I measure awareness campaigns that don’t sell directly?
By reach within the target audience, frequency and completed video views, then by indirect effects: branded search, direct visits, and the performance of retargeting campaigns during the same period.
Conclusion
Measuring ad campaigns starts with the business question and choosing the right metric for it, then setting up accurate tracking with pixels, server-side APIs, Google Analytics, UTM links and WhatsApp tracking. After that comes understanding why numbers don’t match, adopting a single source of truth, following blended return, and calculating profit rather than revenue. Good measurement doesn’t make ads cheaper, but it bases every spending decision on what’s actually happening.
If you’d like measurement set up across your ad accounts or your campaign results reviewed, explore the Facebook ads and TikTok ads services, or get in touch.
