Top TechActivity metrics versus outcome metrics
Start by sorting what you look at into two groups:
- Describes activity: impressions, reach, views, clicks, followers, click-through rate. Useful for diagnosis, not for judgement.
- Describes outcome: qualified leads, cost per lead, cost per acquisition, return on ad spend, revenue attributed to the campaign.
The simple test: could this number rise while your sales stay flat? If yes, it is an activity metric. Views can double without a single extra riyal of revenue; attributed revenue cannot.
The numbers worth tracking
- 1
Cost per lead (CPL)
What you paid per enquiry. Useful, but genuinely misleading on its own — it says nothing about the quality of that enquiry.
- 2
Cost per acquisition (CAC)
What you paid per actual paying customer. This is the number that matters, and it is always considerably higher than cost per lead.
- 3
Return on ad spend (ROAS)
Revenue per unit of spend. Best suited to e-commerce, where the purchase is immediate and trackable.
- 4
Conversion rate
What share of arrivals completed the action. If it falls while clicks hold steady, the problem is after the click rather than before it.
- 5
Lead quality
What share of leads your sales team could actually reach and qualify. The one number no platform provides — it has to come from your team.
The cheap-lead trap
The most common scenario we see: a campaign with an excellent cost per lead and a demoralised sales team, because nobody answers the phone.
What usually happened is that the campaign was optimised toward the wrong objective. Ask a platform for the most form fills at the lowest cost and it will do exactly that — finding the people quickest to fill in a form, who are not necessarily the people most serious about buying.
The fix is rarely different targeting. It is changing what you ask the campaign for: optimise toward a deeper conversion (a completed call, a booked appointment, a purchase) rather than the form, and add a light qualification step that filters out the uncommitted.
Before judging: is your measurement even correct?
Before concluding a campaign has failed, verify that what you are measuring is right. In a large share of the accounts we have audited, the first real improvement came from fixing measurement rather than changing targeting.
- Is the pixel actually firing? Check that events register on the correct pages.
- Are conversions being double-counted? A duplicate installation or an event firing twice makes numbers look better than they are.
- Does attribution add up? If every platform claims the credit, the total will exceed your real sales.
- Are phone and message sales counted? Many conversions in Saudi Arabia and Egypt end on WhatsApp or a call — invisible in the dashboard unless connected.
- Are you comparing like periods? A month containing a season or a holiday does not compare to an ordinary one.
What happens after the click
An excellent ad pointing at a weak page wastes most of what it earned. Before judging the campaign, inspect what the visitor sees after clicking:
- Page load speed, on mobile specifically
- Does the page deliver exactly what the ad promised, or talk about something else?
- How many form fields — every extra field loses a share of applicants
- Is the requested action obvious, or buried below the fold?
- How long does your team take to respond — an hour or three days?
How long before judging?
Judging after a few days misleads; waiting months without intervening wastes money. The more useful criterion is not time but volume of data: has the campaign gathered enough conversions for differences to be meaningful, or could the gap between two ads be chance?
In practice: the opening days are the platform learning, and are not a basis for judgement. Then a period of gathering sufficient data. Then a decision based on repeated patterns rather than one good day.
And changing a campaign daily prevents it from stabilising at all — a more common error than excessive patience.
Questions
Frequently asked
What is a good return on ad spend?
There is no universally correct figure, and anyone who gives you one is guessing. An acceptable return depends on your margin: a product at 70% margin tolerates a far lower return than one at 15%. Calculate your own break-even first — the return at which you cover product and advertising cost — then set your target clearly above it.
The platform says the campaign works and sales have not moved. Why?
Three common causes, in order: broken measurement counting conversions that did not happen; misattribution, where the platform claims credit for sales that would have happened anyway; or poor lead quality, where the journey stops at your sales team. Always start by verifying measurement.
Is an ad that does not sell immediately a failure?
Not necessarily — it depends on the buying cycle. High-value products and business services are not bought from a single ad. But that is not an open excuse: if a campaign is not selling, it should still show progress on a nearer metric — qualified enquiries, booked appointments, or returning visits.
Should we stop a weak campaign or fix it?
Start by diagnosing where it breaks. Poor reach is a targeting problem; good reach with poor clicks is a creative problem; good clicks with poor conversion is a landing page or offer problem. Stopping before diagnosing loses what you learned. More detail on the media buying page.
