Cost per lead is a result, not a cause. It comes from three numbers multiplied together: what it costs to show your ad (CPM), how many people who see it click (click rate), and how many who click go on to enquire (conversion rate). When the cost rises, one or more of these has moved. Finding which one is most of the diagnosis.
An illustration, with invented round numbers. Suppose a thousand impressions cost a hundred rupees. If one in a hundred clicks, that is ten clicks, so a click costs ten rupees. If one in ten clicks becomes a lead, a lead costs a hundred rupees. Now the click rate halves. Nothing else changed, and a lead costs two hundred. You would have called it "Facebook getting expensive". It was the ad tiring.
| If this number moved | It suggests | What to try |
|---|---|---|
| CPM up, others steady | More competition, a seasonal bid rise, or a narrow audience | Widen the audience, or accept a seasonal cost and plan for it |
| Click rate down | Creative fatigue, or an offer that no longer draws attention | New creative, a fresh angle |
| Conversion rate down | The page, form or offer, or a tracking problem | Check the page and form on a phone, check tracking |
| Everything looks normal but leads are fewer | Events not being recorded | Test tracking, compare with actual enquiries |
Seasons and events
Costs often rise in periods when many advertisers compete for the same audience: festival seasons, sales events and the weeks before large holidays. If your costs rise on a predictable calendar and then fall back, the cause is the market, and the answer is to budget for it, not to rebuild the account.
When you have run out of audience
A small audience with a steady budget means the same people see the ad again and again. Frequency rises, and both click rate and cost move the wrong way. That is a saturation problem, covered in when ad frequency is too high.
Do not skip the boring check
Costs can appear to rise when the tracking loses events. A change in the site, in a form or in the consent banner can cut the number of conversions recorded while the real enquiries continue. Compare the platform’s numbers with your own records for the same week before deciding the ads have gone wrong.
Compare with the same period last year
If you have older data, compare this month with the same month a year ago and not with last month. Many businesses have seasonal swings in demand and in the cost of reaching people. What looks like a jump may be the usual pattern. If last year showed the same rise and fall, the answer is planning, not repair.
The audience size problem
An audience that is small relative to the budget gets used up. Reach flattens, frequency rises and costs follow. This happens most with narrow local audiences and with retargeting groups. If your target area is small, consider whether you can widen the geography, the age range or the interests without harming relevance, or whether you should accept a lower budget.
Sudden budget changes
Large increases can push the account into more expensive parts of the auction and unsettle delivery. If you want to scale, increase budgets in steps and watch for a few days between steps. If you cut a budget sharply, delivery can also be disturbed.
A short list to rule out false alarms
- Did your tracking change, or a tag stop firing?
- Did the site slow down, or a form break?
- Did the ad get rejected or limited, so a weaker one took over?
- Did the currency, payment method or account status change?
What not to do
Do not change the budget, the audience and the creative in one afternoon. If the cost then improves, you will not know why, and if it worsens you will not know what to undo. Change one thing, and give it enough time. Our creative testing approach follows the same rule.