CPL and CPA answer different questions, so a lead generation campaign needs both. CPL is what one lead cost, spend divided by leads, and it is the only number that can compare two creatives or two landing pages fairly. CPA is what one customer cost, and it is the only number that says whether the campaign can keep spending.
The two numbers, defined
CPL is the amount of advertising spend required to obtain one lead:
CPL = ad spend ÷ number of leads
If a campaign spends €1,000 and produces 80 leads, the CPL is €12.50. CPA is the downstream cost of a customer, so its denominator sits further down the funnel and its numerator may include more than media. The CPL entry keeps the definition short on purpose, because the hard part is not the formula.
That hard part is which spend and which leads go into it. Campaign spend is an average across every ad inside the campaign, while spend per ad_id is the granular number and the one you can act on. On the lead side, a lead counted by the ad platform and a lead counted by your own tracking are two different numbers, and they never match exactly.
Where each number belongs
| Decision | Metric | Where the numbers come from |
|---|---|---|
| Which creative to keep | CPL per ad, per creative × landing page | spend per ad_id and your first-party lead count |
| Which landing page to send traffic to | CPL per landing page, from the same ad | your own form submissions |
| Whether the offer attracts the right people | cost per qualified lead | your CRM stages, sent back to Meta as events |
| How far to scale | CPA against what a customer is worth | closed deals in your CRM |
| Whether the account is healthy | both, read as a trend | a weekly view, not a single day |
CPL is a delivery metric: it tells you whether the ad did its job. CPA is a business metric: it tells you whether the job was worth doing. An account managed only on CPL will happily buy cheap leads that never answer the phone.
Why CPL alone cannot settle anything
A falling CPL can mean three different things, and only one of them is good. The form got easier, which raises volume and often lowers quality. The audience got cheaper, which may mean it is further from buying. Or the count changed, because the attribution window moved and the same leads are now credited differently.
Meta’s own reporting carries this ambiguity. The Results column follows your attribution setting, so changing the window changes the count without changing a single lead. A custom conversion built on the Lead event is a subset of leads Meta already counted, so adding the two columns together double counts every one of them. The comparison between custom conversions and standard events covers how those subsets are built and where they break.
The fix is not to distrust CPL. It is to fix the denominator: count leads where you can check them, one lead per submission, and read CPL per ad instead of per campaign.
Where CPA gets complicated
CPA needs a funnel, not a form. Three things have to be true before the number means anything:
- Every lead keeps Meta’s lead ID and its ad ID.
- Your CRM records each stage change, and the qualification decision belongs to a person, not to a rule you forgot about.
- Closed deals carry a value, so the denominator is comparable across months.
Send those stages back to Meta and the same data does two jobs. It feeds the business number you report, and it gives Meta’s optimizer a better event to learn from: Meta documents a Conversions API for CRM integration as the route for the qualified leads goal, which is not available for new campaigns without it since April 2026.
The step that catches people is the match. Without Meta’s lead ID, stage events still go through on a hashed email or phone, and the match rate is worse than the one your CRM reports. The setup is in how to set up conversion leads optimization.
A reading order that avoids the usual mistakes
- Count leads first-party, from the submissions your own server received. That is the reproducible number.
- Keep the platform’s count too, and send it back to Meta through the Conversions API so delivery keeps learning.
- Compare creatives on CPL per ad, only after each has spent enough to judge.
- Compare offers on cost per qualified lead, after the same delay for every lead.
- Compare months on CPA, with the deal value included.
- Check the two numbers agree on direction. When they disagree, the funnel changed, not the ads.
The order matters because each step is more expensive to measure than the one before it, and a fault at the top makes every number below it misleading. A routing bug that drops leads shows up as a rising CPA long after it started.
When to use each one
Use CPL whenever you are choosing between ads, creatives, audiences or landing pages. Use cost per qualified lead whenever you are choosing between offers, forms and follow-up speeds. Use CPA when you decide a budget, and keep a second version of it that includes the sales cost, because a cheap customer that takes three hours of calls is not cheap.
One rule saves arguments: never judge a creative on a number it cannot move. A creative changes who clicks and who fills in the form. It does not change how fast the sales team calls back, and that is why a creative can look bad in a month when the follow-up was slow. Compare booking rates between leads called within the same delay before you cut an ad.
DropCPL keeps the denominator honest: it counts leads on the domain that serves the landing page, stores the ad and the landing page on each one, and syncs spend per ad from the Meta Ads API, so CPL per creative × landing page comes from leads you actually received. For how the number behaves across an account, the questions on cost per lead on Meta cover what moves it and what only looks like it does, and CBO vs ABO explains why a campaign-level average hides the pairs that decide the result.