Someone sends you a number and asks whether it is good. Before you can answer, you have to know how they worked it out, because the cost per lead formula is one division sitting on top of three arguments nobody has had yet.
The maths takes two seconds. Spend divided by leads. What takes longer is agreeing what counts as a lead, which costs belong in the spend, and how long a window you measure across. Change any one of those and the same account produces a different answer.
This is for you if you report on lead generation to someone else, or someone reports it to you.
The cost per lead formula in one line
Total spend divided by total leads, over the same time period.
Spend $4,000 on Google Ads in July and record 250 leads in July, and your cost per lead is $16. That is it. There is no weighting, no adjustment, no industry multiplier.
Because it is so simple, people assume the number is objective. It is not. It is a ratio between two things you defined, and most disagreements about performance are really disagreements about those definitions.
So the useful work is not learning the calculation. It is deciding, once, what goes into each half of it, then never quietly changing your mind.
Number one: what the cost per lead formula counts as a lead
Write down your definition before you look at any report.
A form submission is the obvious one. Phone calls are less obvious, and they stay invisible unless you track them. Live chat might count. A newsletter signup probably should not, unless newsletter readers reliably buy from you.
Then decide what to exclude. Duplicates from the same person. Obvious spam. Test submissions from your own team, which sound trivial until you find forty of them in a quarter.
I have never opened an account where the platform’s conversion count matched the client’s inbox. The gap is normal. A gap that keeps widening is a tracking problem, and it is the first thing I check, because it is free to fix and it makes every other number wrong.
Number two: which spend belongs in the cost per lead formula
There are two honest answers here, and you need both.
The platform figure uses ad spend alone. It is what your dashboard shows, it updates hourly, and it is the right basis for deciding which campaign to pause on a Tuesday morning.
The loaded figure adds everything else the leads cost you: management fees, landing page tools, call tracking, the software your CRM sits in. That is the number your finance team means when they ask what a lead costs.
Keep them separate and label them. Trouble starts when someone compares last quarter’s platform figure against this quarter’s loaded figure and concludes performance collapsed.
Number three: the window the cost per lead formula covers
Pick a window long enough to cover your buying cycle, then leave it alone.
Thirty days suits most lead generation accounts. Seven days is useful for spotting a sudden break, but judging performance on a week of data usually means reacting to noise. If your leads take two months to close, a monthly view of leads alone tells you about spending, not about results.
There is a subtler trap. Platforms report conversions on the date of the click, not the date of the conversion. So a lead that arrived today may be credited to an ad from nine days ago, and yesterday’s figures will keep moving for a while after you look at them.
Blended, or by channel?
Blended cost per lead divides everything by everything, organic included.
It is a fair summary for a board meeting and a poor tool for running an account. Free leads pull the average down, which makes paid channels look better than they are, and it hides the channel that is quietly wasting money.
| View | What it answers | Where it misleads |
|---|---|---|
| By channel | Which platform to fund next | Ignores overlap between channels |
| Blended | What the whole engine costs | Flatters paid, hides the weak channel |
| Loaded | What a lead really costs you | Too slow for daily decisions |
Run by-channel for decisions. Report blended upward. Know the loaded figure before you change budgets.
What good inputs look like in practice
For scale, WordStream’s 2026 study of 13,474 US search campaigns running April 2025 to March 2026 found an all-industry average of $66.69 and an 8.18% conversion rate. Ruler Analytics, working from 110 million sessions across 13 industries, put paid search conversion at 5.4% against 2.11% for paid social.
Those are sense-of-scale figures, not targets. They tell you whether you are in the right postcode, not whether you are in the right house.
My own accounts have ranged from $2.68 per lead on US auto transport to $91.18 on US automobile accessories, and both were correct for the business they served. If you want to see the working, the results from accounts I have run lay the numbers out, and there is a longer piece on what separates a good number from a bad one.
Where the cost per lead formula breaks in real accounts
Three failures account for almost every wrong figure I have had to unpick.
The first is untracked phone calls. A home services client once looked like it was paying $140 an enquiry, which was well past what the work was worth. Roughly a third of their enquiries came by phone and none were being counted. Nothing was wrong with the advertising. The measurement was broken.
The second is counting a conversion twice. If your thank-you page fires a conversion and your form also fires one, every lead is recorded as two, and your figure is exactly half what it should be. This flatters the account until someone compares it against the inbox.
The third is mixing windows. Someone pulls spend for the calendar month and leads for the last 30 days, which overlap but are not the same. The error is small, it moves every month, and it makes trends impossible to read.
None of these are strategy problems. They are plumbing, and plumbing is where I start on every audit.
What to change this week
Three things, and none of them require touching a campaign.
Write your lead definition down in one sentence and share it with whoever reads the report. Build both the platform and the loaded version of your figure, so nobody accidentally compares one against the other. Set your reporting window to 30 days and stop looking at Monday-to-Monday swings.
If your paid channels are already split and you are deciding where the next thousand goes, there is a comparison of Google and Meta for lead generation that covers the trade-off. When you want someone to sanity-check your inputs before you act on them, send me the numbers.