Cost per lead versus cost per customer

Every ads report you receive shows the first number. Almost none show the second, because the platform cannot see your invoices. That gap is where good campaigns get killed and bad ones get funded.

I learned this in restaurants before I learned it here. Covers were easy to count and told you nothing on their own. A packed room at the wrong average check is a busy way to lose money. Leads are covers. Customers are the check.

How can cheaper leads make me less money?

Because lead quality and lead cost usually move in opposite directions.

Loosen your targeting, broaden your keywords, drop the qualifying language from your ads, and your cost per lead falls. More people inquire. The inquiries are worse. Some want a service you do not offer, some are outside your area, some are price shopping three bids on a job you would not take.

Your ads report looks better. Your calendar fills with estimates that go nowhere. Your close rate drops and nobody connects the two, because the close rate lives in a different system than the ad account.

A campaign optimized for cost per lead will eventually find you the cheapest possible leads, which are the ones nobody else wanted. That is not a malfunction. It is the metric doing exactly what you asked.

What does that look like in numbers?

Take a dental practice, because the published benchmarks for that category are unusually clean.

LocaliQ, which publishes annual search advertising benchmarks under the WordStream name, puts the 2026 figures for Dentists and Dental Services at $8.00 average cost per click, a 10.67% conversion rate, and $72.97 average cost per lead. Those are medians across a large campaign sample, LocaliQ sells advertising management, and the disclosed sample for the 2025 edition of the same series was 16,446 US campaigns with a minimum of 64 active campaigns per category. Use them as a shape, not as your price.

Month one, tightly targeted. Spend $3,000 at $8.00 per click and you buy 375 clicks (derived). At the 10.67% benchmark conversion rate that is 40 leads (derived), a cost per lead of $75 derived. Your front desk converts 40% of those into booked, attended, paying patients, because they were the right callers. That is 16 patients, at a cost per customer of $188 derived.

Month two, loosened up. Somebody broadens the keywords and drops the qualifiers to get the cost per lead down. It works. Cost per lead falls to $55 derived, so the same $3,000 now produces 55 leads. The callers are worse, and the close rate falls to 25%. That is 14 patients (derived), at a cost per customer of $218 derived.

Cost per lead fell 27%derived. Cost per customer rose 16%derived. You booked two fewer patients for the same money.

Every number in that walk-through is arithmetic on four inputs: spend, click cost, conversion rate, close rate. Substitute your own and it takes five minutes. If your close rate is not written down anywhere, that is the finding.

Here is the restaurant version of the same lesson, since the margins make it starker. The National Restaurant Association's cost structure work puts a typical independent restaurant at roughly 33 cents of food, 33 cents of labor and 29% of sales in other expenses, leaving a pre-tax margin near 5%. At a 5% margin, one dollar of profit requires about twenty dollars of sales. A marketing number that produces revenue without producing margin is not a small problem in that business. It is the whole problem.

Why should I be careful with the benchmark itself?

Because a national median is not your city, and it will be used against you in both directions.

The LocaliQ figures above come from campaigns across the United States. Your auction is one metro, sometimes one zip code, against a handful of specific competitors. A San Francisco dental practice and a rural one are both inside that $72.97, and neither of them is $72.97.

There is a second trap in the 2026 edition specifically. LocaliQ reports that cost per lead decreased across all industries for the first time in five years. If your cost per lead fell this year, some of that movement is the market, not your manager. A benchmark that moves underneath you is not a scoreboard. Ask whether cost per customer moved, because that number has no market-wide tailwind attached to it.

The right use of a published benchmark is as a sanity check on order of magnitude. If your dental cost per lead is $300, something is wrong. If it is $61, you are doing fine and the interesting question is what happened after the phone rang.

How do I calculate cost per customer without new software?

You need three numbers and a spreadsheet. No tooling required for the first pass.

Take a month that has fully closed out, ideally sixty to ninety days back so the sales cycle has finished. Write down total ad spend for that month. Write down leads attributed to ads. Then go into your invoices and count how many of those leads became paying customers.

Spend divided by customers is your cost per customer. Compare it to your average customer value and your gross margin on that work. If cost per customer is a small fraction of margin, the channel works. If it is close to margin, you are running a business to generate revenue for Google.

Do it per campaign if you can, not just per account. The average hides the split. It is common for one campaign to be carrying a second one that loses money quietly.

Do it per service too. A campaign for a low-ticket job and a campaign for a high-ticket job can have identical cost per lead and completely different verdicts.

Where does the close rate problem actually sit?

Often not where the owner assumes, and there is real data on the difference between two businesses running the identical campaign.

ServiceTitan's analysis of its own platform found that a typical home services shop books 42% of the calls it takes, but the gradient underneath is the story. Shops with fewer than five technicians book 24%. Shops with 25 or more book 59%. By trade in June, plumbing ran 43%, electrical 41% and HVAC 38%. That is vendor platform data rather than independent research, and it is observed booking behaviour rather than a survey.

Feed that gradient back into the arithmetic above and the point lands hard. Two shops buying identical clicks at an identical price produce cost per customer figures more than twice apart, and the difference is entirely in who answers and how. No bid adjustment closes that gap.

ServiceTitan puts the value of narrowing it plainly: for a shop with five to fourteen technicians, every 5% improvement in booking rate is worth roughly $100,000 in additional revenue, and it requires less than one extra booked call per weekday.

What do I do when the two numbers disagree?

Tighten the front, then fix the middle.

Tightening the front means writing ads that repel the wrong people on purpose. Naming a minimum job size, a price band, a specific neighborhood, or a specialty costs you clicks and raises your cost per lead. It also raises close rate. Ad copy that filters out the wrong customers is the cheapest way to move cost per customer without touching a bid.

Fixing the middle means looking at why qualified leads do not close. Sometimes the answer is speed. Sometimes it is that the estimate arrives with no reason to pick you. Two things on the website do a surprising amount of that work before a human ever speaks. Trust signals that actually change behavior shorten the decision, and for larger jobs, where financing options sit on the site changes whether the customer says yes now or thinks about it.

The order matters. Filtering first shrinks the pool to people you can actually win, which makes the middle-of-funnel work worth doing.

When does a bad cost per customer mean stop rather than tune?

When the gap is structural rather than tactical.

If your cost per customer is somewhat above target, that is a tuning problem. Negatives, geography, ad copy, landing page. There is room.

If your cost per customer is a multiple of what a customer is worth, tuning will not close that. The click price in your category is what it is, and no amount of optimization changes the auction by that much. Either the offer has to change, the average job value has to change, or the channel is wrong for you. Knowing when to pause a campaign instead of optimizing it is a skill, and most owners wait about two months too long.

Timing is the other explanation people miss. A channel that loses money in your slow season can be your best channel in your busy one, and averaging across the year hides both facts. If your demand has a shape, turning campaigns on and off with the season can fix a cost per customer problem that looked permanent.

Where this whole framework breaks down is at low volume. Four customers in a month is not a rate, it is four events, and the cost per customer you calculate from it will swing wildly for reasons that have nothing to do with your campaign. Below roughly a dozen customers a month from ads, read the number quarterly and resist acting on any single month.

What to do this week

Pick a month that closed at least sixty days ago. Pull ad spend, lead count, and the number of those leads that paid you. Three numbers, one afternoon in your invoices.

Divide spend by customers. Compare it to your gross margin on an average job. Write both numbers on the same line so you cannot look at one without the other.

Then start recording lead source at intake. One field, asked by whoever answers the phone. Without it, next quarter you will be doing this same reconstruction from memory.

If you have more than one campaign, repeat the calculation for each. The account average is the least useful version of this number.

Questions about what a customer is really costing you?

Email me at eric@seod.com with three numbers from a closed month, your ad spend, your lead count, and how many of those leads paid you. I will calculate your cost per customer, compare it against what you tell me a job is worth, and tell you whether the channel is earning its place.

I will also tell you if the honest read is that the numbers are too small to conclude anything yet. That happens often and it is worth knowing before you make a decision on it.

Or keep reading more on paid search for local businesses.