There is a public answer to “what should a lead cost”, and on its own it is close to useless. WordStream’s 2026 benchmark study of more than 13,000 search campaigns across 23 industries puts the average cost per lead in Google Ads at $66.69, with industry averages running from about $26.84 to $131.63. Same platform, same year, a five-fold spread.
What a benchmark cannot tell you is whether your number is good. Two things can: the direction your cost per lead moves once someone is genuinely managing it, and what happens to the lead after it lands. Below are nine campaigns, with the figures as they were reported at the time, and the four levers that moved them.
The benchmark is a starting line, not a target
Cross-industry averages hide the only variables that matter: how much a customer is worth to you, how long they take to decide, and how many people are bidding on the same intent. A €40 lead is expensive for a business selling a €90 product and almost free for one selling a €40,000 contract.
Use the benchmark for one thing only: to tell whether you are in a cheap market or an expensive one. Then stop looking at it. In 2026 the all-industry averages sat at a 6.64% click-through rate, an 8.18% conversion rate and a $5.42 average cost per click, and cost per lead fell year over year for the first time since 2020. None of that is a promise about your account.
Nine campaigns and what they actually paid
These are drawn from our own case studies. Where absolute figures appear they are in the currency the campaign was bought in; percentages travel across borders, absolute numbers do not.
| Campaign | Sector | What moved | Window |
|---|---|---|---|
| Skydent | Dental clinic | ROI 811%, booked-visit cost held in a 400–600 ₽ corridor, metrics up another 30% by mid-month two | ~2 months |
| Elena Clinic | Cosmetology | 151 booked consultations; lead cost 847 ₽ against a 7,607 ₽ plan; non-brand share reached 46% | 3 months |
| SPA resort | Hospitality | CPA down 40% (539 → 323 ₽) while lead volume nearly doubled | 2 months |
| Rodnik Altaya | Resort | Requests 282 → 749 a month, pre-bookings nearly tripled, cost per pre-booking down 42% | Campaign period |
| IT infrastructure provider | B2B technology | ~50 registrations in month one grew to 500–600 leads a month, inside a €30–40 CPL corridor | Months 1–4, held 3+ years |
| E-commerce retailer | Retail | Orders +39% (11,320 → 15,735) with cost per order down 22% (2,371 → 1,848 ₽) | Year over year |
| Displate | E-commerce, US affiliate | 17.5:1 ROAS against a 2:1 goal, 630% revenue growth, 10.7% average conversion rate | 3 months |
| Russkiy Buket | Retail | Profit +29% and performance-channel orders +42% with 500+ campaigns under one dashboard | Year over year |
| THE HEIGHTS Bali | Premium real estate | Warm leads from month one in a niche where the market CPL runs far higher | From month 1 |
Read the column that matters. Almost none of these campaigns won by finding a cheaper click. They won by removing waste, by separating audiences that were being billed as one, by changing where the money was spent, or simply by running long enough to learn something.
The four levers that actually moved the number
1. Audit before you add a single ad
At Skydent the work started with an analytics audit, not a campaign. Measurement was fixed first, so the baseline was real; first-month results then beat that baseline significantly and optimization added another 30% by the middle of month two. If your conversion tracking counts a form view, a bounce and a duplicate submission as three leads, every decision you make afterwards is wrong in the same direction.
2. Separate brand from non-brand, then judge them separately
A blended cost per lead is a comfortable lie. People searching your company name were going to find you anyway, and averaging them in makes an account look efficient while non-brand demand quietly does nothing. At Elena Clinic the entire point was to break that dependency, so the number that mattered was the non-brand share, which reached 46%. Cost per lead came in at 847 ₽ against a 7,607 ₽ plan, but the strategic result was that growth stopped depending on the clinic’s own name.
3. Change the channel before you change the bid
Bid tuning has a floor. Channel choice does not. The SPA resort campaign cut CPA by 40% and nearly doubled lead volume in two months by moving spend to VK Ads, where its audience actually was. Rodnik Altaya did the same thing with display, a format most performance marketers write off, and took cost per pre-booking down 42%. The channel that is unfashionable in your industry is often the one nobody is bidding against you in.
4. Give it a quarter
The IT infrastructure provider began at roughly 50 registrations in month one. By months three and four it was producing 500–600 leads a month inside a €30–40 cost-per-lead corridor, and it held there for more than three years without a single blocking problem. Judged at week six, that campaign was a failure. Most campaigns killed for poor cost per lead are killed before the data exists to judge them.
The number almost nobody reports: cost per answered lead
Here is the uncomfortable part, and it is rarely the media buyer’s fault. Industry research on lead follow-up suggests as few as 27% of leads are ever contacted at all. Response speed compounds it: contacting a lead within a minute has been measured to lift conversion by around 391%, while many businesses average close to 47 hours.
Run that arithmetic on your own account. If you pay $60 for a lead and one in three is ever called, you are not paying $60. You are paying $180 for a conversation, and the dashboard will still say $60.
Cost per lead is a marketing metric. Cost per answered lead is a business metric. Only one of them predicts revenue.
This is why engagements here start with measurement rather than creative. It is the same argument as the five analytics leaks documented on property projects: the budget is rarely the problem, the counting is.
How to set a cost-per-lead target you can defend
- Start at the deal, not the click. Take your average customer value and your gross margin. That is the total you can spend to win one customer.
- Divide by your real close rate, measured from leads that were actually contacted rather than from every form fill.
- Take a third of what is left. The full figure is a ceiling, and ceilings leave no room for a bad month.
- Set it per channel and per audience. One number for brand search, one for non-brand, one for social. Blended targets hide the channel that is carrying the others.
- Fix the review window before launch. Write down when you will judge it and what would make you stop. Deciding that in advance is what stops a working campaign being cancelled in week six.
A worked example: a €4,000 average contract at 40% margin leaves €1,600 to acquire a customer. If one qualified lead in five closes, five leads cost you a customer, so the ceiling is €320 per lead. Target €150–200 and you have room to be wrong.
Frequently asked questions
What is a good cost per lead in 2026?
The all-industry Google Ads average is around $66.69, with industry averages spanning roughly $27 to $132. A good cost per lead is any figure below one third of the margin on the customer that lead becomes, which is usually a very different number from the benchmark.
Why is my cost per lead rising?
Four causes account for most cases: more competitors bidding on the same intent, an audience that has seen the creative too many times, tracking that has started counting differently, or growth into a colder audience. The last one is not a problem, it is the price of scale, but only if you can see it separately, which requires brand and non-brand to be split.
Should I optimize for cost per lead or cost per sale?
Cost per sale, always, once there is enough volume to measure it. Cost per lead is a useful proxy in the first weeks and a dangerous one after that, because the cheapest leads are frequently the least qualified.
How long before a new campaign’s cost per lead settles?
Plan for one full quarter on a considered purchase. Our longest-running paid account went from about 50 registrations in month one to 500–600 leads a month by months three and four. Anything judged before the algorithm has conversion data is being judged on noise.
The short version
Stop asking what a lead should cost and start asking three questions: what a customer is worth, what fraction of leads gets contacted, and which direction the number is moving. In every campaign above the win came from measurement, segmentation, channel choice or patience. Never from a cheaper click.
If you want that arithmetic run against your own account, that is where our paid media and digital strategy work begins.