Ask what is cost per MQL and you will get a clean formula and a lot of confident benchmark numbers, most of which are useless to you. The formula is easy. The benchmark is close to meaningless without one other figure that almost nobody quotes alongside it.
Cost per MQL is defined as: the total marketing spend in a period divided by the number of marketing qualified leads generated in that period. If you spent $20,000 and produced 100 MQLs, your cost per MQL is $200.
That number is worth tracking. Optimizing your ad account toward it is a different proposition, and it is one of the more common ways a B2B paid search program produces a lot of activity and very little revenue.
How it is calculated, and where the ambiguity enters
The arithmetic is trivial. The ambiguity is entirely in the denominator.
An MQL is whatever your company has decided an MQL is. There is no external standard. At one company it means a demo request from a named account. At another it means anyone who downloaded a whitepaper and works somewhere with more than 50 employees. At a third it means a lead score crossed 60, where the scoring model was built 18 months ago by someone who has since left.
Those three definitions produce cost per MQL figures that differ by an order of magnitude at identical spend and identical genuine demand.
This is why cross-company benchmarks for cost per MQL should be treated as weather rather than measurement. When you read that the average B2B SaaS cost per MQL is $200, you are reading an average across companies whose MQL definitions have nothing in common.
Your own trend line over time is useful. Your number against someone else’s is not.
The question you cannot answer without the second figure
Here is the part that matters, and it is the reason a good cost per MQL is unknowable in isolation.
Two companies both spend $20,000 a month and both report a $200 cost per MQL.
The first company converts 3% of MQLs into sales qualified leads. Those 100 MQLs produce 3 SQLs, which means the real cost per SQL is $6,667.
The second converts 25%. The same 100 MQLs produce 25 SQLs, at $800 each.
Identical cost per MQL. An 8x difference in what the money actually bought.
Without knowing your MQL to SQL conversion rate, cost per MQL tells you the price of a step in a process rather than the price of a customer. Companies that report improving cost per MQL quarter over quarter while pipeline stays flat are usually watching that rate quietly decline in the other direction.
Which is the thing to watch: cost per MQL falling while MQL to SQL conversion falls at the same time is not an improvement. It is usually the signature of an ad account that has gotten better at finding people who fill in forms.
Why optimizing an ad account toward it goes wrong
The formula is fine as a report. The damage happens when it becomes the target your bidding chases.
Google’s smart bidding does exactly what you configure it to do. If the conversion action in your account is a form submission, the algorithm’s job becomes finding more people who submit forms, as cheaply as possible. It is extremely good at this.
The population that submits forms most readily is not the population that buys enterprise software. It skews toward the same students, job seekers, researchers and consultants that dominate the search terms report, all of whom will trade an email address for a report without hesitation. A real buyer with budget is more cautious with their contact details, not less. This is the same dynamic that drives wasted ad spend at the query level, arriving one layer further down the funnel.
So the optimization runs in a direction nobody intended. Cost per MQL improves month over month, which reads as success, while the composition of those MQLs degrades. Sales notices first, usually about two quarters before the dashboard does.
There is a second-order effect worth naming. Once sales stops trusting marketing-sourced leads, they stop working them promptly, which drops the MQL to SQL rate further, which makes the marketing numbers look better and the pipeline look worse. That loop is hard to break with reporting alone because every individual number in it is accurate.
What to optimize toward instead
The useful target is cost per sales qualified lead, or better, cost per qualified opportunity. That is the first metric in the chain that sales agrees is real.
Getting your ad account to optimize toward it requires the platform to know which leads became opportunities, which means sending data back from your CRM into Google Ads. The mechanism is offline conversion import, and for most B2B SaaS companies running HubSpot or Salesforce it is a configuration project rather than an engineering one.
Once it is running, three things change.
Bidding chases opportunities rather than form fills, so the algorithm starts hunting a different and much smaller population.
Your search terms and campaigns get scored on what they produced downstream, which frequently reorders your view of which campaigns are working.
And the queries that convert badly on first touch but produce good customers on the third stop looking like waste, because the account can finally see them.
The honest caveat: with a 6 to 14 month sales cycle, the feedback loop is slow. You are not going to see clean optimization signal in week three. Importing the intermediate stages, such as SQL and opportunity created, rather than waiting for closed-won, is how you get a usable signal inside a reasonable window.
What to actually track
Keep cost per MQL. It is a fine operational number and it tells you whether the top of the funnel is functioning.
Track it next to MQL to SQL conversion rate, always, on the same slide. Neither number means anything alone and together they tell you almost everything.
Then add cost per SQL as the number you optimize toward and the number that goes to the board, since it is the first one in the chain that survives contact with sales.
If you can only put one number in front of an executive, cost per SQL is the one that will not embarrass you six months later.
What a qualified conversation costs starts with what a click costs, and that varies enormously by category. I broke that down across 18 B2B software categories and 15 cybersecurity categories.
The short version
Cost per MQL is your spend divided by your MQLs, where MQL means whatever your company decided it means. Cross-company benchmarks are close to useless because the definitions do not match. Your own trend is useful, but only when read alongside your MQL to SQL conversion rate.
The failure mode is quiet: configure your ad account to optimize toward form fills and it will find you people who fill in forms, cheaply and in volume, most of whom will never buy. Connecting your CRM back into the ad account is what changes the target from activity to pipeline.
The fastest way to find out what your account is optimizing toward is to look at its conversion action, which takes about a minute. If it says form submission and you want to know what that has cost you, book a BADASS Discovery Call at bad2badass.com.