Every agency report we've ever been shown by a prospect has cost per lead on the first page, usually in a big number, usually going down. Impressions, reach, click-through, CPL. Fourteen slides, and not one of them says how much money the client actually banked.
Cost per lead is the worst headline metric in marketing, and the reason is simple enough that it should have killed it years ago: you can improve it by making your leads worse.
Widen the targeting. Drop the qualifying question from the form. Swap "book a call" for "download the guide". Do all of that and your CPL halves by Friday. Your agency has a great month. You have more names, more admin, more calls with people who were never going to buy, and less cash than you had in June. Nobody lied to you. The number genuinely went down.
The same trick works on almost every vanity metric in the stack. Reach goes up if you stop caring who sees it. Reply rate goes up if you ask for less. Pipeline value goes up if you let your salespeople guess at the numbers, which they will, generously, on a Friday afternoon. Any metric a supplier can improve without improving your bank balance will eventually be improved without improving your bank balance. That isn't cynicism about agencies; it's just what happens when you pay people for activity.
So what should be on page one instead? Three numbers. We've never needed more than three to work out whether a business is growing or just getting busier.
One: qualified opportunities per week. Not leads. An opportunity is a named person, in a business that fits, who has agreed to a conversation about buying and turned up to it. Weekly, not monthly, because monthly hides a bad fortnight and you'll only notice in the quarter after the one where it mattered. Write down the number that would be enough, then measure against it. For most founder-led businesses turning over £10k to £60k a month, the honest answer is somewhere between three and eight, and they're doing two.
Two: median time to first human contact. Median, not average. One enquiry answered in four days will drag your average into fiction while ten answered in six minutes look fine. This is the cheapest number in the business to move and the one most likely to be quietly awful, because it isn't anybody's job.
Three: cash collected per opportunity created. Not revenue. Not invoiced. Collected, in the bank, from opportunities that entered the pipeline in a given month, measured however long it takes them to close. This is the one that stops all the arguments, because it prices your leads properly. Twenty opportunities that pay you nothing for four months are worth less than eight that pay in three weeks, and no dashboard built around CPL will ever tell you that.
Run those three together and something useful happens: they can't be gamed against each other. Loosen your targeting and volume goes up but cash collected per opportunity falls, so you see it immediately. Cherry-pick only the perfect fits and cash per opportunity looks lovely while the weekly count collapses. You're forced to hold both, which is what actually running a business feels like.
The reason this matters more in 2026 than it did in 2021 is that lead volume has stopped being scarce. Anyone can generate names now, cheaply, at whatever quality you're willing to accept. The scarce thing is a conversation with someone who can pay you and has decided to. Optimising the abundant input while ignoring the scarce one is how good businesses spend two years getting busier and no richer.
Do this before Friday. Open your CRM and pull the last 90 days. Count the opportunities: real ones, people who showed up to a call. Divide by 13 for a weekly number. Then pull the money that's actually landed in the bank from that same cohort, and divide. You now have two of the three, and roughly ten minutes of work will tell you more than the last year of monthly reports.
Most founders who do this find one of two things. Either the weekly opportunity count is fine and the money isn't, which is a sales and offer problem and no amount of new traffic will fix it. Or the money per opportunity is healthy and there simply aren't enough of them, which is the good problem, and the one you can spend your way out of. They need completely different work, and the CPL slide can't tell them apart.
If you'd rather we did that pull with you and told you which of the two you've got, there's a 30-minute call on the site. Bring the CRM export. We'll tell you what we see, whether or not you ever work with us.
Ryan & Ali