Look at the metrics that have survived a decade of monthly reports. Impressions. Reach. Followers. Total engagements. Video views.
Notice the thing they share.
They only move in one direction. Followers are cumulative by construction. Total impressions this quarter can’t be lower than total impressions last quarter unless somebody genuinely stopped working. Reach across a longer window always beats reach across a shorter one.
That’s not an accident and it isn’t really dishonesty either. It’s selection pressure. Numbers that can deliver bad news get questioned, and questioned numbers get quietly dropped from the deck. Numbers that structurally cannot go down survive forever, because nobody ever has to sit in a room and explain them.
A metric that cannot deliver bad news cannot inform a decision. That’s the whole test, and it disqualifies most of what gets reported.
Here are four that can. The most important one isn’t a marketing metric at all.
Cost per qualified lead, with the argument about “qualified” settled first
Cost per lead is nearly useless on its own, because the cheapest way to lower it is to accept worse leads. Every agency knows this and some of them use it.
So the number is cost per qualified lead, and the word doing the work is qualified. Before this metric means anything, somebody has to write down what qualifies a lead, and that somebody should be whoever owns revenue, not whoever owns the campaign.
The definition doesn’t need to be clever. Right company size, real budget, a problem you actually solve, and they took a call. What it needs to be is written down and stable, because a definition that drifts turns the metric back into decoration.
Lead to customer rate
Cost per qualified lead tells you what the top of the funnel costs. This tells you whether the funnel works.
The useful version is segmented by channel, because a blended rate hides the whole story. Paid search leads converting at 8% and social leads converting at 1% is a completely different business from both converting at 4%, and the blended average is identical.
If you track one thing you don’t currently track, make it this one, split by source.
Payback period, which is the one that should lead the report
This is a finance number that got left out of marketing reporting almost everywhere, and it’s the number that decides whether you have a business or an expensive hobby.
Payback period is how many months of gross profit from a customer it takes to earn back what you spent acquiring them. Not revenue. Gross profit, so after cost of delivery.
The benchmarks are reasonably settled. Under 12 months is strong. Twelve to eighteen is fine. Eighteen to twenty-four should worry you. Past twenty-four is a serious problem unless you’re selling six-figure enterprise contracts, where it’s normal. For SMB-focused businesses the target is tighter, somewhere around six to nine months, because SMB customers churn faster and you have less time to make the money back.
Why does this one outrank the others?
Because it sets the speed limit on your growth.
If your payback is eighteen months, every dollar you put into acquisition is locked up for eighteen months before it can be spent again. You can only recycle cash as fast as it comes back. Two companies with identical CAC and identical revenue, one at eight months and one at twenty, are not in the same business. The first can compound. The second needs outside money to grow at all, and in this funding climate that’s a much less comfortable position than it was three years ago.
Nobody puts this in a marketing report because it requires the finance side to hand over gross margin, and those two teams often aren’t speaking. Make them speak.
Share of revenue by channel
The last one is about fragility rather than efficiency.
If 70% of your revenue arrives through one channel, you don’t have a marketing strategy, you have a dependency. Google’s algorithm, Meta’s auction, or one referral partner’s goodwill is holding up your quarter, and you’ll find out how much on the day it moves.
Track it quarterly, and treat concentration as a risk to manage rather than a success to celebrate. Diversifying costs efficiency in the short term. That’s the premium you pay for not being one platform update away from a bad year.
All four are getting harder, and it isn’t your fault
Here’s the context that changes how you should read your own numbers.
Customer acquisition cost has risen roughly 60% over the past five years, and something like 222% over the past eight. That’s structural, not a failure of execution. Signal loss from privacy changes, auction inflation as more spend chases the same inventory, and conversion rates drifting down across the board.
The 2026 figures the benchmark trackers are publishing put median Meta cost per acquisition around $38, with CPMs up about 20% year over year. Google CPCs rose roughly 13%, with median cost per acquisition near $24. Ecommerce retail saw paid CAC jump about 16% in a year, the worst of any sector.
Two things follow from that.
First, a flat CAC is now a win. If your cost per acquisition is the same as last year, you outperformed the market, and a report that presents that as stagnation is reading it wrong.
Second, the businesses holding up best are the diversified ones. Companies leaning primarily on paid channels have seen CAC climb roughly twice as fast as those with a broader mix. Referral in particular pulls blended CAC down meaningfully within the first year, which makes it the highest-leverage channel most small companies still don’t run deliberately.
What I’d actually change
Delete the top four rows of your marketing report and replace them with these four. Not add. Replace, because the decorative numbers survive by taking up space that a difficult number would otherwise occupy.
Then hold the first meeting where somebody has to explain a number that went down. That meeting is the entire point. Everything before it was theater.
I’ve written separately about what your agency genuinely can’t see in their own reporting, which is a different problem from this one and worth reading alongside it. This post assumes the numbers are honest. That post covers what happens when the platforms themselves stopped showing them.
If you want to think through which four numbers fit your business, drop me a line on email, WhatsApp or LinkedIn and we can have a quick chat. I’m contracted full time so this isn’t a pitch. My notes on how I work go into the longer version.