Twenty metrics is more than any team should report on. The useful version is a small set that constrains each other, plus knowing which one you are allowed to sacrifice.
Tracking and reporting are different activities
Most metrics lists conflate two things. There is the set you should be able to produce when asked, which can reasonably run to twenty or more. And there is the set you report on regularly, which should be small enough that a change in any of them prompts an actual decision.
A dashboard with twenty numbers on it is not twenty times more informative than one with five. It is usually less, because attention spreads evenly across the display and nothing on it is clearly the thing that matters. The discipline is choosing which few sit on the front page.
The four that constrain each other
Most of what matters is contained in a small set that cannot be optimised independently. Push one and the others move.
| Metric | What it answers | What it trades against |
|---|---|---|
| CAC | Can we afford to acquire | Falls when you narrow targeting, which caps volume |
| LTV | How much an acquisition is worth | Rises with retention, which is rarely marketing’s remit |
| Payback period | How fast the money comes back | Shortens with worse LTV if you chase quick-converting segments |
| Conversion rate | How efficiently traffic becomes revenue | Rises when you cut traffic, which reduces volume |
The relationship worth internalising is between CAC and LTV, because it sets the ceiling on everything else. If lifetime value is not comfortably above acquisition cost, no channel optimisation fixes the business, and a growth programme built on top of that gap makes the problem arrive faster.
Payback period deserves more attention than it gets. Two businesses with identical ratios behave completely differently if one recovers its acquisition cost in three months and the other in eighteen, because the second needs far more working capital to grow at the same rate.
Which of these is worth attacking first depends on where the funnel is binding, covered in the full funnel piece.
For modelling how a change in one of these moves the others, the SEO ROI calculator is a starting point.
The rest, and what each is actually for
- Acquisition: cost per click, click-through rate, cost per lead, traffic growth. These are diagnostic, not goals. Useful for finding where a change happened, misleading as targets.
- Conversion: step conversions between funnel stages, lead-to-customer rate, revenue per visitor. The step rates matter more than the blended figure.
- Revenue: average order value, return on ad spend, marketing return on investment. Return on ad spend is the most commonly misread of these, because it ignores margin.
- Retention: churn, retention rate, repeat purchase rate, expansion revenue. This is where the largest gains usually sit and where marketing usually has least authority.
The distinction between diagnostic and goal metrics does most of the work. Cost per click is a fine thing to know and a terrible thing to target, because the cheapest clicks are usually the worst ones and a team rewarded for lowering it will deliver exactly that.
Ratios lie in a specific and predictable way
Every ratio can be improved by changing either half, and roughly half the time the easier change is the one you did not intend. Conversion rate rises when you cut traffic. Return on ad spend rises when you stop scaling. Cost per acquisition falls when you narrow targeting to people who were going to buy anyway.
None of those is cheating, and each is sometimes correct. The failure is not noticing which half moved. Any ratio on a front-page dashboard should carry its numerator and denominator next to it, or someone will eventually improve the ratio while making the business smaller and be rewarded for it.
The clearest worked example of this is conversion rate, taken apart in the denominator problem.
Choosing what goes on the front page
The test for whether a metric earns a place: name the decision that changes if it moves ten percent in either direction. If there is no such decision, the metric belongs in the set you can produce on request rather than the set you look at weekly.
For most B2B companies the front page ends up being four or five: pipeline created, cost per acquisition split by channel, payback period, one leading indicator specific to the current constraint, and a quality measure so the leading indicator cannot be gamed. Everything else is available and not displayed.
That last pairing matters. Any leading indicator on its own will be optimised into meaninglessness given enough time. Pair it with a quality measure and it stays honest.
Organic makes that leading indicator harder to choose, for reasons set out in measuring organic when attribution breaks.
Track what you can produce on request, report only what changes a decision. Watch CAC, LTV, payback and conversion rate as a constraint set rather than individually, and always show both halves of a ratio so nobody improves it by shrinking the denominator.
The first thing I ask on a new engagement is CAC, conversion rate and LTV, in that order. If any of the three cannot be produced, that is the finding, and it usually predicts what else is broken.
I removed the specific client figures this post used to carry. They were real to me and unverifiable to a reader, and a post arguing for measurement discipline should not lean on numbers nobody can check.
Frequently asked
What are the most important marketing metrics?
How many marketing metrics should I report on?
What is the difference between a metric and a KPI?
Why is payback period more useful than ROAS?
Working on this yourself?
I design organic discoverability infrastructure for B2B SaaS teams, SEO systems, AI-search visibility, and utility-led acquisition.
Book a call