The marketing metrics that actually matter (and the ones that don't)

Most marketing reports are long because length signals effort. They'd be more useful at a quarter of the size.

Here's what I'd keep.

The four that matter for almost everyone

1. Blended CAC. Total marketing spend ÷ new customers acquired. Ignores attribution entirely, which is why it's trustworthy. If this is rising, something is wrong regardless of what any platform claims.

2. Contribution margin per customer. What's left after cost of goods, fulfilment and payment fees. The number that tells you whether the CAC above is affordable.

3. Payback period. How long until an acquired customer has repaid their acquisition cost. This is a cash flow constraint, and it's the one that kills otherwise healthy businesses.

4. Conversion rate by landing page. The cheapest lever available. Improving conversion rate reduces CAC across every channel simultaneously, without buying anything.

If you tracked only these four, you'd make better decisions than most businesses tracking forty.

Worth watching, in context

ROAS — useful within a platform, misleading across platforms, and meaningless without margin context. 4x ROAS on 20% margin is a loss.

Cost per lead — only alongside lead-to-sale rate. Cheap leads that never close are expensive.

Hook rate on video — genuinely diagnostic for creative, tells you where something failed.

Repeat purchase rate — determines how much you can afford to pay for acquisition.

The metrics to stop reporting

Impressions and reach. They measure how much you spent, dressed as achievement.

Engagement rate. Correlates poorly with revenue. Comments and likes are not intent.

Click-through rate in isolation. A high CTR with low conversion means the ad promised something the page didn't deliver — that's a diagnostic, not a result.

Follower count. Unless you're selling advertising, this is a vanity number.

Bounce rate. Especially in GA4, where the definition changed and most people are still interpreting it as though it hadn't.

Position and impression share. Useful diagnostics occasionally. Not results.

The report I'd actually send

One page:

• Spend, revenue, blended CAC, contribution — this month, last month, same month last year
• The single biggest change and why
• What we're testing next and what would make us stop

Anything beyond that belongs in an appendix nobody opens.

Why this is hard in practice

Vanity metrics survive because they're always available and usually improving. Blended CAC requires connecting marketing data to financial data, which means someone has to reconcile two systems that disagree.

That reconciliation is the work. Skipping it and reporting impressions is how agencies keep clients who are quietly losing money.

A test for any metric

Ask: if this number doubled, would I change anything?

If impressions doubled, would you do anything differently? Probably not. If blended CAC doubled, you'd act immediately.

Metrics that don't change decisions aren't measurements. They're decoration.

For the arithmetic behind affordable CAC, see how to set a CPA target you can actually afford.