Reach, impressions, ROAS: which numbers actually matter

A monthly marketing report is often twelve pages long because twelve pages feels like more work than two. Most of what fills those pages cannot be acted on. Here is a rough sorting of what is worth reading.

Numbers that mean something

Cost per acquisition. What it costs to get one customer, or one qualified enquiry. This is the number to argue about, because it is the one that decides whether the whole thing works. If your average customer is worth £200 and you are acquiring them for £40, everything else is detail.

Return on ad spend. Revenue divided by spend. Useful, with one caveat worth knowing: the platform reports it using its own attribution, which is generous to itself. Meta and Google will both happily claim the same sale. Sanity-check against your actual revenue.

Conversion rate, by landing page. The percentage of visitors who do the thing. When this is low, the problem is usually the page rather than the traffic, and it is the cheapest thing on this list to fix.

Revenue by channel. Not sessions, not clicks. Money. Frequently reveals that the channel everyone is excited about is producing very little of it.

Numbers that are context, not results

Click-through rate. Genuinely useful for judging creative: a low CTR usually means the hook is not landing. But a high CTR on an ad that sells nothing is not a result, and it gets presented as one often.

Cost per click. Worth watching for sudden changes, which usually signal audience fatigue or rising competition. On its own it says nothing about whether the campaign works.

Follower growth. Slow, steady growth is a reasonable sign the content is connecting. A spike is usually a giveaway or one lucky video, and rarely turns into revenue.

Numbers that are mostly decoration

Impressions. How many times something appeared. Buyable in unlimited quantity for very little money, which is exactly why it appears at the top of so many reports.

Reach. The same, deduplicated. Slightly better. Still not evidence anything happened.

Engagement rate. The tricky one, because it feels meaningful. It matters for organic content, where engagement affects distribution. It matters much less for paid, where you are buying distribution directly. A viral post with no sales attached is a hobby.

Video views. A "view" can be three seconds with the sound off. Watch-through rate is the number underneath that actually tells you something.

The number nobody puts in reports

What we learned, and what we are changing. Not a metric, and the most useful line in any report. A month where a test failed and the plan changed as a result is more valuable than a month where every chart went up slightly and nobody knows why.

If your reports never mention anything that did not work, that is worth asking about. Either nothing is being tested, or the failures are being quietly filtered out before you see them. Neither is good.

---

A useful report fits on one page and answers three questions: what did we spend, what came back, and what are we doing differently next month. Everything else is optional.