The Marketing Metrics I Stopped Reporting, and What Replaced Them

I have removed five metrics from my marketing reporting, and every one of them made the deck look worse.

That was the point. A metric that only ever moves up is not measuring anything. It is decorating the meeting.

I want to be precise about what this is and is not. This is not a claim that impressions are meaningless or that nobody should ever look at cost per lead. It is a claim about what belongs on the page you defend in front of the people who fund you. That page has limited space, and the wrong five numbers on it will quietly steer a year of work.

What criterion did I use to cut them?

Can this number go up without the business getting better?

If the answer is yes, it is a diagnostic. Keep it in the working file, look at it when something breaks, and take it off the report. If the answer is no, it belongs on the report.

That single question does more work than any framework I have been taught. It cuts across channels, across company sizes, and across the reporting-tool wars that eat so much of a marketing leader’s time.

The five I stopped reporting

1. Impressions. Replaced with qualified conversations. Impressions are the purest example of the criterion failing. I can double them this afternoon with budget and worse targeting. Nothing improves. The number that cannot be inflated that way is how many conversations happened with someone who could actually buy. That number is small, which is exactly why people avoid putting it on a slide.

2. Cost per lead. Replaced with cost per closed rupee. Cost per lead rewards cheap leads, and cheap leads are usually cheap for a reason. I have watched a cost-per-lead figure improve month after month while the revenue attached to those leads stayed flat, and the reporting gave nobody a reason to ask why. Cost per closed rupee is harder to calculate and slower to arrive. It is also the only version of the question a founder actually cares about.

3. Last-click attribution. Replaced with asking people how they heard about us. Last-click attribution is a story a tool tells you about itself. It reliably credits the final touch, which is reliably the cheapest and least interesting one. A free-text “how did you hear about us” field on the enquiry form is unscientific, messy, and closer to the truth than any dashboard I have run. People tell you about the podcast, the friend, the article from eight months ago. None of that survives a last-click model.

4. Follower count. Replaced with reply rate. I host a podcast that has crossed 1400+ episodes and more than 2 million listens, and neither number predicted a single business outcome on its own. What predicted outcomes was whether people wrote back. A follower is a row in a database. A reply is a person who decided you were worth the effort of typing.

5. MQLs. Replaced with pipeline sales has actually accepted. The marketing qualified lead is the most negotiated number in the profession. It is defined by marketing, counted by marketing, and reported by marketing, which is three too many. Replace it with the count sales has taken into their own pipeline and staked their quarter on. The two numbers will differ, and the gap between them is the most useful thing on the report.

What did not make the cut

Two candidates I considered dropping and kept, because the criterion said keep them.

Website traffic stays. Not as a headline, but as a diagnostic. Traffic cannot be trivially inflated once you exclude paid, and a sudden move in it usually means something real happened, good or bad.

Email list size stays for the same reason, with one condition: it sits next to open rate and reply rate, never alone. A list that grows while replies fall is a list that is being bought rather than earned, and the pair catches it in a way neither number catches on its own.

What it cost me to make the switch

The first reporting cycle after the change was uncomfortable. Numbers that had gone up for four quarters were gone, and the numbers that replaced them started low, because they had never been measured properly and there was nowhere to hide.

The second cycle was where it paid. Every conversation about the report became a conversation about the business instead of a conversation about the report. Nobody argued about the definition of an MQL, because there was no MQL left to argue about.

If you make this change, expect one bad month of optics and then a permanently better standard of argument.

Who this applies to

Marketing leaders and founders who present numbers to someone who funds them. If you are a team of one with no board, the same criterion applies. You are just the person you have to be honest with. My challenge to you: open your last report and find the one metric sitting there out of habit. Delete it before the next cycle.

Frequently asked questions

What marketing metrics should a small team actually report?

Qualified conversations, cost per closed rupee, self-reported attribution, reply rate, and pipeline sales has accepted. Five numbers, all hard to inflate.

Are impressions and reach completely useless?

No. They are diagnostics. Look at them when something breaks. Do not put them on the report you defend, because they can be doubled with budget alone.

Why replace last-click attribution with a survey question?

Because last-click credits the cheapest final touch and hides everything that created the demand. A “how did you hear about us” field is messier and closer to the truth.

What is wrong with MQLs?

Marketing defines them, counts them and reports them. Use the pipeline number sales has accepted into their own forecast instead, and watch the gap between the two.