Every marketing report has numbers that nobody has ever made a decision with. Impressions. Reach. Followers. Time on page. They persist because they are easy to pull and they usually go up.
One test
If this number moves by thirty percent, can you say what should change about the business, and what you would do differently on Monday?
If not, it is not a KPI. It might still be useful for diagnosis. It does not belong at the top of the report.
Run that test across your current dashboard. Most of it will not survive, and that is a good outcome.
The short list
Cost per qualified conversation
Not cost per lead. Cost per conversation with someone who could actually buy. This requires agreeing what qualified means and having someone mark it, which is why most teams avoid it. It is also the number that tells you whether your acquisition is working.
Pipeline created
The value of real opportunities that came from marketing in a period. It moves earlier than revenue, which makes it useful for steering, and it is harder to inflate than lead count.
Close rate by source
The number that changes how you allocate. When one channel produces half the leads at twice the cost and closes at four times the rate, the cost per lead comparison was actively misleading you. This is usually the most surprising number on the list the first time a company pulls it.
Customer acquisition cost against customer value
Acquisition cost alone means nothing. Against what a customer is worth over their life, it becomes the number that tells you whether you can afford to grow faster or need to stop.
Time from first touch to closed deal
Under-tracked and very useful. It tells you how far ahead to plan, whether your nurture is the right length, and whether a change actually accelerated anything or just moved the timing of a report.
Two more worth adding now
Branded search volume. As more research happens in places you cannot see, people who become aware of you often arrive by typing your name. Rising branded search is one of the few visible signals that awareness work is landing.
Response time to inbound. Not a marketing metric by tradition, and it affects marketing results more than most marketing metrics do. Measure it honestly, including nights and weekends.
The metrics that are hardest to produce are usually the ones nobody can argue with.
Why the bad metrics survive
Three reasons, and they are all human.
They are easy. A platform hands them to you formatted. Pipeline attribution requires someone to maintain a system and make judgment calls.
They flatter. Impressions almost always rise. A report full of rising numbers makes for a comfortable meeting.
They are safe. A metric nobody can act on is a metric nobody gets held to. That is not usually a conscious choice, but it is a real dynamic in a lot of reporting relationships.
What a good report looks like
One page. Five or six numbers at the top with the period comparison. A short written note explaining what changed and what is being done about it. Diagnostics underneath for anyone who wants them.
The written note is the part most reports skip and the part most leaders actually read. A number without an interpretation just generates questions in a meeting.
Where to start
Pick one. Cost per qualified conversation is usually the highest-value first move, because producing it forces the team to define qualified, which is a conversation most businesses have never properly had.
That definition improves the ads, the follow-up, and the sales process, all before you have reported the number once.