Marketing analytics can feel like a test you did not study for. There are many metrics, several platforms and a dashboard that someone else built. The good news is that you do not need all of them. You need the few that help someone make a decision.
Start with the decision
Before choosing a metric, finish this sentence:
Because of this number, we will decide whether to ______.
If you cannot finish it, the metric may be interesting but it is not yet useful. A decision could be to change a budget, rewrite a page, pause a campaign, contact a group of leads or investigate a drop in traffic.
This question is useful because it gives the number a job. It also stops a report from becoming a museum of everything the tools can measure.
Four questions that keep a dashboard honest
A dashboard is a communication tool, not a storage unit. These four questions help keep it focused:
- Who is reading this? A campaign manager, a founder and a finance lead need different levels of detail.
- What decision are they making? Each section should support a decision.
- What is the comparison? A number alone is not a result. Compare it with a target, a previous period or a different audience.
- What is missing? A dashboard should say what it cannot tell you, especially when tracking is incomplete.
Metrics to watch at each stage
It helps to group metrics by the stage of the customer journey rather than by platform.
Reach and discovery
These metrics describe who saw the marketing: impressions, reach, search visibility, new visitors. They are useful for understanding whether the message is being placed in front of the intended audience. They do not prove that anyone cared.
Interest and engagement
Saves, replies, shares, video completion, time on page and repeat visits suggest that someone found the content useful enough to act. Engagement is most useful when you can link it to a specific format, topic or audience.
Conversion
Conversions are the actions you have defined as valuable: a sign-up, a download, an enquiry, a purchase. The definition matters. A “lead” is only a lead if the business agrees on what counts.
Retention and quality
If you can see what happens after conversion, include it. A large number of low-quality sign-ups can look better than a small number of useful ones. Where possible, connect acquisition data with later outcomes such as activation, replies or renewals.
Avoid the trap of comparing unrelated numbers
Two numbers can rise at the same time without causing each other. A post can go viral while sales stay flat. Traffic can fall while revenue grows because the remaining visitors are more qualified.
When you present a comparison, name the relationship you are testing. If you do not know the relationship, describe the pattern and avoid claiming a cause. “Sessions fell after the navigation change” is an observation. “The navigation change caused sessions to fall” requires stronger evidence.
A simple weekly analytics routine
A short, repeatable routine is often more useful than a large monthly report no one reads.
- Monday: check the few numbers linked to current decisions.
- Midweek: look for anomalies and note possible explanations.
- Friday: write three lines: what changed, what it might mean, what you will do next.
Keep a log. Memory is a poor analyst. A dated note about a campaign change or a tracking update makes future changes much easier to interpret.
What to do when the numbers are flat
Flat numbers are information, not a verdict. Before changing everything, check:
- Was the campaign actually live for the full period?
- Did tracking break or change?
- Is the comparison fair? Seasonality, holidays and promotions can distort a simple week-on-week comparison.
- Is the audience large enough for a meaningful difference?
- Is the goal still the right goal?
Then change one meaningful thing at a time. A calm test is easier to learn from than five simultaneous changes.
The habit that matters most
The best analysts are not the people who know the most metrics. They are the people who can turn a number into a clear question, a sensible decision and a useful next step. Start with the decision. The metric list will become shorter, and the work will become clearer.
