RevOps
The Marketing Metrics That Quietly Lie
Marketing metrics mislead when a field means something different from its name, when the denominator is wrong, or when an average hides drift. One automated report showed a 115.9% click rate because 'unique clicks' counted per-link clicks, not people; the true figure was 13.9%. Define every metric against the platform's schema, choose honest denominators and watch trends, not averages.
Key Takeaways
- Never clamp a number you don't understand. Find out why it is impossible.
- Check field definitions against the platform's own API documentation, not the label in the dashboard.
- Pick the denominator that reflects reality, such as messages seen rather than messages sent.
- Averages hide drift. Watch the latest sends against the baseline.
The impossible click rate
While automating a daily email report (case study 05), one campaign showed 401 'unique clicks' from 346 delivered emails: a 115.9% click rate. The tempting fix was to cap it at 100% and move on.
Checking the field against Mailchimp's own API schema showed the real issue: 'unique clicks' counted unique clicks per link, summed across links, not unique people. The subscriber-level field gave the true figure, 13.9%. A capped number would have looked fine and stayed wrong.
The wrong denominator
LinkedIn DMs are usually reported as replies divided by messages sent. In one founder-led campaign (case study 10), 104 people were contacted and 7 replied: 6.7%. But only 43 had seen the message, so the reply rate among people who actually read it was 16.3%.
Both numbers are true. Only one tells you whether the message works. The other mostly tells you about delivery and timing.
The average that hides drift
A newsletter averaging a 42.2% open rate (case study 03) looks healthy. Its September sends opened at 35.4% and 27.7%. Without a feedback loop comparing each send with the baseline, that slide would have stayed invisible until it became a problem.
How to keep metrics honest
- Define each metric in writing, against the platform's schema, before you report it.
- Pin known cases with tests. In case study 05, a regression test fixed to the exact 401/346 numbers stops the error from returning.
- Report the denominator next to every rate.
- Compare against a baseline, not just the all-time average.
Why this is a RevOps job
Reports drive decisions about budget, channels and people. If the numbers are defined loosely, every decision built on them inherits the error. RevOps is where those definitions live, so that marketing, sales and leadership read the same number the same way.
Put It To Work