Open any analytics dashboard and you will find hundreds of numbers. Impressions, reach, plays, replays, profile visits, link taps, average watch time, audience retention and many more. Most of them describe what happened. Only a few help you decide what to do next. Here is how we separate the two.
Start with the job of the content
A metric is only useful in relation to a goal. Before choosing what to measure, decide what each piece of content is for. We group content into three jobs and give each one its own scorecard.
Reach content
Built to find new people. Watch three second hold rate, shares per thousand views and follows per thousand views. Raw views are useful context, but they swing wildly with the algorithm.
Trust content
Built to deepen the relationship with people who already know you. Watch saves, comments and average watch time. A save is a strong signal that someone found real value.
Conversion content
Built to drive an action. Watch link clicks, sign-ups, sales and cost per result. Use tracked links and codes so you can connect posts to outcomes.
Report on the job each post was hired to do, and the dashboard stops being noise.
Use rates, not totals
Totals make big accounts look good and small accounts look bad. Rates make content comparable. Shares per thousand views tells you whether a video was worth passing on, whatever its reach. Follows per thousand views tells you whether people wanted more of you. Convert your key metrics into rates and the trends become much clearer.
Look at formats over time
Individual posts are noisy. One goes out during a big news event, another gets shared by a larger account. Group posts by format and look at the average over six to eight weeks. That is where the real patterns show up, and where decisions about what to make more of should come from.
The metrics you can safely ignore
- Impressions on their own. They count the same person many times and say nothing about attention.
- Likes as a headline number. Easy to give, easy to inflate and only weakly connected to business results.
- Follower count week to week. Watch the rate of new follows from content instead.
A simple monthly report
Our monthly reports fit on one page: the three or four rate metrics that match each content job, a comparison by format, the top three posts with a sentence on why they worked, and the decisions we are making next month. If a number does not lead to a decision, it does not make the report.
