how to measure content performance.
likes lie. here are the signals that tell you which content actually worked — and what to make next.
measure by goal: reach for awareness, saves for value, clicks for intent, conversions for sales. then make more of what wins.
knowing which content worked is how you stop guessing and start compounding. but "worked" depends on the goal — and reading the wrong signal sends you making more of the wrong thing.
— 01match metric to goal.
awareness? reach and new-follower reach. value? saves and shares. intent? profile and link clicks. sales? conversions. judge each post by the job it was meant to do, not a blanket number.
pick the metric from the objective, not from what the dashboard shows first. reach measures distribution, saves measure intent, and enquiries measure whether any of it worked — reporting all three as though they are the same thing is how content decisions get made badly.
for anything meant to sell, the only honest measures are enquiries and their source. for anything meant to build credibility, saves and branded search volume are the closest available proxies.
and ignore likes entirely. they are the cheapest action available on any platform, which makes them the least informative.
— 02find your patterns.
review monthly: which formats, topics and hooks consistently win? the patterns tell you what to make more of. a good report surfaces these, not just raw numbers.
look at your top ten posts by saves and shares, then the bottom ten. the pattern across both usually contradicts what the team assumed was working, and that contradiction is the whole value of the exercise.
separate format from topic. a subject that performed in one format and failed in another is telling you about the format, and it is the most common thing conflated in content reviews.
and check hold rate rather than views on video. views measure the hook; hold rate at three seconds measures whether the rest was worth making.
— 03feed it back into content.
measurement is only useful if it changes what you make. double down on winning formats, retire the losers, and let data shape the next content cycle.
make five more of whatever worked, immediately. most accounts treat a strong post as luck and move on, when it is the clearest signal the platform will give about what the audience wants.
and cull rather than add. dropping the format that consistently underperforms improves a calendar faster than introducing a new one, and it frees the production time to do the winner properly.
then hold the change for a quarter before judging it. reacting weekly produces constant adjustment and no compounding, which is the state most content calendars are permanently in.
— 04what a useful content review contains.
four things: what performed and why, what did not, what changes next month, and what needs a decision from you. everything else belongs in an appendix nobody reads.
and it should be written by whoever makes the content. a review assembled from a dashboard by someone else carries no judgment, which is the only part worth reading.