Last quarter I sat with a growth team staring at a dashboard full of green arrows. Views up. CTR up. "Engagement rate" up.
Everything looked great. Until we asked the only question that matters:
"What's actually working, and what should we do next?"
Silence. We had data, not direction. I've been in that meeting too many times as a person on the hook for results.
This post is a practical answer to how to know what content works - and how to act on it.
Marketing shouldn't be a guessing game
Strategy should be driven by clarity and decisions, not vanity metrics.
The goal isn't to admire charts; it's to ship content, faster, with outcomes you can defend in a board meeting.
Read this:
Make Better Marketing Decisions
The telltales of content that isn't (and is) working
- LinkedIn: 3% engagement, 0 qualified clicks. Feels good in the feed; moves nothing.
- YouTube: 64% retention to 0:15, 35% drop at the CTA -> the hook works, the ask doesn't.
- Campaign dashboards: tens of charts, thousands of rows of data, no single sentence that explains what not to do and what to double down on next week.
"Working" content isn't the post with most likes. It's the piece that does the job you hired it to do.
The mistake: confusing activity with impact
Views, likes, and impressions are activity. Useful, but incomplete.
Define the job first, then measure against it:
- Awareness -> qualified reaches, brand searches, follows
- Engagement -> saves, shares, meaningful comments, watch retention, profile visits
- Action -> clicks, replies/DMs, email signups, demo requests, revenue influence
Pick one primary job per post. If you measure everything, you learn nothing.
A 5-step way to know if content's working
(in < 30 minutes)
1) Write the job-to-be-done (one line)
- "This LinkedIn post should drive qualified profile visits."
- "This YouTube video should retain viewers to the first value reveal."
2) Set a personal baseline, not an internet benchmark
- Use your last 20 posts/videos. Track medians, not averages.
- Compare each new piece against your baseline for its job.
3) Look at quality signals, not just counts
- LinkedIn: saves, profile visits, outbound clicks, DM rate, comment quality
- YouTube: retention to first 30s + first key point, click-through rate on the title/thumbnail, % viewers who reached your CTA chapter
4) Tag the pattern drivers
For each piece, tag quickly:
- Topic (problem-first, story, how-to)
- Hook (question, stat, contrarian)
- Format (text, doc, carousel; short vs. long; chapters)
- CTA (comment, click, DM, subscribe)
- Timing (day/time)
You are looking for repeatable combinations that beat your baseline.
5) Make a decision: keep / iterate / kill
- Keep: repeat the winning combo next week
- Iterate: keep the idea, change hook/format/CTA
- Kill: archive it; don't spend cycles defending a dud
Document your decision in one sentence to force clarity.
Two quick examples (what "working" looks like)
LinkedIn
- Post A: 150 likes, few saves, no clicks.
- Post B: 35 likes, 12 saves, 7 site clicks, 3 DMs from ICPs.
-> Post B "worked" even with lower vanity metrics. Job = action.
YouTube
- Video A: 5k views, 25% drop in first 10s, most viewers bail before value.
- Video B: 1.2k views, 65% reach first value point, CTR above your median.
-> Video B "worked" and deserves iteration. Job = engagement/retention.
The mini worksheet (steal this)
Get this google spreadsheet and use it for each post/video.
Click the link: Content Performance Google Sheet
Run this for your last 20 pieces. You'll see the story.
Common traps to avoid
- Measuring a conversion goal with awareness metrics
- Chasing platform "benchmarks" instead of your own baseline
- Calling a one-off spike a strategy
- Turning postmortems into deck-building exercises
Do this now (10-minute sprint)
- Open your last 10 LinkedIn posts and 3 YouTube videos.
- Write the job for each (one line).
- Mark wins where the job beat your median baseline.
- Tag topic, hook, format, CTA, timing for each win.
- Ship one follow-up this week reusing the winning combo.
You'll learn faster than any "best practices" listicle can teach you.
Where RevScope fits
Want the loop without the manual tagging?
Drop a YouTube or LinkedIn link into RevScope and get a short narrative of what's working, what's not, and what to do next—so you can make the keep/iterate/kill call in minutes, not weeks.
Try it now: Drop a YouTube or LinkedIn link -> get your takeaways in minutes.
