Reading analytics

Goal

Understand what your analytics show, diagnose where a video lost viewers, and decide what to change, using your own data rather than rules of thumb.

What analytics are for

Analytics answer one question: what did viewers actually do? They do not tell you whether the idea was good, whether you are talented or whether you should quit. They tell you where attention went. That is enough to improve.

Metric names differ between platforms and change over time. The concepts below are stable; find their equivalents in your app’s insights section.

The key metrics

Concept	What it tells you	Typical name
Views / plays	How many times the video started	views, plays
Reach	How many unique accounts saw it	accounts reached
Hook retention	How many stayed past the first seconds	3-second views, hook rate, retention at 3 s
Average watch time	How long people watched on average	avg watch time
Completion rate	Share who watched to the end	watched full video, finish rate
Retention curve	Where viewers drop off, second by second	audience retention graph
Shares / sends	How often people send it to others	shares, sends
Saves	How often people keep it for later	saves
Likes / comments	Reactions, a weaker signal	likes, comments
Follows from the video	Whether it won new followers	follows, new followers
Traffic source / audience	Followers vs non-followers, demographics	audience, sources

As lesson 1.1 showed, the platforms point to finishing a video and sharing as strong signals. Look at those first.

Read the retention curve

If your platform offers a retention graph, it is the most useful chart you have. Where the line falls tells you what went wrong.

Pattern	Likely problem	What to try
Steep drop in the first 2-3 seconds	The hook failed or the first frame was weak	Rewrite the hook; start in motion; cut the warm-up; add on-screen text
Gradual decline through the middle	The middle is slow or unfocused	Tighten beats; cut repeated ideas; add a turn; vary the visuals
Cliff at a specific moment	Something happened there: a boring segment, a loud sound, a confusing line	Fix that moment
Low completion but good early retention	The promise was kept too early or the ending is flat	Shorten; end on the payoff; add a loop
Retention is fine but shares are low	People like it but have no one to send it to	Make the idea more specific to a person or situation
Good watch time but few new followers	People enjoy it but do not see a reason to follow	Make your thread clearer; mention what the account offers; show a series feel
Views are tiny with good retention	Few people saw it; the system may still be testing; the topic may be narrow	Wait 24-48 hours; try the idea again with a different hook

Followers versus non-followers

If most of your views come from followers, the video is circulating inside your existing audience. If many come from non-followers, it is reaching new people. Both matter: the first measures loyalty, the second growth. Compare the hook rate and completion between the two groups if your analytics allow.

How to avoid fooling yourself

• Small numbers are noisy. A reel with 40 views tells you little. Look at patterns across 5-10 reels.
• Don’t compare yourself to others. Their audience and history differ.
• Beware the one hit. A spike can be luck. Repeating the structure tells you whether it was a pattern.
• Look at the same time horizon. Compare videos at 24 hours and at 7 days.
• Remember what you cannot see. Analytics don’t show why someone left, only that they did.

Build your analytics table

After each post, add a row at 24 hours and update at 7 days.

Date	Title	Hook (exact)	Length	Views	Hook retention	Avg watch	Completion	Shares	Saves	Follows	Notes

At the end of two weeks, sort by completion and by shares. Compare the top three and the bottom three. What do they have in common?

Questions to ask each video

1. Where did people leave?
2. What was on screen then?
3. What did I say at that moment?
4. What’s one change to test next?

Common mistakes

• Focusing on likes.
• Judging after one hour.
• Changing five things at once.
• Chasing the format of one hit without understanding it.
• Ignoring your own gut: use data to inform, not replace, your voice.

Exercise

1. Open the analytics of your first posts.
2. Fill in your table for each.
3. For each, identify the drop-off pattern and write one hypothesis.
4. Pick a single change to try in the next reel (lesson 7.3).

Daily practice

• Minimum (keeps the streak): a voice note of any length.
• Today’s take (about 30 seconds): Tell the lens your top three observations from your analytics.
• Weeks 1-2 everything stays private; from week 3 follow your publishing plan.
• Don’t watch the take right away (lesson 4.1). Tick the day in your tracker.

Reading (optional)

This week’s book, about 15-20 minutes: Austin Kleon, Show Your Work!. A loose suggestion; it does not have to match the lesson. More in resources and the schedule.

Checklist

• I know the key metrics and their names in my app.
• I can read a retention curve.
• I filled my analytics table.
• I wrote one hypothesis per video.
• I recorded today’s voice note and take.

---

← 7.1: Posting mechanics | Course index | 7.3: The iteration loop →