Analytics dashboards and the iteration loop

Transcript of the narration, chapter by chapter.

0:00 Analytics and iteration

Here's a question worth sitting with. If you posted fifty videos last year, what did you learn from them? For many creators, the honest answer is: not much. They checked views in the first hour, felt good or bad, and moved on. The creators who grow fastest aren't necessarily the ones who post the most. They're the ones who learn the most per post. In this lesson, you'll learn where to find your numbers, the five metrics that matter, how to build a simple dashboard in Google Sheets, and a weekly iteration loop.

0:40 Where the numbers live

Where do you find your numbers? On TikTok, TikTok Studio, in the app and on the web, shows views, average watch time, how many watched the full video, retention, traffic sources like For You, Search and Profile, and audience data. On Instagram, the professional dashboard and post insights show reach, plays, watch time, likes, comments, shares, saves and follows, and Trial Reels show results among non-followers. The Edits app surfaces insights too. On YouTube, YouTube Studio shows views and engaged views, viewed versus swiped away for Shorts, average view duration, retention and subscribers gained. Menus change often, but these concepts stay stable.

1:24 Five vital signs

Now the five metrics that matter most. Think of them like a doctor's basic checkup: pulse, blood pressure, temperature. A few vital signs tell you most of what you need. One, early retention: the share still watching after about three seconds, or viewed versus swiped away. That's your hook. Two, average watch time or completion: did the middle hold? Three, shares and sends per thousand views: was it worth passing on? Four, saves per thousand views: was it worth keeping? Five, follows per thousand views: did it make people want more? Each metric points to a specific part of your video to fix.

2:09 Compare fairly

Why per thousand views? Because it lets you compare fairly. A video with two thousand views and a video with two hundred thousand views are hard to compare on raw shares. But shares per thousand views tells you which one people valued more. Same with saves and follows. And use medians rather than averages when you summarize, because one viral video can skew an average completely. Here's the key idea. You're not looking for your biggest video. You're looking for patterns in what works for you.

2:46 Example 1: A vs B

A simple example. Video A has five thousand views, sixty shares and forty saves. Video B has fifty thousand views, two hundred shares and a hundred saves. Raw numbers say B wins. Now per thousand views. A: twelve shares and eight saves per thousand. B: four shares and two saves per thousand. So people who saw A valued it much more. Maybe B got lucky with a trending sound. A might deserve a sequel, or even a paid boost.

3:20 Watch me: the dashboard

Watch me build the dashboard. I open Google Sheets and paste the header from the lesson: date, platform, pillar, format, hook type, length in seconds, views, three-second retention percent, average watch seconds, completion percent, shares, saves, follows, and notes. Now I add calculated columns. Shares per thousand equals shares divided by views, times a thousand. Same for saves and follows. And watch ratio equals average watch time divided by video length. I fill in my last twenty videos. Then I insert a pivot table with hook type as rows and the median of each rate as values. Immediately I can see result-first hooks beat question hooks on retention for me.

4:08 Example 2: UK finance creator

Now a realistic scenario, illustrative. Emma is a personal finance creator in the UK. She logs twenty-four videos over six weeks. Her dashboard shows that result-first hooks, like I saved three hundred twelve pounds last month with this spreadsheet, have a median three-second retention of seventy-one percent, versus fifty-eight percent for question hooks. Her budget template videos get four times more saves per thousand views than her opinion videos. So she changes one thing: every tutorial now opens with the result. Two weeks later, median retention on tutorials is up, and follows per thousand views have improved. She adds a templates series to her calendar.

4:54 The iteration loop

Here's the iteration loop. Review weekly, at the same time, and record numbers at seven days for fairness. Pick one lesson from the data. Change one variable in next week's videos, like hook type, length or format, not five things at once. Use built-in tests where they exist: Instagram Trial Reels, alternate hooks on similar posts, and YouTube's Test and Compare for long-form thumbnails and titles. And keep a winners library of your best hooks, formats and topics. It's like a science experiment. If you change the temperature and the ingredients at the same time, you'll never know which one made the cake rise.

5:39 Misleading numbers + mistakes

Beware of misleading numbers. Views alone can mislead, especially on YouTube Shorts, where views have been counted differently since March 2025, so check engaged views too. One viral video skews averages, so use medians. Different platforms define metrics differently, so compare within a platform. And small samples mislead, so look for patterns across several posts, not one. The common mistakes: checking stats obsessively in the first hour and deleting videos that seem to fail. Changing many variables at once. Tracking only views and likes. And never writing learnings down, so the same mistakes repeat.

6:20 Reading a retention graph

Let me show you how to read a retention graph, because it tells you exactly where to edit. Open the retention chart for a video in TikTok Studio or YouTube Studio. A steep drop in the first second or two means the hook didn't land. A gentle, steady slope is normal. A sudden cliff in the middle usually means something specific happened at that moment: a slow explanation, a confusing cut, or a moment where the promise felt broken. So scrub to that exact second and watch it with fresh eyes. And if the line bumps up near the end, people are rewatching, which often means you have a great loop or a moment worth replaying.

7:10 Recap

Recap. Use TikTok Studio, Instagram insights and YouTube Studio for early retention, watch time, shares, saves and follows. Normalize per thousand views and use medians. Build a simple sheet and pivot by hook type and pillar. And iterate weekly by changing one variable at a time, using built-in tests like Trial Reels.

7:33 Try this now

Try this now. Build the dashboard sheet from the lesson. Log your last ten to twenty videos. If you don't have enough yet, log a competitor's public numbers where you can see them, like views and comments. Create a pivot by hook type and pillar. Then write one sentence: next week, I will change this one variable, and I'll judge it by this one metric. Next module: creating brand-safe UGC for campaigns.

Posting is half the job

Every video is an experiment. The creators who grow fastest are not the ones who post the most; they are the ones who learn the most per post. That requires a small set of metrics, a place to record them and a habit of changing one thing at a time.

Where to find your numbers

TikTok: TikTok Studio (in the app and on the web) shows views, average watch time, watched full video, retention by second, traffic sources (For You, Search, Profile, Following), and audience data.
Instagram: the professional dashboard and post insights show reach, plays, watch time, likes, comments, shares, saves and follows; Trial Reels show results among non-followers. The Edits app also surfaces Reels insights.
YouTube Shorts: YouTube Studio shows views and engaged views, "viewed vs swiped away" for Shorts, average view duration, retention, and subscribers gained.

Names and locations of metrics change as apps update, but the concepts below stay stable.

The five metrics that matter most

Metric	What it tells you	What to change if it is weak
Early retention (share still watching after ~3 seconds, or viewed vs swiped away)	Did the hook work?	First frame, text hook, first line
Average watch time / completion	Did the middle hold attention?	Pacing, cut dead air, deliver sooner
Shares and sends per 1,000 views	Was it worth passing on?	More useful, surprising or relatable payoff
Saves per 1,000 views	Was it worth keeping?	Checklists, tutorials, reference value
Follows per 1,000 views	Did it make people want more?	Clear niche, series, CTA for part two

Normalize by views (per 1,000 views) so you can compare a video with 2,000 views to one with 200,000.

Hands-on: a simple content dashboard

Create a Google Sheet with one row per video:

date,platform,pillar,format,hook_type,length_s,views,retention_3s_pct,avg_watch_s,completion_pct,shares,saves,follows,notes

Add calculated columns:

shares_per_1k   =K2/G2*1000
saves_per_1k    =L2/G2*1000
follows_per_1k  =M2/G2*1000
watch_ratio     =I2/F2          (average watch time ÷ video length)

Then build a pivot table by hook_type and pillar showing the median of each rate. After 20–30 videos, patterns appear: perhaps "result-first" hooks beat "question" hooks for you, or your myth-busting pillar drives follows while tutorials drive saves.

The iteration loop

Review weekly at the same time. Record numbers at 7 days for fairness.
Pick one lesson from the data ("result-first hooks held better").
Change one variable in next week's videos (hook type, length, format), not five.
Use built-in tests where available: Instagram Trial Reels, alternate hooks on similar posts, YouTube's Test & Compare for long-form thumbnails and titles.
Keep a "winners" library of your best hooks, formats and topics.

Worked example (illustrative)

A UK personal-finance creator logs 24 videos over six weeks. Her dashboard shows that videos with a result-first hook ("I saved £312 last month with this spreadsheet") have a median 3-second retention of 71 percent versus 58 percent for question hooks, and that her "budget template" videos get four times more saves per 1,000 views than her opinion videos. She changes one thing: every tutorial now opens with the result. Two weeks later, median retention on tutorials has risen and follows per 1,000 views have improved. She adds a "templates" series to her calendar.

Reading a retention graph

Open the retention chart for a single video in TikTok Studio or YouTube Studio:

A steep drop in the first one or two seconds means the hook did not land. Test a new first frame and text hook.
A gentle, steady slope is normal; no video keeps everyone.
A sudden cliff in the middle points to a specific moment: a slow explanation, a confusing cut or a broken promise. Scrub to that second and fix it in the next video.
A bump near the end means people are rewatching, often a sign of a good loop or a moment worth replaying.

Beware of misleading numbers

Views alone can mislead, especially on Shorts since view counting changed in 2025.
One viral video skews averages; use medians.
Different platforms define metrics differently; compare within a platform.
Small samples mislead; look for patterns across several posts.

Common mistakes

Checking stats obsessively in the first hour and deleting "failed" videos.
Changing many variables at once, so you cannot tell what worked.
Tracking only views and likes.
Never writing down learnings, so the same mistakes repeat.

Analytics dashboards and the iteration loop