TikTok analytics explained: how to read every metric that matters

Updated July 2026

Short answer: TikTok analytics live in TikTok Studio under the Analytics tab. The metrics worth reading are watch time, average watch time, completion rate, traffic sources, and follower activity — they explain why a video performed, not just that it did. Check them in a 10-minute weekly review and use retention graphs to find the exact second viewers leave.

Most creators open their analytics, see a views number, feel good or bad about it, and close the tab. That's not reading analytics — that's checking a scoreboard. The actual value in TikTok Studio is diagnostic: it tells you why a video did what it did, and what to change in the next one. Here's how to read it like a practitioner, in ten minutes a week.

Where TikTok analytics actually live

Everything runs through TikTok Studio. In the main app, go to your profile, tap the menu in the top corner, and choose TikTok Studio — Analytics is right there. There's also a standalone TikTok Studio app, and a desktop version that's genuinely worth using because retention graphs are much easier to scrub with a mouse than a thumb. Any public account gets access; you don't need a business account to see the numbers that matter.

You'll get account-level data across 7-day, 28-day, and custom date ranges, plus per-video analytics behind the three-dot menu on any video you've posted. Account-level trends are for your weekly review. Video-level data is where the real lessons live.

The metrics that matter and what each one actually tells you

TikTok shows you a lot of numbers. Five of them earn your attention. Here's what each one is actually measuring.

Views

A view registers roughly the moment your video starts playing, which makes views closer to an impression count than a quality signal. Views tell you how widely TikTok distributed the video — nothing about whether anyone wanted it. Treat views as the output of the system, not the input. When views are low, the cause is almost always buried in the retention data, not in the view count itself.

Watch time and average watch time

Total watch time is distribution multiplied by interest, so it's noisy. Average watch time is the one to read: compare it against your video's length. If your video is 30 seconds and average watch time is 8 seconds, most people are gone by the first third — that's a hook problem or a pacing problem, and it's fixable. Average watch time relative to length is the fastest single read on whether a video held people.

Completion rate

The percentage who watched the full video is the strongest quality signal TikTok gives you, with one big caveat: it's length-dependent. A 7-second video will always complete at a higher rate than a 60-second one, so never compare completion across videos of different lengths. Compare your 30-second videos against your other 30-second videos. When a video completes unusually well for its length, make more like it.

Traffic sources

Traffic sources break down where views came from, and the mix tells you what kind of asset you made. A For You-heavy mix means the algorithm distributed it to strangers — that's the growth engine working. A Search-heavy mix means the video answers something people actively type, which usually means it keeps earning views for months. Profile-heavy views mean people who found one video went digging through your catalog — a sign your content binges well. Following-heavy views with little For You traffic means the video served your existing audience but didn't break out.

Follower activity

The Followers tab shows when your audience is most active. Be honest about what this is worth: posting time matters less than most creators think, because the For You feed serves videos for days, not hours. The more useful read in this tab is new followers per video — it tells you which content actually converts viewers into an audience, which is a different skill from getting views.

How to find the exact second people leave

Open any video's analytics and scroll to the retention graph. This is the single most useful chart on the platform, and most creators never look at it. It plots the percentage of viewers still watching at every moment of the video.

Every retention graph drops steeply in the first two to three seconds — that cliff is normal and happens to everyone, because a chunk of feed viewers swipe on reflex. Don't panic about the cliff. Judge the slope after it.

Look at the curve after the opening cliff. A gentle, steady slope is healthy. What you're hunting for is an elbow — a spot where the line suddenly steepens.
Note the timestamp where that elbow starts. That's the second people decided to leave.
Rewatch your own video at exactly that moment. There's almost always an identifiable cause: a lull, a topic change, a long-winded explanation, a mid-video ask for follows, or the payoff arriving and giving people permission to swipe.
Check for bumps or spikes near the end — those usually mean people rewatched or looped a moment. Loopable endings are worth engineering on purpose.
Write down the lesson and apply it to the next video. Don't re-edit and repost the old one — the reach is already spent.

That last point is the honest limitation of analytics: retention graphs are a post-mortem. The video already used its shot at distribution by the time you can see where it bled viewers. If you want a similar read before you post, ReelTok — an iOS app from Viral App Labs — analyzes your video before publishing, scoring it 0 to 100 for virality with predicted reach and hook feedback from its Surge AI coach, all processed on-device with a 3-day free trial and no account required. Pre-post analysis plus post-post retention graphs covers both ends of the feedback loop.

The numbers you can ignore

Half the battle of reading analytics is knowing what not to read. These numbers eat attention and give almost nothing back.

Total likes as a standalone number. Likes divided by views is a mildly useful ratio; the raw count is a vanity metric that mostly tracks distribution.
Daily follower fluctuation. Followers drift up and down every day for reasons that have nothing to do with your content. Read the weekly trend, ignore the daily noise.
Profile views day to day. Useful as a monthly trend, meaningless as a Tuesday number.
First-hour view counts. Some videos ramp over days. Judging a video at hour one leads to deleting content that was still being tested by the algorithm.
Anyone else's numbers. Completion rates, view counts, and ratios vary wildly by niche and video length. Your only fair comparison is your own recent videos of similar length.

The 10-minute weekly review

Analytics checked daily produce anxiety. Analytics reviewed weekly produce decisions. Same day every week, ten minutes, this sequence:

Minute 1: Open TikTok Studio and set the range to the last 7 days.
Minute 2: Find your best and worst video of the week by views. These two are your entire syllabus for the session.
Minutes 3-4: For the best video, check traffic sources and average watch time. You're answering one question: what did this video do that the others didn't?
Minutes 5-7: For the worst video, open the retention graph and find the elbow. Rewatch that exact second. Name the cause in plain words.
Minutes 8-9: Scan completion rates across the week's videos, grouping by similar length. Flag anything that overperformed its group.
Minute 10: Write one sentence starting with 'Next week I will...' based on what you found. One change per week compounds; five changes per week just makes noise.

Your analytics checklist

Analytics live in TikTok Studio — use the desktop version for retention graphs.
Read average watch time against video length, not in isolation.
Compare completion rates only between videos of similar length.
Check traffic sources to learn what kind of asset each video became.
Find the retention elbow on your worst video every week and name the cause.
Ignore daily follower noise, raw like counts, and first-hour panic.
Run the 10-minute review weekly and change exactly one thing per week.

That's the whole system. TikTok doesn't publish its exact ranking weights, and it never will — but retention-style signals are generally understood to drive distribution, and every number above feeds back into one skill: holding attention one second longer than you did last week. Read the graphs, name the cause, fix it in the next video.

Know your score before you post

ReelTok's AI analyzes your video on your iPhone — a 0–100 virality score, predicted reach, hooks, and fixes before you share. Free 3-day trial.

Frequently asked questions

Where do I find TikTok analytics?

TikTok analytics live in TikTok Studio: tap your profile, open the menu, choose TikTok Studio, then Analytics. You can also use the standalone TikTok Studio app or the desktop version for a bigger view of retention graphs. Individual video stats sit behind the three-dot menu on each video you've posted.

What is a good completion rate on TikTok?

There is no official benchmark — TikTok doesn't publish one — but a useful working rule is that short videos under about 15 seconds should see roughly half of viewers finish, while longer videos naturally land lower. Compare completion against your own videos of similar length rather than against other creators or other niches.

How do I read a TikTok retention graph?

Open a video's analytics, scroll to the retention section, and look for the point where the curve suddenly steepens — that timestamp is where viewers left, so rewatch your video at that exact second to see what triggered the exit. Every graph drops sharply in the first seconds; that early cliff is normal, so judge the slope after it.

What do traffic sources mean in TikTok analytics?

Traffic sources show where each view came from: the For You feed, Following feed, Search, your Profile, or a Sound page. A For You-heavy mix means the algorithm is distributing your video, a Search-heavy mix means it answers a query people keep typing, and Profile-heavy views mean viewers are binging your catalog.

How often should I check TikTok analytics?

Once a week for ten minutes is enough for most solo creators — daily checking mostly produces anxiety, not insight, because single-video numbers swing too much day to day. Do one weekly review comparing your best and worst videos, and only open individual video analytics mid-week when something clearly outperforms or flops.

Do views or watch time matter more on TikTok?

Watch time matters more, because views only tell you how widely TikTok distributed a video, while watch time and completion tell the algorithm whether people actually wanted it. TikTok doesn't publish exact ranking weights, but retention-style signals are generally understood to drive further distribution, so treat views as the output and watch time as the input.

Related guides

Keep going: try the free TikTok hook generator and the virality score checker, browse hook examples and video ideas for your niche, or look a term up in the creator glossary.