Analyse Another TikTok Channel: 6 Metrics to Read

To analyse another TikTok channel, such as a competitor's or a creator you are about to book, you do not need their TikTok Analytics. You read the public numbers on their latest videos: median views, views per follower, engagement per view, the gap to the peak video, posting rhythm and the trend by date.

How does your own channel differ from someone else's?

For your own channel, TikTok Analytics in TikTok Studio shows traffic sources, average watch time, finish rate, and the inferred gender, age and location of viewers. TikTok itself says these demographic statistics are "based on predictive models rather than confirmed user data" (TikTok Creator Academy), meaning they come from predictive models, not from confirmed user data. To read your own channel, use Analytics. CIA cannot replace it.

For someone else's channel, CIA reads the numbers TikTok shows publicly: views, likes, comments, shares and posting date for each video, plus the follower count and total likes on the profile. Two common cases are a competitor's channel and a KOC's (key opinion consumer, a consumer with influence) before you pay for a booking.

Every figure below is an estimate from public data. It is not the channel owner's internal number and it is not audited. Use the figures to compare and to ask questions, and do not close a contract on a single stat card.

Which six metrics are worth reading in public data?

CIA's channel analysis tool reads up to 45 of the latest videos. At the highest setting, a channel that posts daily gives a sample of about a month and a half; a channel that posts twice a week gives about five months. Check the posting date of the oldest video to see how long the sample runs.

Metric	Ready-made or self-calculated	Question it answers
Median views	Self-calculated: the MEDIAN function on the export file	How many views does a typical video get?
Views per follower	Self-calculated: average views divided by the follower count at the top of the page	Do videos reach beyond the follower base?
Engagement rate	Ready-made (stat card)	Do viewers react?
Peak video vs the rest	Highest and average views are ready-made; median is self-calculated	Is the channel steady, or living off a few hits?
Posting rhythm	Ready-made (posts/week card)	Is the channel still posting regularly?
Trend by date	Read the date column, or chart the export file	Is the channel rising or flattening?

Source: the tool's results screen, checked against the code on 28 Oct 2026.

Does median views show what a "normal" video gets?

The median is the middle number when views are sorted from low to high. The screen shows average views; you get the median yourself with the MEDIAN function on the views column of the CSV or XLSX file.

You need both because one hit video pulls the average up. An illustrative example, not real data: nine videos get 10,000 views each, and the tenth gets 1,000,000. The average is 109,000 and the median is still 10,000. Booking on 109,000 means paying for a video that may not repeat. The article median views matter more than follower count goes deeper.

What do views per follower tell you?

Take the average views, or the median, and divide by the follower count at the top of the page; the screen does not show this ratio. A ratio above 1 means videos are being pushed to people who do not follow the channel. This article gives no "good" or "bad" threshold because there is no reliable source for one. Use the ratio to compare channels in the same niche.

Engagement per view: divided by what?

CIA calculates engagement rate on the total views of the videos in the sample, not on followers. Many articles online divide by followers, so two numbers with the same name cannot be compared. The formula is in how to calculate TikTok engagement rate; what counts as a healthy level is in what a good engagement rate is.

How far above the rest is the peak video?

Put the highest-views card next to the average and the median. If the peak is many times the median, the channel may be leaning on a few hits. When booking, use the median as the expectation for the video you order, not the peak. How to use this when booking is in evaluate a creator before you book.

Posting rhythm: how many posts a week?

The posts/week card is an average over the whole span of the sample, so it hides recent change: a channel that stopped posting two weeks ago can still show a healthy number. Check the date column of the ten or so newest videos. Which rhythm suits which kind of channel is in TikTok posting frequency.

Are views rising or falling over time?

CIA has no stat card for the trend. Read the date column, or draw a line chart from the export file. Several weeks in a row below the earlier level is the point to worry; channel engagement dropping lists the causes to rule out.

How do you read single videos instead of the whole channel?

The table on the screen shows each video's posting date, description (truncated), views, likes, comments, shares and engagement rate, but it cannot be sorted. To analyse the videos in the export file, follow three steps:

1. Open the CSV or XLSX file and sort by the Views column. Take the 3 highest and the 3 lowest videos.
2. Put the description (complete in the file) and posting date of the six videos side by side. Do they share a topic, a format, a day of the week?
3. Compare each video's engagement rate with the channel's overall rate. A video with many views but below-average engagement may have travelled far while viewers scrolled past.

There is no separate sound or hashtag column; in the file, hashtags appear inside the description. To find out what viewers say, read the comments: analyse TikTok comments shows how, and CIA's Comments Analysis tool costs 12 credits and reads up to 300 comments per run.

What can CIA not read?

• No Analytics numbers. CIA cannot read watch time, finish rate, traffic sources or viewer demographics. A video with many views whose viewers left after two seconds looks identical to one that was watched to the end.
• A sample of up to 45 recent videos, not the whole channel. A long-running channel may once have looked very different.
• A private channel gives nothing to read. When no video can be retrieved, the credits are refunded.
• Where followers come from is covered in part, through the Audience Analysis tool (12 credits): it counts the region and language of about the 60 newest followers. That is not a random sample, and it has no age or gender. How to read it properly is in check creator audience overlap.
• The screen does not show saves; the export file has that column. Saves are not part of the engagement rate.

A TikTok channel belongs to a real person. The account owner can send a request to delete data about them, or to object to analysis of their account, on the data request page; an account that has requested a block is refused before any credit is charged. CIA does not look for the channel owner's email or phone number; to contact a creator, use the public contact details they publish themselves.

How do you run a channel analysis with CIA?

1. Paste the channel's @name or the TikTok profile link into the tool.
2. Choose how many videos to read (default 30). Whatever you choose, one run does not read more than 45 videos.
3. Run it. Each run costs 5 credits.
4. Look at the four stat cards (average views, engagement rate, posts/week, highest views), the header (followers, total likes, video count) and the per-video table. The export file holds the video rows, so take the follower count from the header.

What do you use the six metrics for: a rival or a creator?

A competitor. Read the same channel again after a few weeks, then compare posting rhythm, median and the videos that newly entered the top. The change shows which direction the competitor is turning. Track competitors on a schedule covers how to keep that reading rhythm.

A KOC before booking. Look first at three metrics: the median, views per follower, and the gap between the peak video and the median. An unusually high engagement rate is a reason to look closer, not proof that the channel buys engagement; signs of bought engagement lists the signs to check further. More articles are on the KOC topic page.

Frequently asked questions

How do I see another TikTok channel's numbers?

Public TikTok numbers include views, likes, comments and shares for each video, and the follower count on the profile. CIA reads these numbers, calculates the engagement rate and average views for you, and you calculate the median from the export file. CIA cannot read the numbers in their Analytics.

Do I need to log in to TikTok to analyse a competitor's channel?

Not with CIA. You need a CIA account and the public channel name; you do not need to log in to TikTok or ask the channel owner for permission. Pasting the @name or the profile link is enough.

Can I see another channel's video completion rate?

Not with CIA: the tool does not read finish rate or watch time. If you need this number for a creator before booking, ask the channel owner for a screenshot of their Analytics.

How is TikTok engagement rate calculated?

CIA calculates engagement on the total views of the videos in the sample, not on followers, so it cannot be compared with sources that divide by followers. The full formula is in how to calculate TikTok engagement rate.

What does it mean when a channel has few followers but high views?

Videos are reaching people who do not follow the channel: a good sign for reach, though it does not say whether viewers buy. Check the median as well to be sure one hit is not pulling the average up.

Can a private channel be analysed?

No. A private channel has no public videos to read, so the tool returns no numbers. When no video can be retrieved, the credits are refunded.

How many credits does it cost to analyse a TikTok channel?

The channel analysis tool costs 5 credits per run and reads up to the 45 latest videos of a public channel. Credit pack prices are on the pricing page.

Want to try it on a specific channel?

CIA's Channel Report tool reads up to the 45 latest videos of a public TikTok channel, 5 credits per run, with a CSV/XLSX file so you can calculate the median yourself. Every figure is an estimate from public data, not the channel owner's internal number.