TikTok Analytics for Other Accounts: What You Can Actually See, and What Every Tool Is Estimating
- TikTok publishes engagement counts on every public video but never publishes another account's reach, retention, follower demographics or revenue — those are the numbers tools estimate.
- Estimated figures are not useless, but they are model output. Two tools looking at the same account will disagree, and neither can tell you which is closer.
- The genuinely reliable competitor data is free and public: posting cadence, format mix, hook structure, comment themes and how the account is classified by search.
- The highest-value competitor research is qualitative — read the comments under their best videos. That is where audience demand is stated in the audience's own words.
You want to know what's working for the accounts beating you. So you look for TikTok analytics tools, and you find plenty — dashboards showing competitor growth rates, engagement rates, estimated earnings, audience demographics, best posting times.
Some of that is real. A surprising amount of it is a model producing a confident-looking number from very little.
Knowing which is which changes both what you pay for and what you act on.
What's actually visible about someone else's account
| Data | Real or estimated? | Source |
|---|---|---|
| Views, likes, comments, shares on a video | Real | Displayed publicly by TikTok |
| Follower count, video count, bio | Real | Public profile |
| Posting frequency and timing | Real (observable) | Anyone can record it |
| Sounds, hashtags, captions, formats | Real | Visible on each post |
| Engagement rate | Calculated | Real inputs, but the formula differs per tool |
| Reach and impressions | Estimated | Not published for other accounts |
| Watch time / retention | Estimated | Never public |
| Audience demographics | Estimated | Never public — inferred from commenters or panels |
| Estimated earnings | Modelled | Guessed rate × guessed qualified views |
Everything above the line is a fact anyone can collect. Everything below it is a model. Tools rarely draw that line for you, because the numbers below it are the ones that justify the subscription.
"Estimated earnings" deserves a specific warning. It's typically an assumed rate multiplied by an assumed qualified-view count — and both inputs are unstable. Payouts run on qualified views, not the number on the video, and that ratio varies wildly by format. Anyone quoting a competitor's income from public data is stacking two guesses and reporting the product as a fact.
Why two tools give different numbers for the same account
Because they're doing different things and calling it the same thing.
- Different formulas. Engagement rate might be divided by followers, by views, or by views on recent posts only. Same account, three different rates, all "correct."
- Different sampling. Some read the last 10 videos, some the last 100. An account that changed strategy two months ago looks completely different depending on the window.
- Different collection cadence. A tool that sampled a video on day one and a tool that sampled it on day thirty are measuring different points on a curve.
The practical rule: pick one tool and never compare its numbers to another tool's. Internal consistency is the only property these estimates reliably have. It's the same discipline that applies to any third-party scoring product — the checkers that claim to detect reach restrictions have exactly this problem, and so do keyword tools.
The competitor data that's free, reliable, and mostly ignored
Here's the reframe. The most decision-useful competitor information isn't a metric at all, and you can gather it in an afternoon with a spreadsheet.
1. Cadence and consistency
How often do they actually publish, and has it changed? An account that quietly went from three a week to daily six weeks before it took off has told you something a growth-rate chart obscures.
2. Format mix
Count their last 30 videos by type — talking head, voiceover, text-on-screen, duet, over a minute, under thirty seconds. The mix tells you what they've found repeatable. Note especially how many exceed a minute, since that's tied to what earns under Creator Rewards.
3. Hook structure of their top five
Transcribe the first sentence of their five best-performing videos. Patterns show up fast — and unlike engagement rate, this is directly copyable as a structure rather than as content.
4. Comment themes — the highest-value item on this list
Read the comments under their best videos. Not the count. The content.
People state what they wanted and didn't get. "But what about X." "This doesn't work if you're Y." "Where do you find Z." Every one of those is a video someone has already told you they want, sourced from an audience that has proven it's interested in the topic. No analytics product surfaces this, because it isn't a number.
5. Search classification
Search the terms you both want and see whether they appear. This tells you how the system has classified them, and it's the closest thing to seeing their positioning from the algorithm's side. Keyword research on TikTok works alongside this.
When paying for a tool is justified
Three cases, all about scale rather than insight:
- Tracking many accounts over time. Manual observation doesn't scale past a handful. If you're monitoring thirty, automated collection earns its cost — the value is the time series, not the estimates.
- Discovery. Finding accounts in your niche you didn't know existed is genuinely hard by hand and something these tools do well.
- Reporting to someone else. If a client expects a dashboard, that's a legitimate reason — as long as the estimated figures are labelled as estimates in the deck.
Not justified: buying a subscription to answer "why are they beating me." That answer is almost always in the qualitative list above, and the number you're paying for is a modelled proxy for it.
Our own bias here is probably obvious: we build production and distribution tooling, not analytics dashboards, and the reason is that competitor numbers have never been the bottleneck we see. The bottleneck is how many genuinely different pieces of content someone can ship per week. Research that doesn't change what you publish is just a nicer-looking form of procrastination.
FAQ
Can I see another account's TikTok analytics?
Not their actual analytics — TikTok's analytics are private to the account owner. What's available publicly is engagement counts on each video plus profile-level facts. Any tool showing you a competitor's reach, retention or demographics is presenting an estimate, whether or not it labels it as one.
Are free TikTok analytics viewers accurate?
For the public numbers — views, likes, comments — they're simply reading what TikTok displays, so accuracy isn't the issue. For anything derived, free tools tend to use simpler models than paid ones, which mostly means they're wrong differently rather than more. Judge on whether the tool tells you which numbers are estimated; that's a better signal of quality than the numbers themselves.
What's the best free way to research TikTok competitors?
A spreadsheet with five columns: date, format, first line, view count, top comment themes. Fill it in for a competitor's last 30 videos. It takes an hour or two and produces more actionable material than most paid dashboards, because every column is something you can copy as a structure rather than admire as a statistic.
One thing to take away: before subscribing to anything, spend twenty minutes reading the comments under your top competitor's five best videos and write down every question that appears more than once. That list is your next five videos, and it cost nothing.