How to calculate TikTok engagement rate in 2026

TikTok engagement rate has two formulas and they produce very different numbers. Dividing interactions by views is the right approach for TikTok, because videos are pushed to people who do not follow the account. Dividing by followers gives a much higher number that sometimes exceeds 100%. Before comparing two channels, make sure both use the same denominator.

Two formulas, and why TikTok differs from Instagram

Divided by views:

Engagement rate = (likes + comments + shares + saves) ÷ views × 100

Divided by followers:

Engagement rate = (likes + comments + shares + saves) ÷ followers × 100

On Instagram or Facebook, most people who see a post already follow the account, so a follower denominator is defensible. TikTok is different: the For You page pushes videos to non-followers, so a channel with 10,000 followers can have a video with 500,000 views. Dividing that video's interactions by 10,000 produces a meaningless number.

This is not theoretical. In a test published by Kuli, author Michael Hodara ran the same channel through five popular engagement rate calculators and got results from 1.07% to 112.52% — a hundredfold spread, purely because each tool chose a different denominator (Kuli, accessed 21/09/2026).

A figure of 112% does not mean the channel is a hundred times better. It means the tool is dividing by the wrong thing.

Which benchmark to compare against

The two most-cited benchmark sets measure different things, which is exactly why they do not contradict each other:

Source	Denominator	Figure	Sample
Socialinsider, Q2 2026	views	3.85%	2 million videos from 214,507 profiles
Quid (formerly Rival IQ), 2026 report	followers	2.01% (all-industry median)	150 companies per industry, 18 industries

Sources: Socialinsider and Quid, both accessed 21/09/2026.

Socialinsider also records that the Q2 2026 figure of 3.85% is down 8% on the previous quarter, after sitting around 4.20%–4.30% from Q1 2025. Which means a benchmark taken from last year's article is already wrong, and will drift further.

Using a whole platform's number as the benchmark for your channel is another trap. The spread between industries is wider than the spread between platforms, and small channels almost always rate higher than large ones.

A worked example

A channel with 48,000 followers. Three recent videos:

Video	Views	Likes	Comments	Shares
1	120,000	4,800	210	350
2	35,000	1,500	64	90
3	810,000	32,000	1,400	4,100

Totals: 965,000 views, 44,514 interactions.

• Divided by views: 44,514 ÷ 965,000 × 100 = 4.61%
• Divided by followers (average per video): 14,838 ÷ 48,000 × 100 = 30.9%

The 30.9% looks impressive but all it says is that the third video was pushed beyond the follower base. The 4.61% is the number that compares to other channels.

Note also that the third video accounts for 84% of total views. If you average each video's rate instead of pooling totals, you get yet another number. When you report, say whether you pooled or averaged.

Three mistakes that cost money

Taking the media kit number on trust. A media kit is written by the channel, and no standard requires it to state the denominator. An honest creator can still send you the 30.9% from the example above because that is what their tool reports.

Comparing channels from different periods. Engagement moves with seasons and campaigns. December for channel A against April for channel B compares two different things.

Measuring across too few videos. One viral video drags the whole average up. Reading a few dozen recent videos gives a truer picture than three.

Checking a channel in a few minutes

Doing this by hand for dozens of videos is something nobody does twice. CIA's channel report reads a channel's 60 most recent videos and returns engagement on both denominators, median views, posting frequency and the spread of interactions across videos. One run costs 5 credits.

For a channel posting daily, 45 videos is roughly the last six weeks — enough to see the current trend, not enough to reconstruct a year. If you need longer history, run it periodically and keep the results, or use competitor watch for regular snapshots.

Every figure here, both CIA's and the cited reports', is an estimate from public data the platform displays. It is not the channel's internal analytics and it is not audited.

Frequently asked questions

What engagement rate is good? It depends on the denominator and the industry. With views as the denominator, Socialinsider's Q2 2026 reference point is 3.85%. But comparing against your own industry and your own channel size is far more useful than a platform-wide figure.

Why do two tools give different numbers for one channel? Almost always the denominator, or the number of videos read. Ask the tool what it divides by and how many videos it reads; if it cannot answer, the number compares to nothing.

Do saves and shares count? In the common formula, yes. Some tools count only likes and comments, which is another reason two numbers diverge.

Small channels rate higher, so are they better? Not necessarily. A high rate on a small audience can still produce fewer orders than a low rate on a large one. Read both the rate and median views before setting a budget.

Do I have to pay to work this out? No. The formula is above and you can add it up yourself. Paying means not opening 45 videos and typing each number by hand.

For more on this topic see Creators & KOLs. If your real job is choosing where to put a budget, see the ads a competitor is running covers the other side of the same decision.