TikTok engagement rate: what viral TikToks get per view, and how their hooks and captions compare

There is no single good rate: it depends on what you divide by and which posts you compare. Per view, the median viral TikTok in our library, which holds posts found by platform search, got likes, comments and shares equal to 11% of its views (middle half 6.3% to 17%), against 5.8% for the TikToks the same searches found that did not go viral. The two groups are not matched: viral posts had far more views, and among posts under 100,000 views the gap was much smaller, 6.8% against 5.6%. Likes are most of it: a median like-to-view ratio of 9.7% against 5.1%, and about half of viral TikToks (49%) reached 10 likes per 100 views. Divided by followers instead, the median viral TikTok's engagement came to 508% of its follower count, because a viral post's views run far past that count, so measure per view. Most viral TikToks (85%) opened with what our rule calls a statement, its catch-all for any opening that is not a question, a number, a how-to or a negative, against 79% of the posts that did not go viral; across all the forms the mix showed no clear difference (p = 0.08). These describe what search shows, not causes or targets.

Original data from CreatorsCompanion. Sample: 1,696 viral and 306 non-viral TikToks from the same searches (mostly October 2026), an older sample of 2,817 and 388, 501 viral Reels, and 1,519 Reels from 41 tracked Instagram creators. Figures computed on from public post metrics, published as aggregates only. A post counts as viral when its views are higher than its creator’s follower count (counted as at least 1,000). Most posts were found through platform search, which shows popular posts first, so the figures describe that sample, not a typical post; the limits are listed below. How we collected and counted them: methodology.

Published . We make CreatorsCompanion, a tool that finds viral posts, so we have an interest in the topic; the method and limits are stated so you can judge the numbers yourself.

11%

median engagement rate of a viral TikTok: likes, comments and shares over views

n = 1,696 viral TikToks; middle half 6.3% to 17%

5.8%

the same rate for TikToks from the same searches that did not go viral

n = 306 TikToks; middle half 3.8% to 9.1%; under 100,000 views: 6.8% vs 5.6%

9.7% vs 5.1%

median likes per view, viral vs not viral TikToks

49% of viral and 14% of non-viral TikToks reached 10 likes per 100 views

1.7%

likes and comments per view on a typical Reel of the Instagram creators we track

41 creators, 1,519 Reels; breakout Reels got 1.01 times their creator’s usual rate

The finding: viral TikToks got more likes and shares per view

Engagement rate is the number creators use to judge whether a post did well. We measured it on the TikToks our library holds, posts found by platform search, and compared the viral ones, whose views passed their creator’s follower count, with the posts the same searches found that did not.

Per view, viral TikToks drew more. Their median engagement rate was 11%, likes, comments and shares over views, with a middle half of 6.3% to 17%. The TikToks that did not go viral had 5.8% (3.8% to 9.1%). Pick one of each at random and the viral one had the higher rate 73% of the time (p < 0.001). The gap is in likes and shares; comments per view showed no clear difference (0.05% against 0.08%, p = 0.06).

Two results temper that. Among posts under 100,000 views, where most of the non-viral TikToks sit, the gap was much smaller: 6.8% against 5.6%. And within the same Instagram creators, Reels that broke out showed no clear difference in likes and comments per view from their creator’s other Reels (a median of 1.01 times the creator’s usual rate). Post age and account size do not explain the gap on their own: matched one to one on age, viral TikToks had 10% against 5.8%, and matched on follower count 10% against 5.9%. A high rate goes with the posts search shows as viral; our data does not show it as the cause.

Openings told us less than captions. Most viral TikToks (85%) opened with what our rule files as a statement, its catch-all for any opening that is not a how-to, a question, a number or a negative, and the mix of openings showed no clear difference from the posts that did not go viral (p = 0.08). Captions and first sentences did differ: the shorter they were, the larger the share of the posts found that were viral.

How to read these numbers. Every TikTok here was found through platform search, which shows popular posts first, so the figures describe what search surfaces, not a typical post. Most TikToks were published in October 2026 and most captions are in English. Each post was measured once, when a search found it, and the posts that did not go viral lean toward larger accounts. These are descriptions of groups, not causes; the full list is under limits of this data, below.

How we measure engagement rate

Our engagement rate is per view: a post’s likes, comments and shares added up, divided by its views, times 100. We compute it for each post, then report the median post and the middle half (the 25th to 75th percentile), so one huge post cannot move the figure. Views are the public play count when a search found the post. The search library records no saves, so saves are not in the rate; our older sample has them, further down.

Why the rate in your analytics or a calculator may differ from ours:

  • Divided by followers. Many calculators divide by follower count. That suits a post seen mostly by followers and breaks for a viral one, whose views pass the follower count by definition. Divided by followers (counted as at least 1,000), the median viral TikTok’s engagement came to 508% of the count, and 74% of viral TikToks passed 100%; the non-viral ones came to 0.91%. By followers, the number mostly measures reach against account size.
  • Divided by people reached. One person can play a video several times. A rate over unique accounts reached divides by a smaller number than a rate over plays, so it comes out higher.
  • More kinds of engagement. Some rates add saves, and some add other actions. In our older sample, adding saves moved the median viral rate from 12% to 14%.
  • An average instead of a median. All engagement over all views leans toward an account’s biggest posts. The median post is closer to what a typical post gets.
  • A different moment. We read each post once, a median of 13 days after it went up for viral TikToks. Your dashboard shows today’s count.
The same engagement divided by followers instead of views
GroupPosts25th percentileMedian75th percentileAbove 100% of followers
Viral, search library1,69695%508%2,544%74%
Not viral, same searches3060.21%0.91%2.7%0%
Older sample, viral2,81776%408%3,280%71%
Older sample, not viral3880.99%2.7%6.0%0.26%

Followers are counted as at least 1,000, as in our viral test. A viral post’s views pass its follower count by definition, so its engagement over followers can pass 100%.

Engagement rate of viral and non-viral TikToks

Across the 1,696 viral and 306 non-viral TikToks from the same searches, likes carried most of the engagement, shares came next and comments were the smallest part. Viral TikToks had more of both likes and shares per view (0.58% shares per view against 0.22%; p < 0.001).

Engagement rate by views: likes, comments and shares over views, per post
GroupPosts25th percentileMedian75th percentileLikes per view, medianComments per view, medianShares per view, median
Viral, search library1,6966.3%11%17%9.7%0.05%0.58%
Not viral, same searches3063.8%5.8%9.1%5.1%0.08%0.22%
Older sample, viral2,8176.8%12%18%11%0.06%0.78%
Older sample, not viral3885.4%9.0%14%7.9%0.12%0.38%

Each post’s likes, comments and shares are added up and divided by its views. The percentiles describe posts, not accounts. The older sample was collected with a most-liked sort, which favours posts with many likes.

An older, separate sample from our first viral finder, collected with a most-liked search sort from November 2025 to March 2026, agrees in direction: a median rate of 12% for its 2,817 viral TikToks against 9.0% for its 388 non-viral ones (p < 0.001). That sort picks posts with many likes, which can push its rates up, and they do run higher. It also records saves: a median of 0.99% saves per view for viral TikToks against 0.68%, and with saves added the median rates come to 14% and 10%.

Saves per view, older TikTok sample
GroupPostsSaves per view, 25th percentileMedian75th percentileRate without saves, medianRate with saves, median
Viral2,8170.45%0.99%2.0%12%14%
Not viral3880.29%0.68%1.7%9.0%10%

The search library records no saves, so only the older sample can show them.

What like-to-view ratio is good?

The like-to-view ratio is the simplest version of the rate. The median viral TikTok had 9.7% (middle half 5.6% to 16%); the median non-viral one 5.1% (3.0% to 8.4%). About half of viral TikToks, 49%, reached 10 likes per 100 views, against 14% of the others. At 20 likes per 100 views it thins out: 14% of viral TikToks and almost none of the others (0.33%).

Share of TikToks reaching each like-to-view line49% of viral TikToks reached 10 likes per 100 views, against 14% of the TikToks from the same searches that did not go viral.
  • Viral (n = 1,696)
  • Not viral (n = 306)
  • 5% or more78% Viral (n = 1,696)50% Not viral (n = 306)
  • 10% or more49% Viral (n = 1,696)14% Not viral (n = 306)
  • 20% or more14% Viral (n = 1,696)0.33% Not viral (n = 306)

Search library, same searches: 1,696 viral and 306 not viral TikToks.

Likes per view, and the share of TikToks reaching a like-to-view line
GroupPosts25th percentileMedian75th percentile5% or more10% or more20% or more
Viral, search library1,6965.6%9.7%16%78%49%14%
Not viral, same searches3063.0%5.1%8.4%50%14%0.33%
Older sample, viral2,8175.9%11%16%80%54%13%
Older sample, not viral3884.7%7.9%13%72%37%4.9%

So a like rate near 10% put a TikTok in the middle of the viral posts search showed, and above most of the posts that did not go viral. It is a description of those posts, not a line that makes a video travel: judge your own against your median post.

By view count, grade and account size

Among viral TikToks, the rate rose with the view count: 6.8% for posts under 100,000 views, 10% from 100,000 to 1 million and 12% at 1 million or more. By our viral grade, Early had the lowest median (7.8%) and Exceptional the highest (12%). Account size barely moved it: viral TikToks from accounts under 10,000 followers had 10%, from 10,000 to 100,000 followers 11%, and from 100,000 to 1 million 11%.

The view count also gives a fairer comparison of viral and non-viral posts. Most non-viral TikToks (261 of 306) had under 100,000 views, so comparing within that band takes out much of the difference in views. It does not take out account size or age: by our definition a viral post under 100,000 views comes from an account with fewer followers than that (a median of 3,210 followers, against 78,000 for the non-viral posts in the band), and the viral posts in the band were measured at a median of 7 days against 3 days. There the gap shrank to 6.8% against 5.6% (218 and 261 posts). It is still clear (p < 0.001), but a random viral TikTok in the band had the higher rate 59% of the time, against 73% overall. Among accounts of 100,000 to 1 million followers, the one account band with enough posts in both groups, non-viral TikToks had 6.4% against 11% for viral ones.

Age and account size do not close the gap. We paired each of the 306 non-viral TikToks with the unused viral TikTok closest to it in age when measured, so both groups had a median age of 4 days: the medians were 10% against 5.8%, and a random viral TikTok in the pairs had the higher rate 70% of the time (p < 0.001). Paired on follower count instead (305 pairs, a median of 144,000 followers in both groups), they were 10% against 5.9%, with the viral TikTok higher 71% of the time (p < 0.001). In the one age band with 100 posts in both groups, TikToks measured 2 to 7 days after posting, the viral ones had 11% against 6.5% (347 and 106 posts). Matching on age or follower count barely moved the gap; keeping to posts under 100,000 views shrank it most.

Median engagement rate by views, by the post’s view count, the account’s size and the post’s age, and in pairs matched on age or follower count
GroupViral postsViral, median rateNot viral postsNot viral, median rate
Under 100,000 views2186.8%2615.6%
100,000 to 1 million views71310%under 100under 100
1 million views or more76512%under 100under 100
Under 10,000 followers78510%under 100under 100
10,000 to 100,000 followers64011%under 100under 100
100,000 to 1 million followers24211%1196.4%
Measured under 2 days after posting1168.2%under 100under 100
Measured 2 to 7 days after posting34711%1066.5%
Measured 7 to 14 days after posting45611%under 100under 100
Measured 14 to 30 days after posting77711%under 100under 100
Matched one to one on age when measured30610%3065.8%
Matched one to one on follower count30510%3055.9%

“Under 100” marks a group with fewer than 100 posts, too few for a median. Accounts of 1 million followers or more are left out: under 100 posts in both groups. Matched pairs: each non-viral TikTok is paired with the unused viral TikTok closest to it in age when measured, or in follower count, so both groups have the same number of posts and almost the same ages or follower counts.

Engagement rate by views of viral TikToks, by viral grade
Viral gradePosts25th percentileMedian75th percentileLikes per view, median
Early1374.9%7.8%13%6.9%
Promising1464.8%8.0%12%7.1%
Strong1326.2%10%15%9.6%
Very strong3536.0%9.8%15%8.8%
Exceptional9287.3%12%19%11%

The grade measures how far a post’s views passed its creator’s follower count, from Early to Exceptional; the cut-offs are on the methodology page.

Instagram Reels: likes and comments per view

The Instagram posts our searches find come without like or comment counts, so for Reels we use the creators our users track and compare each creator with their own Reels. For each of 41 creators with at least 10 Reels at least 3 days old, leaving out paid and pinned Reels, we took the median likes and comments per view of their Reels, then the median across creators. A typical Reel got likes and comments equal to 1.7% of its views (middle half of creators 1.2% to 2.3%): 1.5% likes and 0.09% comments per view. Divided by today’s follower count, the same Reels came to 0.15%.

These numbers do not compare with the TikTok ones above. The creators are mostly large accounts (a median of 158,000 followers) chosen by our users, Reel views are play counts, which can include replays, and our data has no share or save counts for these Reels, so the Reels rate is likes and comments only.

Instagram Reels of tracked creators: a typical Reel’s engagement per view, across creators
Measure25th percentile creatorMedian creator75th percentile creator
likes and comments per view1.2%1.7%2.3%
likes per view1.1%1.5%2.1%
comments per view0.03%0.09%0.14%
likes and comments per follower0.06%0.15%0.29%

Each creator’s median Reel first, then the spread across 41 creators (1,519 Reels at least 3 days old, paid and pinned Reels left out). Our data has no share or save counts for these Reels. Followers are today’s count.

Did breakout Reels get more engagement per view? We marked a Reel as a breakout when it reached at least 3 times its creator’s median views, over the last 365 days, the rule our other studies use. Then we divided each Reel’s likes and comments per view by its creator’s median rate. The 214 breakout Reels, from 38 creators, came in at a median of 1.01 times their creator’s usual rate (middle half 0.67 to 1.53); the 1,206 other Reels with a like count at 1.00. 50% of breakouts were above their creator’s median rate, against 49% of the other Reels, and the two groups showed no clear difference (p = 0.89). So a breakout Reel’s median rate per view sat close to its creator’s usual one, with no clear difference from the creator’s other Reels.

Breakout and other Reels: likes and comments per view, against the creator’s usual rateNo clear difference: breakout Reels at a median of 1.01 times their creator’s usual rate, other Reels at 1.00.
  • Breakout Reels1.01 (0.67 to 1.53)
  • Other Reels1.00 (0.75 to 1.32)

Dot: median. Bar: middle half (25th to 75th percentile). Line: 1, the creator’s usual rate. 39 tracked Instagram creators.

Breakout Reels against their own creator’s usual engagement per view
GroupReels with a rateCreators25th percentileMedian75th percentileAbove the creator’s median rate95% interval, low95% interval, high
Breakout Reels214380.671.011.5350%44%57%
Other Reels1,206390.751.001.3249%46%52%

Each Reel’s likes and comments per view divided by its creator’s median rate: 1 means the creator’s usual rate. A breakout is a Reel with at least 3 times its creator’s median views, over the last 365 days. Reels showing no likes, which usually means the creator hid the count, have no rate and are left out.

A breakout is picked by its views, the same number the rate divides by, so read this as a description of these Reels and creators, not a test of what makes a Reel travel.

Hooks: how viral TikToks open

We sort each post’s opening by a fixed rule on its first sentence, taken from the transcript when we had one and otherwise from the caption’s first line. In this order: a how-to (it starts with “how to”, “how I” or similar), a question (it has a question mark, or starts with “why”), a number or list (it starts with a digit, a price or a number word), a negative (it starts with “stop”, “don’t”, “never”, “the biggest mistake” or similar), and a statement, which is everything else. The rule covers English, Spanish, French and Portuguese. We never publish the text itself.

Most viral TikToks opened with a statement: 85%, against 8.1% with a question and 3.5% with a number or list; how-to and negative openings held under 30 posts each. Among the TikToks that did not go viral, 79% opened with a statement and 12% with a question; across the forms the two groups showed no clear difference (p = 0.08; we call a difference real only at p < 0.01). Taken on its own, the statement share looked different (85% against 79%; p = 0.005 in a test of statements against every other opening), but we ran that test after seeing the data, so we report it and build nothing on it. Viral Reels also opened mostly with statements: 77% statements, 8.4% questions, 7.0% numbers or lists and 6.8% how-tos. Too few Reels in the search results missed the viral line to compare them.

How the first sentence opens, viral and non-viral TikToks85% of viral and 79% of non-viral TikToks opened with a statement; the mix showed no clear difference (p = 0.08).
  • Viral (n = 1,696)
  • Not viral (n = 306)
  • Statement (anything else)85% Viral (n = 1,696)79% Not viral (n = 306)
  • Question8.1% Viral (n = 1,696)12% Not viral (n = 306)
  • Number or list3.5% Viral (n = 1,696)under 30 posts, not shown Not viral (n = 306)

Search library, same searches: 1,696 viral and 306 not viral TikToks. A form with fewer than 30 posts in a group has no bar.

How the first sentence opens: share of posts by hook form
Hook formViral TikToksNot viral TikToksViral Reels
Statement (anything else)85%79%77%
Question8.1%12%8.4%
Number or list3.5%under 307.0%
How-tounder 30under 306.8%
Negative (stop, never)under 30under 30under 30
No first sentenceunder 30under 30under 30

Posts per column: 1,696, 306 and 501. “Under 30” marks a form with fewer than 30 posts, which we do not show. Each share is rounded on its own.

Seen the other way, most TikToks the searches found were viral whatever the opening (1,696 of 2,002), so compare the forms with each other: 86% of those opening with a statement were viral, 79% of questions, 81% of numbers or lists and 83% of how-tos. The mix test above found no clear difference (p = 0.08), so no form stands out.

Share of the posts the searches found that were viral, by hook form
Sample and hook formFoundViralShare viral95% interval, low95% interval, high
TikTok, Statement (anything else)1,6841,44386%84%87%
TikTok, Question17513879%72%84%
TikTok, Number or list746081%71%88%
TikTok, How-to35under 3083%67%92%
TikTok, Negative (stop, never)under 30under 30under 30under 30under 30
TikTok, No first sentenceunder 30under 30under 30under 30under 30

A form is shown when the searches found at least 30 TikToks with it; a count under 30 prints as under 30. Reels are left out: too few Reels in the search results missed the viral line.

Among viral TikToks, the 138 that opened with a question had a lower median engagement rate than the 1,443 that opened with a statement: 8.6% against 11% (p < 0.001). That describes the two groups. Questions may be more common in some niches or formats, so it does not show that a question costs engagement.

Engagement rate by views of viral TikToks, by hook form
Hook formPosts25th percentileMedian75th percentileLikes per view, medianComments per view, median
Statement1,4436.5%11%18%10%0.05%
Question1384.4%8.6%13%7.4%0.06%

The other forms hold fewer than 100 viral TikToks each, too few for a median.

What this cannot tell you: whether speaking to the viewer, a contrast or a curiosity gap works better. Our rule does not sort openings that way, and we do not publish hook text. For openings to try, the free hook generator writes ten from one idea, in eight types.

Captions, hashtags and the first sentence

Viral TikToks had shorter captions: a median of 119 characters against 210, with every character counted, hashtags included. 39% of viral TikToks had a caption under 100 characters, against 17% of the others, and 13% had 500 characters or more, against 26%. Seen the other way, the share of the posts found that were viral fell as captions got longer: 93% with a caption under 100 characters, down to 72% with 500 characters or more (p < 0.001 for the mix).

Caption length, viral and non-viral TikToks39% of viral TikToks had a caption under 100 characters, against 17% of the TikToks from the same searches that did not go viral.
  • Viral (n = 1,696)
  • Not viral (n = 306)
  • Under 100 characters39% Viral (n = 1,696)17% Not viral (n = 306)
  • 100 to 199 characters31% Viral (n = 1,696)31% Not viral (n = 306)
  • 200 to 499 characters18% Viral (n = 1,696)26% Not viral (n = 306)
  • 500 characters or more13% Viral (n = 1,696)26% Not viral (n = 306)

Search library, same searches: 1,696 viral and 306 not viral TikToks. Characters include hashtags.

Hashtags: both groups had a median of 5, and most posts used 3 to 5 hashtags (81% of viral and 67% of non-viral TikToks). The edges went less often viral: 73% of the TikToks found with no hashtag, and 72% of the 57 with 6 hashtags or more, against 87% with 3 to 5 hashtags (p < 0.001 for the mix). 23% of non-viral TikToks had no hashtag, against 11% of viral ones.

The first sentence was shorter on viral TikToks too: a median of 8 words against 12 (p < 0.001). Of the TikToks found whose first sentence ran 1 to 5 words, 90% were viral, against 77% at 21 words or more.

Captions, hashtags and the first sentence: viral and non-viral TikToks from the same searches
MeasureViral, shareNot viral, shareFoundShare viral95% interval, low95% interval, high
Caption: Under 100 characters39%17%71993%91%94%
Caption: 100 to 199 characters31%31%61285%82%87%
Caption: 200 to 499 characters18%26%37779%75%83%
Caption: 500 characters or more13%26%29372%67%77%
Hashtags: No hashtag11%23%26373%67%78%
Hashtags: 1 or 2 hashtags5.0%under 309787%78%92%
Hashtags: 3 to 5 hashtags81%67%1,58487%85%89%
Hashtags: 6 hashtags or more2.4%under 305772%59%82%
First sentence: 1 to 5 words29%17%53690%88%93%
First sentence: 6 to 10 words34%28%65587%84%90%
First sentence: 11 to 20 words23%31%48781%77%84%
First sentence: 21 words or more14%23%31177%73%82%

Viral TikToks: 1,696; not viral: 306. Caption length counts every character, hashtags included. The first sentence comes from the transcript when we had one, otherwise from the caption. “Under 30” marks a cell with fewer than 30 posts.

These are patterns in what search shows, not rules for your captions. The posts that did not go viral lean toward larger accounts and were measured younger, and a caption written for search or for a brand can differ in many ways at once.

Did the two groups differ? Rank tests
ComparisonFirst groupSecond groupChance the first is higherp
Engagement rate by views: viral vs non-viral TikToks1,69630673%< 0.001
The same, posts under 100,000 views only21826159%< 0.001
The same, pairs matched on age when measured30630670%< 0.001
The same, pairs matched on follower count30530571%< 0.001
Likes per view: viral vs non-viral TikToks1,69630673%< 0.001
Comments per view: viral vs non-viral TikToks1,69630647%0.06
Shares per view: viral vs non-viral TikToks1,69630670%< 0.001
Engagement rate by views: older sample, viral vs non-viral2,81738861%< 0.001
Engagement rate by views, viral TikToks: question vs statement openings1381,44338%< 0.001
Caption length: viral vs non-viral TikToks1,69630633%< 0.001
Words in the first sentence: viral vs non-viral TikToks1,69030039%< 0.001
Hashtags: viral vs non-viral TikToks1,69630655%0.004
Breakout vs other Reels, rate over the creator’s median2141,20650%0.89

A Mann-Whitney rank test. “Chance the first is higher” is how often a random post from the first group beats a random post from the second (50% means no tendency). We call a difference real only at p < 0.01.

Did the mix differ? Chi-square tests between viral and non-viral TikToks
ComparisonFirst groupSecond groupChi-squareDegrees of freedomp
Hook form (negative and no first sentence pooled as other)1,6963068.3840.08
Statement vs every other opening (run after seeing the data)1,6963067.7610.005
Caption length band1,69630679.783< 0.001
Hashtag band1,69630641.733< 0.001

Each test asks whether viral and non-viral TikToks were spread differently across the forms or bands. The Reels hook test is left out: with so few non-viral Reels, some expected cells fall under 5, too few for the test.

What this means for your posts

  • Measure per view, against yourself. Add likes, comments and shares, divide by views, and compare each post with your own median post. A follower-based rate tells you more about reach than about the post. To check one of your own, the free TikTok engagement rate calculator does the sum and places the result against these medians.
  • Read a like rate in context. Around 10 likes per 100 views sat in the middle of the viral TikToks search showed, and above most of the posts that did not go viral. It is a description, not a target. Within the same Instagram creators, breakout Reels showed no clear difference in rate per view from their creator’s other Reels, so a breakout showed in its views, not its rate.
  • Read a competitor the same way. Our guide on how to see which posts are working best for your competitors applies the views test to a rival’s recent posts.
  • No opening form stood out across the mix. Viral and non-viral TikToks showed no clear difference in how they opened (p = 0.08), and our rule cannot sort openings by style. Test a few openings with the hook generator, and build the rest of the video with the Reel script guide.
  • Read caption length as a pattern, not a rule. Shorter captions and first sentences were more common among the viral TikToks search showed; we did not test whether shortening a caption changes anything. The TikTok caption generator writes short captions, and the TikTok hashtag generator a small mixed set.
  • Start from what travels in your niche. The free niche check ranks recent TikToks and Reels by views against followers. For the view numbers behind “viral”, see how many views is viral, and for length, viral video length. Our guide on how to go viral on TikTok puts these findings next to what TikTok says its For You feed weighs.

What our data cannot tell you

  • Cause. A higher rate, a short caption or a statement opening went with viral posts in what search showed. None of it was tested, and the groups differ in account size and age.
  • Watch time. Public metrics carry no watch time or completion rate, so we cannot say how long people stayed.
  • Hook styles the rule does not sort. Speaking to the viewer, a contrast, a curiosity gap or a story: our rule sorts openings only into the five forms above.
  • Your niche. We did not split these rates by niche: each niche category comes from a few searches taken once, which our methodology treats as too thin to publish. The free niche check shows what is viral in yours right now.
  • Instagram Reels from search, YouTube and LinkedIn. Our Instagram search results carry no like or comment counts, too few YouTube videos passed the viral line, and we have no LinkedIn data.

Limits of this data

  • Search results, not a random sample. Every TikTok here was found by platform search, which shows popular posts first, so even the non-viral posts are ones search chose to show. The rates describe that sample, not a typical post.
  • Viral and not viral are not matched. A post counts as not viral when its views do not pass its follower count (counted as at least 1,000), so the non-viral group leans toward larger accounts. It was also measured younger: a median of 4 days after posting, against 13 days for viral TikToks. Matched one to one on age or on follower count, viral TikToks still had the higher rate; among posts under 100,000 views the gap was much smaller.
  • One reading per post. We read each post’s numbers once, when a search found it. Engagement per view can move as a post ages.
  • What we count. TikTok rates add likes, comments and shares; the search library has no save counts, so saves appear only for the older sample. Our data has likes and comments only for tracked Reels, and we left out 17 Reels that showed no likes, which usually means the creator hid the count.
  • Hook forms are a simple rule. We sort the first sentence by a fixed rule, and “statement” holds everything the rule does not catch. Punctuation in an automatic transcript can be wrong, which affects the question form. We never publish the text.
  • Short windows, one niche mix. Search library TikToks were published between September 10 and November 11, 2026, mostly in October; the older sample was collected from November 2025 to March 2026. Business, marketing, fitness, real estate and money topics are over-represented, and most captions are in English.
  • Tracked creators are not a random sample. The Reels figures use creators our users chose to follow, mostly large Instagram accounts, with today’s follower counts.
  • Descriptive, not causal. Every comparison here describes groups of posts. None shows that a higher rate, an opening or a caption length made a post travel.

The full method, sampling and definitions are on the methodology page.

See what is viral in your niche

These numbers describe many niches at once. To see the posts beating their creator’s following in yours, type your niche: the free niche check ranks recent TikToks and Reels by views against followers, the same test this research uses. It needs no account.

Frequently asked questions

What is a good engagement rate on TikTok?

There is no single line, and it depends on what you divide by. Per view, the median viral TikTok in our search data got likes, comments and shares equal to 11% of its views (middle half 6.3% to 17%), and the TikToks the same searches found that did not go viral got 5.8% (3.8% to 9.1%). The groups are not matched: among posts under 100,000 views, where most of the non-viral ones sit, the gap was much smaller, 6.8% against 5.6%. These are posts that platform search chose to show, mostly in English and from October 2026, so read them as what search surfaces, not as a target that makes a post travel. The best comparison is your own median post.

What is a good like-to-view ratio on TikTok?

In our data the median viral TikTok had 9.7% likes per view, and the median non-viral one 5.1%. 49% of viral TikToks reached 10 likes per 100 views, against 14% of the others. 20 likes per 100 views is uncommon: 14% of viral TikToks and 0.33% of non-viral ones got there.

How do you calculate engagement rate on TikTok?

Add a video's likes, comments and shares, divide by its views and multiply by 100. That is the rate this study uses for each post, then we report the median post. Some calculators add saves or divide by followers instead; in our older sample, adding saves moved the median viral rate from 12% to 14%. Whichever you choose, use the same formula for every post you compare.

Why is my TikTok engagement rate over 100%?

Because it is divided by followers. A video that reaches beyond your followers can collect more likes than you have followers. Divided by followers (counted as at least 1,000), the median viral TikTok in our data came to 508%, and 74% of viral TikToks passed 100%. Divide by views instead and the same posts have a median of 11%.

What is a good engagement rate for Instagram Reels?

Across the 41 Instagram creators our users track, a typical Reel got likes and comments equal to 1.7% of its views (middle half of creators 1.2% to 2.3%), or 0.15% of the creator's follower count. These are mostly large accounts (median 158,000 followers), our data has no share or save counts for them, and Reel views are play counts, which can include replays, so compare your Reels with your own median Reel rather than with this figure.

Do question hooks get more engagement?

Not in our data. Among viral TikToks, those that opened with a question had a median engagement rate of 8.6%, against 11% for those that opened with a statement (p < 0.001). Questions were also no more common among viral TikToks than among the rest: 8.1% against 12%. That describes the groups; it does not show that a question hurts.

How long should a TikTok caption be, and how many hashtags?

Viral TikToks in our data had shorter captions, a median of 119 characters against 210, and 5 hashtags in both groups. Of the TikToks the searches found with a caption under 100 characters, 93% were viral, against 72% of those with 500 characters or more. That is what search shows, not a rule: the two groups also differ in account size and in how old the posts were when measured.

How to cite this

CreatorsCompanion, “TikTok engagement rate: what viral TikToks get per view, and how their hooks and captions compare,” published , https://www.creatorscompanion.app/research/tiktok-engagement-rate

You are welcome to quote these numbers with a link to this page. Every figure on the page, with the file and field it comes from, is in the CSV download. Spotted an error? Email support@creatorscompanion.app and we will correct the page and date the change.

More research