OwlScran research · Quarterly edition
Creator Index: Q3 2026 edition
Observed benchmarks from 83 active OwlScran creators with a qualifying account. This edition preserves the published figures and methodology from one computation, so you can cite the same evidence later.
By OwlScran · Edition and methodology captured on . Computed .
The date identifies the computation, not a shared reporting window. This self-selected sample is not representative of all creators. Changes in accounts or methods can change later results; this is not a time-series comparison.
Published editions are quarterly. Q3 2026 identifies the publication, not a quarter-long reporting window. These figures come from the 17 September 2026 computation. The current Index can update between published editions.
The published figures
Each sample counts the accounts contributing to that metric. The middle half runs from the 25th to the 75th percentile; it is not a confidence interval. Results below 30 accounts are withheld. Unavailable metrics and withheld values are left empty in the CSV, with their status preserved.
Combined followers per creator
| Measure | Sample | Median | Middle half |
|---|---|---|---|
| Combined followers | 83 | 43,272 | 19,066.5 to 129,618.5 |
| Measure | Sample | Median | Middle half |
|---|---|---|---|
| Followers | 62 | 31,634.5 | 14,998 to 77,791.75 |
| Engagement per view | 52 | 6.47% | 4.73% to 9.26% |
| Average post performanceAccounts reached per post | 52 | 29,252.5 | 4,514.25 to 76,579 |
| Views from non-followers | 49 | 78.31% | 69.05% to 90.24% |
| Video views per follower | 40 | 0.67× | 0.28× to 1.27× |
| Top video’s share of views | 40 | 24.22% | 19.43% to 41.33% |
| Saves per view | 35 | 0.54% | 0.25% to 0.9% |
| Shares per view | 40 | 0.46% | 0.31% to 1.19% |
TikTok
| Measure | Sample | Median | Middle half |
|---|---|---|---|
| Followers | 59 | 22,542 | 10,885.5 to 74,459 |
| Engagement per view | 59 | 8.42% | 6.01% to 10.18% |
| Average post performanceViews per post | 59 | 17,101 | 3,186 to 127,789 |
| Views from non-followers | 0 | Unavailable | |
| Video views per follower | 37 | 0.37× | 0.14× to 1.12× |
| Top video’s share of views | 37 | 33.86% | 18.35% to 46.08% |
| Saves per view | 0 | Unavailable | |
| Shares per view | 37 | 0.3% | 0.13% to 0.54% |
YouTube
| Measure | Sample | Median | Middle half |
|---|---|---|---|
| Followers | 26 | Withheld | |
| Engagement per view | 26 | Withheld | |
| Average post performanceViews per post | 26 | Withheld | |
| Views from non-followers | 22 | Withheld | |
| Video views per follower | 16 | Withheld | |
| Top video’s share of views | 16 | Withheld | |
| Saves per view | 0 | Unavailable | |
| Shares per view | 14 | Withheld | |
| Measure | Sample | Median | Middle half |
|---|---|---|---|
| Followers | 1 | Withheld | |
| Engagement per view | 1 | Withheld | |
| Average post performanceAccounts reached per post | 1 | Withheld | |
| Views from non-followers | 0 | Unavailable | |
| Video views per follower | 1 | Withheld | |
| Top video’s share of views | 1 | Withheld | |
| Saves per view | 0 | Unavailable | |
| Shares per view | 0 | Unavailable | |
Percentages in the tables are ratios multiplied by 100. JSON and CSV preserve the original ratios and unrounded values. Followers are counts, not deduplicated people; views per follower is a ratio. The CSV includes a unit for every row.
Methodology preserved with this edition
1. Who is included
We use creators marked active in the OwlScran creator index who have at least one connected platform account with 2,000 followers or more. The floor applies separately on each platform, to every metric. Smaller accounts are excluded even if the creator qualifies elsewhere. Combined followers add only qualifying accounts. Active is an account status, not a measure of how often someone posts or whether they earn money. Platform metrics come from connected accounts. Each creator contributes at most one account-level value to each platform and metric; the sample size beside a result counts those contributions.
2. Engagement and post performance
Account engagement is average engagements divided by average views, multiplied by 100. Account averages use up to 20 eligible videos posted at least seven days ago. Interactions include likes and comments, with saves, shares or a platform interaction total where available. Our engagement methodology [1] explains the calculation. Post performance uses average views on TikTok and YouTube, and average accounts reached on Instagram and Facebook. Reach and views measure different things and should not be ranked against each other.
3. Views from non-followers
For each creator, we divide views labelled non-follower by all views returned in the follower-status breakdown, including any unknown category. Instagram uses account Insights; YouTube uses channel Analytics and subscription status at the time of the view [2]. Both request the preceding 30 days at retrieval. We then take the median of these account percentages. Repeat views count, and someone who does not follow an account may have watched it before. This measures distribution beyond followers, not new viewers or follower conversion.
4. Concentration, views per follower, saves and shares
These four analyses use exactly the 20 most recent stored videos posted at least seven days before the edition date. We use the latest stored snapshot available at computation for each video. Each snapshot must have been captured at least seven days after publication, with a view count on all 20. Videos marked unavailable by the integration cannot contribute stored counts. If any selected video is unavailable or fails the snapshot rule, the account does not contribute to these analyses; we do not substitute an older video. These are accumulated counts at different ages, not performance measured over a fixed seven-day period. The top-video share is the largest view count divided by their total views. Views per follower is their mean views divided by the platform follower count in the creator index. Save and share rates divide total saves or shares by total views, and require that interaction on every video. Zero-view totals cannot contribute to concentration or interaction rates. Views per follower can be zero. Saves are available through our Instagram integration; shares through Instagram, TikTok and YouTube. Missing values are never backfilled from older snapshots or replaced with zero.
5. Medians and ranges
Each creator has equal weight in the published median, regardless of their views or followers. The band runs from the 25th to the 75th percentile: the middle half of account values. It describes variation between creators, not a confidence interval or a forecast for the next post. Percentiles use linear interpolation, the type-7 method documented by R [3]. We do not remove large values just because they are unusual.
6. Missing values and publication threshold
Engagement cells include finite rates of zero or more; view-share cells include percentages from 0% to 100%. Missing values are excluded, not replaced with zero. Follower and post-performance distributions currently include only positive values. We withhold a result below the minimum sample shown on this page. That is a publication rule, not proof of representativeness, statistical precision or guaranteed anonymity. This edition introduces the 2,000-follower floor and includes measured zero engagement rates; earlier editions used a different cohort and excluded zero engagement.
7. Dates and limitations
Account refreshes are scheduled nightly and the page is cached for up to a day. The edition date is the computation date, not a shared observation period or the last successful sync of every metric. A failed split refresh can leave the previous value in place. This is a self-selected sample of OwlScran users. We do not control for niche, country, account size, content format or paid distribution. Available interactions differ by platform, so differences cannot establish that one platform performs better. The 20-video analyses favour accounts with a longer stored history. Videos differ in age, so older videos have had more time to accumulate views. Views and interactions can update at different speeds. The follower count is current, not the count when each video was posted. Changes between editions can reflect changes in the sample as well as performance.
These method notes were captured alongside the published computation. References to earlier editions describe the source methodology; this is the first fixed edition in this archive.
Sources and measurement definitions
- OwlScran. How we calculate engagement rate. Formulas, eligible videos and available interactions. Read the calculation method
- YouTube Analytics API. Subscribed status: whether the viewer was subscribed when the view occurred. Check the platform definition
- R documentation. Sample quantiles, including the type-7 interpolation used here. Check the percentile method
The JSON includes the captured measurement explanations, units, exclusions and references. External source pages may change after this edition.
Cite this edition
OwlScran Ltd (2026). The OwlScran Creator Index, fixed Q3 2026 edition (17 September 2026). https://www.owlscran.com/creator-benchmarks/editions/2026-09-17 (computed 2026-09-17T12:55:54.690Z).
Published aggregates are available under CC BY 4.0. Credit OwlScran and link to this edition. No account-level records are included.
Source provenance
Captured from the public JSON download on the Creator Index. SHA-256 of the original UTF-8 JSON download:
ef231c8f52d0de037a4a9802657bb92289252789bd15c0031ec03a99d53affc7Download the original source exportThe JSON download records this source hash and computation timestamp. The CSV is a table of those same aggregates.