Datasets:
id string | metric string | platform string | period string | value float64 | unit string | qualifier string | source_id string | confidence_label string | notes string |
|---|---|---|---|---|---|---|---|---|---|
twitch_hours_q3_2026 | hours_watched | Twitch | 2026-Q3 | 4,510,000,000 | hours | null | S12 | primary | Up 4% on Q2 2026; strongest since Q2 2025 |
twitch_tracked_share_q3_2026 | share_of_tracked_hours | Twitch | 2026-Q3 | 14.2 | percent | null | S13 | secondary | Denominator includes YouTube and TikTok LIVE |
kick_hours_q3_2026 | hours_watched | Kick | 2026-Q3 | 2,040,000,000 | hours | null | S12 | primary | Up 30.7% on Q2 2026 |
global_hours_q3_2026 | hours_watched | All tracked platforms | 2026-Q3 | 31,698,000,000 | hours | null | S12 | primary | Highest quarterly total recorded by Streams Charts |
twitch_hours_q1_2026 | hours_watched | Twitch | 2026-Q1 | 4,550,000,000 | hours | null | S15 | secondary | Stream Hatchet via secondary compilation |
kick_hours_q1_2026 | hours_watched | Kick | 2026-Q1 | 1,270,000,000 | hours | null | S15 | secondary | null |
twitch_hours_q4_2025 | hours_watched | Twitch | 2025-Q4 | 4,380,000,000 | hours | null | S15 | secondary | null |
twitch_hours_q3_2025 | hours_watched | Twitch | 2025-Q3 | 4,600,000,000 | hours | null | S16 | secondary | null |
twitch_hours_q2_2025_initial | hours_watched | Twitch | 2025-Q2 | 4,900,000,000 | hours | null | S11 | secondary | Initial Stream Hatchet report; lowest since Q3 2020 |
twitch_hours_q2_2025_later | hours_watched | Twitch | 2025-Q2 | 4,640,000,000 | hours | null | S16 | secondary | Later compilation of the same provider's series |
kick_hours_q2_2025 | hours_watched | Kick | 2025-Q2 | 983,000,000 | hours | null | S11 | secondary | null |
twitch_hours_q1_2025 | hours_watched | Twitch | 2025-Q1 | 5,270,000,000 | hours | null | S15 | secondary | null |
twitch_gaming_share_q2_2025 | market_share | Twitch | 2025-Q2 | 54.2 | percent | null | S11 | secondary | Share of tracked gaming platforms by hours |
twitch_ccv_drop_2025_08_22 | viewership_change | Twitch | 2025-08-22 | -24 | percent | null | S7 | secondary | Concurrent viewership vs previous week |
twitch_ccv_drop_2025_08_24 | viewership_change | Twitch | 2025-08-24 | -23 | percent | null | S7 | secondary | Concurrent viewership vs previous week |
twitch_decline_range_low | viewership_decline_range_low | Twitch | 2025-08 | 5 | percent | null | S6 | secondary | Analyst Zach Bussey; varies by time of day |
twitch_decline_range_high | viewership_decline_range_high | Twitch | 2025-08 | 22 | percent | null | S6 | secondary | Analyst Zach Bussey; varies by time of day |
twitch_channels_suspicious_q2_2025 | channels_with_suspicious_stream | Twitch | 2025-Q2 | 41,000 | channels | > | S2 | primary | Channels with 50+ average viewers and at least one suspicious stream |
twitch_streamers_persistent_q2_2025 | streamers_persistent_viewbotting | Twitch | 2025-Q2 | 4,400 | streamers | > | S2 | primary | null |
twitch_streamers_persistent_pct_q2_2025 | streamers_persistent_viewbotting_share | Twitch | 2025-Q2 | 10.8 | percent | null | S2 | primary | Denominator not stated identically across sources; may be measured against flagged channels |
twitch_fake_hours_q2_2025 | fake_hours_watched | Twitch | 2025-Q2 | 30,000,000 | hours | > | S2 | primary | Generated by suspicious streams |
analytics_users_viewbotting_pct | streamers_using_analytics_who_viewbot | Twitch and Kick | 2025-Q2 | 10 | percent | null | S2 | primary | null |
kick_streamers_flagged_pct_q2_2025 | streamers_flagged_share | Kick | 2025-Q2 | 16 | percent | > | S3 | primary | Streamers averaging 50+ live viewers |
kick_confirmed_viewbotters_q2_2025 | confirmed_viewbotters | Kick | 2025-Q2 | 3,000 | streamers | ~ | S3 | primary | Number tripled in Q2 2025 |
kick_fake_hours_q2_2025 | fake_hours_watched | Kick | 2025-Q2 | 20,000,000 | hours | ~ | S2 | primary | Attributed to botted streams |
twitch_hours_aug_2025 | hours_watched | Twitch | 2025-08 | 1,410,000,000 | hours | null | S10 | secondary | Down 9% on July |
twitch_peak_ccv_change_aug_2025 | peak_concurrent_viewership_change | Twitch | 2025-08 | -73 | percent | null | S8 | secondary | Month on month |
twitch_avg_ccv_aug_2025 | average_concurrent_viewers | Twitch | 2025-08 | 1,900,000 | viewers | ~ | S9 | secondary | Down 8% on July |
web_malicious_bot_share_2024 | malicious_bot_share_of_web_traffic | Web (all) | 2024 | 37 | percent | null | S21 | secondary | Not Twitch-specific; vendor page |
twitch_top10_hours_share_q3_2026 | top_channel_hours_share | Twitch | 2026-Q3 | 4.8 | percent | null | S14 | primary | Top 10 channels |
twitch_top100_hours_share_q3_2026 | top_channel_hours_share | Twitch | 2026-Q3 | 17 | percent | null | S14 | primary | Top 100 channels |
twitch_top1000_hours_share_q3_2026 | top_channel_hours_share | Twitch | 2026-Q3 | 40.2 | percent | null | S13 | secondary | Top 1,000 channels |
twitch_active_channels_q3_2026 | channels_with_watch_time | Twitch | 2026-Q3 | 5,800,000 | channels | ~ | S14 | primary | null |
Twitch Bot Viewership and Viewbotting Measurements (2025-2026)
Dataset summary
Twitch publishes no audited figure for the share of its viewers that are bots. This dataset collects the published measurements that exist, with one row per statistic, a source ID, and a confidence label, so that readers can see which metric each percentage describes (concurrent viewers, channels, or hours).
It accompanies the research note What Percent of Twitch Viewers Are Bots? Evidence and Estimates (2026) by Abigail Says (BriefLedger), last updated 4 November 2026.
Files
| File | Rows | Description |
|---|---|---|
measurements.csv |
33 | One row per reported statistic (hours watched, viewership drops, viewbotting channel counts, Kick comparisons, audience concentration) |
derived_estimates.csv |
3 | Fake watch time as a share of total hours, calculated from published figures |
timeline.csv |
9 | Dated events from Q2 2025 to October 2026 |
sources.csv |
23 | Source register: publisher, title, date, URL, source type |
unverified_claims.csv |
4 | Widely repeated claims with no traceable primary source |
Columns
measurements.csv
id: unique row identifiermetric: what is measuredplatform: Twitch, Kick, all tracked platforms, or web (all)period: quarter, month or date the figure coversvalue: numeric valueunit: hours, percent, channels, streamers, viewersqualifier:>(greater than),~(approximate), or emptysource_id: key intosources.csvconfidence_label:primary(published by the producing organisation),secondary(reported by an outlet citing another party)notes: caveats
derived_estimates.csv
fake_hours,total_hours,fake_share_percent: calculated as fake hours divided by total hoursconfidence_label: alwaysderived
Key figures
- Sitewide concurrent viewership fell by 24% (22 Aug 2025) and 23% (24 Aug 2025) after detection tools were deployed; the decline ranged from 5% to 22% by time of day.
- Q2 2025: over 41,000 Twitch channels with 50+ average viewers had a suspicious stream; over 4,400 streamers (10.8%) showed persistent viewbotting; suspicious streams generated 30 million+ fake hours.
- Derived fake share of Twitch hours in Q2 2025: about 0.61% to 0.65%. Kick: about 2.0%.
- Twitch hours watched in Q3 2026: 4.51 billion.
Intended uses
Fact-checking claims about Twitch bot percentages, teaching examples on metric definitions, and as a starting point for tracking updated measurements.
Limitations
- Most audience figures are secondary. No primary dataset from Twitch was available.
- The 10.8% persistent-viewbotting share may be measured against flagged channels rather than all 50+ viewer channels; press coverage states the denominator differently.
- Derived estimates combine two providers (Streams Charts numerators, Stream Hatchet denominators) and are indicative only.
- Q3 2026 hours come from Streams Charts; earlier quarters come from Stream Hatchet, so rows are not directly comparable across providers.
- Rows in
unverified_claims.csvcome from companies that sell or analyse viewbots and are not evidence of the bot share. - Do not add rows together; they measure different things.
Source and licensing
Figures are facts reported by the third-party sources listed in sources.csv. The compilation, labels and derived estimates are released under CC BY 4.0. Please cite the original sources for individual statistics.
Citation
Says, A. (2026). Twitch Bot Viewership and Viewbotting Measurements (2025-2026) [Dataset]. BriefLedger. Author profile: https://briefledger.co.uk/author/abi-gail/ also check Statistics https://briefledger.co.uk/twitch-statistics/
- Downloads last month
- 86