Updated Oct 11: 688 view-growth comparisons crossing YouTube's Aug 24 counting change were excluded and the main view-growth figures were recalculated. Details
A recent survey of 311 Japanese VTubers found that about 8 in 10 had kept streaming while feeling unwell. Behind that is a familiar fear: "if you take a break, your channel stops growing." Is that actually what happens? We looked at the YouTube activity history of the VTuber channels TubeSaku tracks and pulled out 20,002 breaks of 7 days or more with no uploads or streams.
VTubers who got back to their own pace after a break roughly maintained their view-growth pace.
After excluding 688 comparisons whose 28-day windows crossed YouTube's Aug 24 counting change, 1,265 clean comparisons remained. Overall median post-break view growth was 60.0% of pre-break. In the 246 cases where the creator returned to at least their pre-break number of active days, median view growth was 101.6%. In the 1,019 cases where activity stayed lower, it was 50.7%. What looks like "the algorithm punishing breaks" may instead reflect, at least in part, creators being unable to return to a pace that was hard to sustain.
TubeSaku analyzed 20,002 breaks across 9,006 VTuber channels. For view growth, we excluded 688 comparisons whose 28-day windows crossed YouTube's Aug 24 counting change, leaving 1,265 clean comparisons. Where creators returned to at least their pre-break number of active days (246 breaks), median post-break view growth was 101.6% of pre-break. Where activity stayed lower (1,019 breaks), it was 50.7%. The same pattern held at 4–5 days a week. The data did not support a simple "breaks cause growth to drop" explanation. This is observational data and does not establish causation. Source: TubeSaku "VTuber Hiatus Study 2026" https://tubesaku.com/research/vtuber-hiatuses-research-2026_en/
On Sep 28, 2026, haico Online Clinic (uyet Inc.) published a survey of 311 VTubers and V-livers in Japan. About 70% said they had experienced mental health problems related to streaming, and about 80% had kept streaming while feeling unwell. Many saw taking time off as a risk to their activity.
YouTube's own Help Center says that, looking at channels that took breaks, it found no consistent relationship between the length of a break and the change in views afterwards, while noting that audience behavior can change during a break.
So instead of a survey, we used actual activity logs, focused on VTubers who had been streaming or uploading at a high pace, and compared activity and view growth before and after each break.
We used periodic snapshots of each channel's total view count to compare views gained in the 28 days before the break with the 28 days after returning. Of the old 1,953 usable comparisons, 688 crossed YouTube's Aug 24 counting change and were excluded, leaving 1,265 breaks across 1,193 channels for the main view-growth analysis.
| Break length | Breaks | View growth after return (median, before = 100%) |
|---|---|---|
| All | 1,265 | 60.0% |
| 7–13 days | 1,030 | 62.4% |
| 14–29 days | 188 | 53.5% |
| 30–59 days | 38 | 40.3% |
Longer breaks went with lower ratios. That differs from YouTube's statement, but we measured something different: we started from periods when creators were especially active. Growth may also have slowed first and prompted the break, so this can't be read as "the break caused the drop."
The 30–59 day group has only 38 breaks; treat it as indicative.
We split the 1,265 clean comparisons into those where the creator returned to at least their pre-break number of active days, and those where activity stayed lower.
Median view growth in the 28 days after returning, with the 28 days before = 100%. Bar length is relative to 130%.
| Activity after return | Breaks | Active days (median, before → after) | View growth |
|---|---|---|---|
| Same as before or more | 246 (19.4%) | 18 → 21 | 101.6% |
| Less than before | 1,019 (80.6%) | 18 → 12 | 50.7% |
Among the 64 breaks where creators returned to exactly the same number of active days, median view growth was 95.4%. In the 182 breaks where activity was higher than before, it was 108.3%. The 101.6% figure is not driven only by a handful of extreme ramp-ups.
This does not prove that "getting your activity back makes your views grow." Channels that were already doing well may simply have found it easier to keep going, and things like health, motivation and ideas for new content can affect both activity and views.
| Pace before the break | Returned | Didn't return |
|---|---|---|
| 4–5 days/week | 101.2% (168) | 50.7% (626) |
| 5–6 days/week | 100.2% (53) | 52.5% (268) |
| 6–7 days/week | 104.5% (25, indicative) | 49.6% (125) |
The pattern held regardless of the starting pace. What separated the groups wasn't how often they streamed, but whether they could get back to their own pace.
To avoid counting the same channel twice, we used only each channel's first observed break of 7–13 days.
Share of channels with fewer active days in the 28 days after returning than in the 28 days before.
| Group | Channels | Fewer active days | Active days (median, before → after) |
|---|---|---|---|
| First 1–2 week break (all) | 7,080 | 77.2% (5,464) | 18 → 14 |
| Of which 6–7 days/week before | 803 | 80.4% (646) | 26 → 20 |
This doesn't mean "the break made them unable to stream." They may have gone on break because the pace was already becoming hard to keep. And because we start from a particularly active period, some decline is expected even without a break (regression to the mean). A near-daily pace may simply have been hard to sustain from the start.
We reviewed three channels by hand. All three started content aimed at viewers outside their existing fan base.
These are individual cases, not a statistical trend. None of them went back to their old level of activity; they changed what they offered. Where a comparison spans YouTube's Aug 24, 2026 view-count change, it includes that effect (about +20%).
This is not "never take a break" or "stream more."
Eight in ten creators who were running at 6–7 days a week couldn't get back to that pace, which suggests it was hard to sustain in the first place. Meanwhile, creators at 4–5 days a week who returned to their own pace kept growing, and some grew with less activity by trying something new. Some breaks happen because someone hit their limit; others are planned, followed by a fresh start.
We found no support for a simple "breaks cause growth to drop" story. Creators who got back to their own pace kept growing.
4–5 days a week worked just as well for those who returned to it. It's better to design a schedule you can return to after resting than to hold onto a pace that breaks you.
Even with streams and uploads paused, a short YouTube community post (text, image or poll) can share how you're doing and when you plan to be back. Posts can appear on subscribers' Home and Subscriptions feeds, and image posts can appear in the Shorts feed. It doesn't need to be an obligation; one line is enough.
When you come back, try a new format or kind of content. Channels that rely only on a small circle of regulars tend to feel the most pressure not to rest. Reaching new audiences helps growth, and it also makes it easier to take time off.
Public upload lists can miss members-only videos and streams, which could make an active period look like a break.
We collected additional data that captures members-only activity and checked 306,781 extra video IDs on the channels involved. Of the 21,727 breaks in our original analysis, 7.94% turned out to have activity during the "break," and in 4.36% the break no longer qualified. We rebuilt the activity history and reran the analysis; the members-only correction itself barely changed the main patterns. View-growth metrics were then recalculated again after excluding comparisons that crossed the Aug 24 counting change.
We only used metadata such as video IDs. We did not view or use any members-only content.
When first published (Oct 8, 2026), this report and our press release said that breaks starting in August 2026 (401 breaks) showed median view growth of 108.7% in the 28 days after returning.
However, the post-break period for that group falls after YouTube's Aug 24, 2026 change to how public view counts are tallied, which our separate analysis estimated at roughly +20%. When we reran the analysis using only breaks whose post-return window ended before Aug 24, the pattern did not hold.
| Window (before and after) | Breaks | View growth (median, before = 100%) | Share at or above pre-break |
|---|---|---|---|
| 7 days | 130 | 72.9% | 29.2% |
| 10 days | 64 | 72.4% | 31.3% |
| 14 days | 16 | 78.7% | 37.5% |
We therefore removed the August claim and recalculated the main view-growth analysis. Of the old 1,953 usable comparisons, 688 (35.2%) crossed the Aug 24 change and were excluded, leaving 1,265 clean comparisons. The 688 excluded comparisons include all 401 August-start breaks plus 287 breaks that started earlier but whose post-return 28-day window crossed Aug 24.
Because the rerun only includes windows that end before Aug 24, it uses 7–14 day windows instead of 28 days and skews toward creators who returned quickly. Not all of the gap between 108.7% and 72.9% is necessarily due to the view-count change.
| Main 28-day comparison | Breaks | View growth (median, before = 100%) |
|---|---|---|
| Activity returned to pre-break level or higher | 246 | 101.6% |
| Activity stayed below pre-break level | 1,019 | 50.7% |
After excluding the 688 cross-regime comparisons, the same qualitative split remained: view growth was roughly maintained where activity returned, and about half of pre-break growth where it did not.
We also corrected a channel-count typo in the previous page: the 1,953 usable events came from 1,770 unique channels, not 3,870. The revised main analysis contains 1,265 events across 1,193 channels.
Aggregated figures used in this report are available as CSV (column names in English, labels in Japanese). Free to use for articles, videos and research with attribution to this page. Intended license: CC BY 4.0.
CSVView growth by whether activity returnedCSV files contain aggregates only, with no data that identifies individual channels or videos.