Learnings journal

What the data has and hasn't supported

Common YouTube creator beliefs, checked against one channel's tracked data — the current verdicts, the recent changes, and the full history of how each belief has fared over time.

Every verdict here is checked against @iamartharrison’s tracked data, through Aug 23, 2026.

Data throughAug 23, 2026

agreesthe belief matches the example channel's data
disagreesthe data runs the other way here
inconclusivethe sample is too thin to read
not currently testablethe data needed isn't shared

Common advice, checked here

Common creator advice, checked against this one channel's data.

These cards do not compare many channels. They only ask whether a common piece of creator advice matches this channel. Each card reports agrees, disagrees, inconclusive, or not currently testable.

Beliefinconclusive

“Older channels become less dependent on algorithmic recommendations.”

On this channel: correlation0.28 — the belief is inconclusive here.

How this is computed

r = 0.28· n = 5 — n = 5, low confidence, 0.7× its own threshold.

Across the shared pool. For each channel, the share of views over its most
recent 28 covered days that arrived from sources the viewer did not choose —
Browse, Suggested, Search, hashtags, Shorts, ads — set against the channel's
age in days. The complement (direct, external links, playlists, end screens,
the channel page, notifications) is where the viewer came back on their own,
the same split the public database already uses. Pearson correlation across
channels; agrees at r <= -0.4, disagrees at r >= 0.4.

Source: Maturation-creator folklore.

Beliefinconclusive
“Distribution happens in two phases — test (low impressions, narrow audience) then scale.”
On this channel: — the belief is inconclusive here.
How this is computed
— n = 0, low confidence.
```
Per video, fit a piecewise function to the impression-by-day curve.
Identify the 'elbow' where slope changes most. Agrees if ≥50% of
mature videos show a detectable elbow within days 3–14.
```
Source: Two-phase-algorithm folklore.
Beliefinconclusive
“Videos with average view percentage above 50% receive sustained impression delivery.”
On this channel: 0 — the belief is inconclusive here.
How this is computed
0 — n = 1, low confidence.
```
Per video, compare impression delivery rate in days 8–14 against days 1–7.
Bucket by retention tier (<30%, 30–50%, 50%+). Agrees if 50%+ bucket has
impression-rate ratio ≥0.7 AND lower-retention buckets ≤0.4.
```
Source: TubeBuddy creator guides.
Beliefnot currently testable
“Browse / Homepage impressions are weighted toward existing subscribers.”
Not testable on this channel right now — owner-DB-only data (subscribed-state / search-query tables live in the private owner view, not the public DB).
How this is computed
```
Browse views / Browse impressions broken out by subscribed state. Agrees
if subscribed-state share of Browse views ≥50%.
```
Source: Common assertion; conflicts with documented YouTube guidance.

Beliefdisagrees

“Browse pickups are more durable than Suggested-driven impressions.”

On this channel: 0.32 — the belief disagrees here.

How this is computed

0.32· agrees ≥ 1.50 — n = 4, low confidence, 2.8× its own threshold.

Across the shared pool, durability read as a half-life. Per channel and per
source: take the day that source delivered the most views, fit a line through
the log of the next 14 days, and read off how long a halving takes. A source
that is not fading over that stretch has no half-life and sits out —
reporting a cut-off window as a long half-life would flatter whichever source
happened to be climbing. The reading is the pool's median of (Browse
half-life over Suggested half-life); agrees at >= 1.5, disagrees at <= 1/1.5.
The shared data carries source views per day at the channel level, so this
is a channel-level durability read rather than the per-video one the belief
describes.

Source: Pickup-pattern folklore.

Beliefnot currently testable
“Cards contribute roughly 1% of traffic.”
Not testable yet — the data it needs hasn't been wired up here. Awaiting Round 4 cards-source data ingestion.
How this is computed
```
card_views / total_views. Cards are usually grouped under Direct/Unknown
or a specific source.
```
Source: YouTube official help docs.
Beliefinconclusive
“Channel-page-driven views come from highly engaged viewers.”
On this channel: 0.00 — the belief is inconclusive here.
How this is computed
0.00 — n = 0, low confidence.
```
Compare AVD, retention pct, sub conversion, comment rate of Channel Page
traffic vs Suggested traffic. Agrees if all four are higher for Channel Page.
```
Source: Common assertion.
Beliefinconclusive
“As library grows, Channel Page share of views falls.”
On this channel: 0.00 — the belief is inconclusive here.
How this is computed
0.00 — n = 0, low confidence.
```
Per channel, Pearson correlation between days_since_first_video and
trailing-7d Channel Page view share. Agrees if r ≤−0.5.
```
Source: Library-growth folklore.
Beliefinconclusive
“Comments-per-view tracks audience resonance better than likes-per-view.”
On this channel: 1of the 5 this test needs — the belief is inconclusive here.
How this is computed
n = 1· min 5 — n = 1, low confidence.
```
For each engagement metric, Pearson correlation with sub conversion rate
at the video level. Agrees if comments/v has the highest correlation.
```
Source: Engagement-depth folklore.
Beliefinconclusive
“Comments are weighted more than likes by the algorithm.”
On this channel: 0of the 8 this test needs — the belief is inconclusive here.
How this is computed
n = 0· min 8 — n = 0, low confidence.
```
For videos with comparable view counts, regress impression-rate-after-day-7
on (comments per view) and (likes per view) separately. Agrees if
|β_comments| > 2 × |β_likes|.
```
Source: Various creator coaching.

Beliefinconclusive

“High click-through-rate drives video success more than high retention.”

On this channel: — the belief is inconclusive here.

How this is computed

— n = 1, low confidence.

Per video, scale CTR and retention to channel-relative units (each
value's distance from the channel's median, divided by the channel's
typical spread). Fit a linear regression of total views against both.
If the magnitude of the CTR coefficient is at least 1.5× the retention
coefficient, the data agrees that CTR dominates on this channel.

Source: MrBeast school of creator advice; widely contested.

Beliefnot currently testable
“Neither click-rate nor retention alone matters; the product (or geometric mean) does.”
Not testable yet — the data it needs hasn't been wired up here. Awaiting Round 4 retention-curve-derived precision.
How this is computed
```
Define combined = sqrt(ctr_z × retention_z) per video. Regress total
views on combined vs on CTR and retention separately. Agrees if combined-
only regression has higher R² than either-alone.
```
Source: Combined-quality folklore.

Beliefnot currently testable

“Channels publishing daily need pre-recorded inventory; those without it burn out.”

Not testable on this channel right now — the size of each channel's unreleased upload queue — scheduled and unlisted videos.

How this is computed

Across the shared pool — but this belief is conditional, and the condition
is not in the data. It says daily publishers WITH pre-recorded inventory
hold their cadence and those WITHOUT it break, so splitting the pool needs
the size of each channel's unreleased queue: scheduled and unlisted videos.
Only published videos are ever shared, so that split cannot be made. How
often daily publishers break cadence IS measurable, but it answers a
different question — reporting it here would credit the break to a cause
nothing observed. The test names the field it needs and returns a verdict
the day that field exists.

Source: Common creator advice.

Beliefinconclusive
“YouTube's algorithm tests a video primarily in its first 1–3 days, after which performance is largely set.”
On this channel: — the belief is inconclusive here.
How this is computed
— n = 0, low confidence.
```
For each video with ≥14 days tracked, what fraction of total lifetime
views came from days 1–3? Channel median across qualifying videos.
```
Source: VidIQ and TubeBuddy guides; widely repeated in creator literature.
Beliefinconclusive
“Once 30 days post-publish, a video's daily view rate stabilizes and varies little.”
On this channel: — the belief is inconclusive here.
How this is computed
— n = 0, low confidence.
```
For videos ≥60 days old, coefficient of variation of daily views in
days 30–60. Channel median. Agrees if median CoV ≤0.5.
```
Source: Common analytics-tool framing.
Beliefinconclusive
“Editing title or thumbnail post-publish kills momentum.”
On this channel: 0.00 — the belief is inconclusive here.
How this is computed
0.00 — n = 0, low confidence.
```
Same as #12, but agrees if post-change views stay ≤50% of pre-change
baseline for 14+ days. Tests for sustained damage rather than reset+recovery.
```
Source: Conventional wisdom (the 'no-fiddle' rule).
Beliefnot currently testable
“Videos with high end-screen click-through have higher overall retention.”
Not testable yet — the data it needs hasn't been wired up here. Awaiting Round 4 cards/end-screen data ingestion.
How this is computed
```
Per video, end-screen view share of total views vs lifetime weighted
retention. Agrees if r ≥0.4. Requires Round 4 cards/end-screen data.
```
Source: End-screen optimization folklore.
Beliefnot currently testable
“End-screen impressions and clicks contribute 5–15% of traffic.”
Not testable yet — the data it needs hasn't been wired up here. Awaiting Round 4 retention/cards data ingestion.
How this is computed
```
end_screen_views / total_views channel-wide. Reads from per-source breakdown.
Note: requires retention curve / cards data from Round 4 to test fully.
```
Source: YouTube official help docs (creator-side guidance).
Beliefinconclusive
“Viewers from external sites have the highest engagement rates.”
On this channel: 0.00 — the belief is inconclusive here.
How this is computed
0.00 — n = 0, low confidence.
```
AVD, retention, sub conversion of External vs all other sources. Agrees
if External AVD ≥ p75 of cross-source AVD distribution. Sample-size flag
triggers when External views < 100 channel-wide.
```
Source: Common assertion.
Beliefinconclusive
“Days to first 100 subscribers is the slowest sub-acquisition leg.”
On this channel: 0 — the belief is inconclusive here.
How this is computed
0 — n = 0, low confidence.
```
Days from 1 sub to 100 subs vs days from 100 to 200, 200 to 300, etc.
Channel-relative growth rates per sub-bucket. Agrees if days-per-sub
in 0–100 is ≥2× days-per-sub in 100–500.
```
Source: Beginner-creator coaching.

Beliefinconclusive

“Performance trajectory of the first 10 videos predicts month-12 channel state.”

On this channel: 0 measured — the belief is inconclusive here.

How this is computed

n = 0 — n = 0, low confidence.

Across the shared pool. The belief never fixed a definition of "predicts",
so this one is stated plainly: the TRAJECTORY is the slope of first-week
views across a channel's first ten videos (rising, flat or falling), and the
month-12 STATE is that channel's total views one year after its first
upload. Pearson correlation across channels; agrees at r >= 0.5, disagrees
at r <= -0.5. A channel needs ten videos with a full first week and a shared
day on or after its own first anniversary; those short of either are counted
in the evidence rather than filled in.

Source: Long-arc folklore.

Beliefinconclusive
“A video's lifetime trajectory is largely determined by its first 24–48 hour performance.”
On this channel: 0of the 5 this test needs — the belief is inconclusive here.
How this is computed
n = 0· min 5 — n = 0, low confidence.
```
Pearson correlation of day1_views vs total_views across all videos with
≥30 days tracked. Same correlation for day1_ctr vs lifetime_weighted_ctr.
Higher r = belief agrees on this channel.
```
Source: Closely related to #2; cited as 'velocity matters'.

Beliefinconclusive

“First video's day-30 performance correlates with month-3 channel trajectory.”

On this channel: correlation0.28 — the belief is inconclusive here.

How this is computed

r = 0.28· n = 3 — n = 3, low confidence, 0.6× its own threshold.

Across the shared pool. For each channel: the views its FIRST video had
banked by day 30, against that channel's own total views over its first 90
days. Pearson correlation across channels; agrees at r >= 0.5, disagrees at
r <= -0.5. Both windows must lie wholly inside a channel's shared coverage —
a channel that shared only part of either window sits the round out rather
than contributing an understated total.

Source: TenfoldGrowth-style coaching.

Beliefinconclusive
“Publishing video N consumes impressions that would otherwise have gone to video N−1, slowing N−1's trajectory.”
On this channel: — the belief is inconclusive here.
How this is computed
— n = 1, low confidence.
```
For each video pair (N−1, N) where N−1 is still in days 1–7 of its
lifecycle when N publishes, compare N−1's view trajectory in the 3 days
after N's publish to N−1's trajectory in the 3 days before. Channel
median of (views_3d_after / views_3d_before) across eligible pairs.
```
Source: Common creator advice; cited across creator coaching channels.
Beliefnot currently testable
“The 0–15-second retention drop is the strongest single predictor of overall performance.”
Not testable yet — the data it needs hasn't been wired up here. Awaiting Round 4 retention curve ingestion.
How this is computed
```
From video retention curve, retention_15s_pct. Pearson correlation with
total views and weighted_ctr_post_day_3. Agrees if r ≥0.7. Requires
retention curve data (Round 4).
```
Source: Hook-economy folklore.
Beliefnot currently testable
“Hour-1 view rate predicts the video's lifetime performance.”
Not testable yet — the data it needs hasn't been wired up here. YouTube Reporting API does not expose hourly granularity.
How this is computed
```
Requires hourly data not exposed by the YouTube Reporting API.
```
Source: Velocity-creator folklore.
Beliefinconclusive
“Longer videos accumulate more total watch minutes per viewer.”
On this channel: 1of the 5 this test needs — the belief is inconclusive here.
How this is computed
n = 1· min 5 — n = 1, low confidence.
```
Pearson correlation of duration_seconds and (total_watch_minutes / total_views).
Agrees if r ≥0.5.
```
Source: Logical inference (bounded by retention).
Beliefinconclusive
“Videos 8–10 minutes long perform best for ad revenue and watch time.”
On this channel: 0 — the belief is inconclusive here.
How this is computed
0 — n = 0, low confidence.
```
Bucket videos by duration (60/180/360/600/1200/1800/3600+ sec). Median
weighted CTR, AVD, retention pct, sub conversion per bucket. Agrees if
600s bucket has the highest median weighted retention pct.
```
Source: Creator-economy folklore; pre-2018 mid-roll-ad threshold.
Beliefinconclusive
“A new publish steals impression budget from the rest of the catalog.”
On this channel: — the belief is inconclusive here.
How this is computed
— n = 1, low confidence.
```
When video N publishes, change in non-N videos' impression share over
the next 3 days. Agrees if median impression-share-of-others drops
≥15% in the post-publish window.
```
Source: Catalog-cannibalization folklore.

Beliefnot currently testable

“Channels with narrow topical focus outgrow broad-topic channels.”

Not testable on this channel right now — per-video topic signal (video titles); the shared data carries one category per channel, which cannot describe how varied that channel's own videos are.

How this is computed

Across the shared pool — but "narrow focus" has to be measured across a
channel's OWN videos, which needs per-video topic signal (titles are the
usual stand-in). The shared data carries one category per channel and no
per-video text, and a single label cannot describe variation inside the
channel it labels: two channels in one category may be a specialist and a
generalist. Grading by category and calling it focus would be an invented
measurement. The test names the field it needs; no number of extra channels
supplies it.

Source: Niche-down folklore.

Beliefnot currently testable
“Titles with numbers (5 Ways… / 3 Things…) perform better.”
Not testable on this channel right now — owner-DB-only data (subscribed-state / search-query tables live in the private owner view, not the public DB).
How this is computed
```
Median weighted CTR and median total views for title_has_number = 1 vs 0.
Agrees if both metrics ≥1.2× the non-numbered group.
```
Source: Listicle-era folklore.
Beliefinconclusive
“Late-afternoon publishing optimizes day-0 reach.”
On this channel: — the belief is inconclusive here.
How this is computed
— n = 2, low confidence.
```
Median day-1 views by hour-of-publish. Identify mode. Agrees if mode
falls in 14:00–18:00 local time. Many channels lock to one hour, making
this inconclusive.
```
Source: Publishing-time folklore.
Beliefinconclusive
“Channels that post on a consistent cadence outperform those that don't.”
On this channel: 2.00 — the belief is inconclusive here.
How this is computed
2.00 — n = 1, low confidence.
```
Cadence regularity score = the typical day-to-day spread of
inter-publish gaps, divided by the mean gap (coefficient of
variation). Lower = more consistent. Correlate against
trailing-30-day view growth.
```
Source: Universal creator-coaching advice.
Beliefinconclusive
“Publishing too often cannibalizes per-video performance.”
On this channel: — the belief is inconclusive here.
How this is computed
— n = 1, low confidence.
```
Same calculation as #1.
```
Source: Aliased to #1 frequency_cannibalization (different phrasing of same belief).
Beliefinconclusive
“Long gaps between uploads cause the algorithm to deprioritize the channel.”
On this channel: — the belief is inconclusive here.
How this is computed
— n = 0, low confidence.
```
For gaps of ≥5 days between publishes, compare the post-gap video's
day-1/day-3 performance against the channel's typical day-1/day-3.
```
Source: Common creator coaching.
Beliefnot currently testable
“Titles ending in ? outperform on click-through rate.”
Not testable on this channel right now — owner-DB-only data (subscribed-state / search-query tables live in the private owner view, not the public DB).
How this is computed
```
Median weighted CTR for videos where title_is_question = 1 vs 0.
Owner-DB only — not in the public DB (titles are coarsened in public).
```
Source: Title-optimization folklore.
Beliefinconclusive
“Once a channel has 30+ videos, key ratios (CTR, AVD, sub conversion) settle.”
On this channel: — the belief is inconclusive here.
How this is computed
— n = 2, low confidence.
```
Coefficient of variation of weekly weighted CTR and AVD across the
channel's history. Compare CoV in weeks 1–10 vs weeks 11+. Agrees if
both CoVs drop by ≥30% after the 30-video mark.
```
Source: Maturation folklore.
Beliefinconclusive
“High retention drives video success more than high click-through-rate.”
On this channel: — the belief is inconclusive here.
How this is computed
— n = 1, low confidence.
```
Same regression as #7; agrees if |β_retention| > 1.5 × |β_ctr|. Note:
can be `inconclusive` simultaneously with #7 if neither dominates.
```
Source: Sean Cannell / Think Media school; opposite of #7.
Beliefinconclusive
“Videos with growing Search-source share over time become long-term assets.”
On this channel: — the belief is inconclusive here.
How this is computed
— n = 0, low confidence.
```
Per video, slope of weekly Search-source view share over the video's
lifetime. Channel-level: share of mature videos with positive slope.
```
Source: SEO-creator advice.
Beliefinconclusive
“Shorter videos accumulate more views than longer ones.”
On this channel: 1of the 5 this test needs — the belief is inconclusive here.
How this is computed
n = 1· min 5 — n = 1, low confidence.
```
Pearson correlation of duration_seconds and total_views. Agrees if r ≤−0.3.
```
Source: Pre-Shorts-era folklore.
Beliefinconclusive
“Most subscribers come from a few breakout videos, not from many videos contributing equally.”
On this channel: — the belief is inconclusive here.
How this is computed
— n = 2, low confidence.
```
Gini coefficient on subs_gained distribution across videos. Agrees if
Gini ≥0.7.
```
Source: Power-law folklore.

Beliefagrees

“Channels with more subs per view have more engaged audiences.”

On this channel: correlation1.00 — the belief agrees here.

How this is computed

r = 1.00· n = 3 — n = 3, low confidence, 3.3× its own threshold.

Across the shared pool. Subscribers per view (lifetime subscribers over
lifetime views) against how far into a video the audience typically watches
— the median halfway-mark watch ratio across that channel's videos. Pearson
correlation across channels; agrees at r >= 0.3, disagrees at r <= -0.3.
The belief names three engagement reads; the shared data carries retention
only, and the evidence records comment rate and average view duration as
absent rather than substituting something else for them.

Source: Engagement-quality folklore.

Beliefnot currently testable
“Subscribed-state viewers have higher retention than non-subscribers.”
Not testable on this channel right now — owner-DB-only data (subscribed-state / search-query tables live in the private owner view, not the public DB).
How this is computed
```
AVD per (video, subscribed_state). Agrees if subscribed AVD median is
≥1.3× non-subscribed AVD.
```
Source: Common assertion.
Beliefnot currently testable
“Day-0 traffic is dominated by subscribed-state viewers.”
Not testable on this channel right now — owner-DB-only data (subscribed-state / search-query tables live in the private owner view, not the public DB).
How this is computed
```
Per-video, what % of day-0 views came from subscribed-state viewers?
Channel median. Reads from summary_video_subscribed (owner-DB only — not in the public DB).
```
Source: Common assertion in creator coaching.
Beliefnot currently testable
“Suggested impressions reach mostly non-subscribed viewers.”
Not testable on this channel right now — owner-DB-only data (subscribed-state / search-query tables live in the private owner view, not the public DB).
How this is computed
```
Suggested views per subscribed state. Agrees if non-subscribed share of
Suggested views ≥70%.
```
Source: Common assertion.

Beliefinconclusive

“Channels accelerate after 1,000 subscribers.”

On this channel: 5.00 — the belief is inconclusive here.

How this is computed

5.00 — n = 0, low confidence.

Across the shared pool. For each channel that crossed 1,000 subscribers
inside its own shared coverage: mean daily views over the 30 days AFTER the
crossing, divided by the mean over the 30 days before it. The reading is the
pool's median of those ratios; agrees at >= 1.5, disagrees at <= 1/1.5. Both
windows must lie inside coverage, so a channel already past 1,000 on its
first shared day has no crossing to read.

Source: Monetization-threshold folklore.

Beliefinconclusive
“Renaming a video resets its testing cycle.”
On this channel: 0.00 — the belief is inconclusive here.
How this is computed
0.00 — n = 0, low confidence.
```
For videos with a title change, compare daily views/impressions/CTR for
the 7 days before vs 7 days after the change. Agrees if median ratio
drops below 0.5 then recovers above 1.0 within 14 days.
```
Source: Common YouTube growth-channel advice.
Beliefinconclusive
“Most videos peak in their first three days.”
On this channel: — the belief is inconclusive here.
How this is computed
— n = 0, low confidence.
```
Distribution of peak_views_day across mature videos. Agrees if ≥60% of
videos peaked on days 0–2.
```
Source: Pre-2020 algorithmic-period folklore.

Beliefagrees

“YouTube ranks by session length, not per-video metrics.”

On this channel: correlation0.85 — the belief agrees here.

How this is computed

r = 0.85· n = 3 — n = 3, low confidence, 2.8× its own threshold.

Across the shared pool, on a stand-in: a session spans videos and often
channels, and nothing shared here can see one. The stand-in is the minutes a
typical view spends watching — each video's length times how far its
audience watched, weighted by first-week views — set against the channel's
recent view growth (its last 28 covered days over the 28 before them).
Pearson correlation across channels; agrees at r >= 0.3, disagrees at
r <= -0.3. Confidence stays low at every pool size: this stands in for
session length, it does not measure it.

Source: Session-watch folklore (post-2018 algorithm framing).

Beliefinconclusive
“Saturday and Sunday uploads get less engagement than weekday uploads.”
On this channel: 2 — the belief is inconclusive here.
How this is computed
2 — n = 2, low confidence.
```
Median weighted CTR, AVD, sub conversion for videos published on Sat/Sun
vs Mon–Fri. Agrees if all three are lower on weekend uploads. Inconclusive
with <10 weekend uploads.
```
Source: Day-of-week folklore.

Recent changes

Beliefs whose verdict moved in the last 90 days.

No belief changed verdict in the last 90 days.

A change is logged here only when a belief moves between agrees, disagrees, inconclusive, or not currently testable. A long, quiet stretch with no change is the common case on a single channel.

Belief history

Every belief on record, with how it has held or moved over time.

Each belief checked against one channel's tracked data, with its current verdict, confidence, number of changes, last change date, and current streak.
Row	Belief	Status	Confidence	Last changed	Streak
01	Older channels become less dependent on algorithmic recommendations.	inconclusive	low	—	2 days
02	Distribution happens in two phases — test (low impressions, narrow audience) then scale.	inconclusive	low	—	2 days
03	Videos with average view percentage above 50% receive sustained impression delivery.	inconclusive	low	—	2 days
04	Browse / Homepage impressions are weighted toward existing subscribers.	not currently testable	low	—	2 days
05	Browse pickups are more durable than Suggested-driven impressions.	disagrees	low	—	2 days
06	Cards contribute roughly 1% of traffic.	not currently testable	low	—	2 days
07	Channel-page-driven views come from highly engaged viewers.	inconclusive	low	—	2 days
08	As library grows, Channel Page share of views falls.	inconclusive	low	—	2 days
09	Comments-per-view tracks audience resonance better than likes-per-view.	inconclusive	low	—	2 days
10	Comments are weighted more than likes by the algorithm.	inconclusive	low	—	2 days
11	High click-through-rate drives video success more than high retention.	inconclusive	low	—	2 days
12	Neither click-rate nor retention alone matters; the product (or geometric mean) does.	not currently testable	low	—	2 days
13	Channels publishing daily need pre-recorded inventory; those without it burn out.	not currently testable	low	—	2 days
14	YouTube's algorithm tests a video primarily in its first 1–3 days, after which performance is largely set.	inconclusive	low	—	2 days
15	Once 30 days post-publish, a video's daily view rate stabilizes and varies little.	inconclusive	low	—	2 days
16	Editing title or thumbnail post-publish kills momentum.	inconclusive	low	—	2 days
17	Videos with high end-screen click-through have higher overall retention.	not currently testable	low	—	2 days
18	End-screen impressions and clicks contribute 5–15% of traffic.	not currently testable	low	—	2 days
19	Viewers from external sites have the highest engagement rates.	inconclusive	low	—	2 days
20	Days to first 100 subscribers is the slowest sub-acquisition leg.	inconclusive	low	—	2 days
21	Performance trajectory of the first 10 videos predicts month-12 channel state.	inconclusive	low	—	2 days
22	A video's lifetime trajectory is largely determined by its first 24–48 hour performance.	inconclusive	low	—	2 days
23	First video's day-30 performance correlates with month-3 channel trajectory.	inconclusive	low	—	2 days
24	Publishing video N consumes impressions that would otherwise have gone to video N−1, slowing N−1's trajectory.	inconclusive	low	—	2 days
25	The 0–15-second retention drop is the strongest single predictor of overall performance.	not currently testable	low	—	2 days
26	Hour-1 view rate predicts the video's lifetime performance.	not currently testable	low	—	2 days
27	Longer videos accumulate more total watch minutes per viewer.	inconclusive	low	—	2 days
28	Videos 8–10 minutes long perform best for ad revenue and watch time.	inconclusive	low	—	2 days
29	A new publish steals impression budget from the rest of the catalog.	inconclusive	low	—	2 days
30	Channels with narrow topical focus outgrow broad-topic channels.	not currently testable	low	—	2 days
31	Titles with numbers (5 Ways… / 3 Things…) perform better.	not currently testable	low	—	2 days
32	Late-afternoon publishing optimizes day-0 reach.	inconclusive	low	—	2 days
33	Channels that post on a consistent cadence outperform those that don't.	inconclusive	low	—	2 days
34	Publishing too often cannibalizes per-video performance.	inconclusive	low	—	2 days
35	Long gaps between uploads cause the algorithm to deprioritize the channel.	inconclusive	low	—	2 days
36	Titles ending in ? outperform on click-through rate.	not currently testable	low	—	2 days
37	Once a channel has 30+ videos, key ratios (CTR, AVD, sub conversion) settle.	inconclusive	low	—	2 days
38	High retention drives video success more than high click-through-rate.	inconclusive	low	—	2 days
39	Videos with growing Search-source share over time become long-term assets.	inconclusive	low	—	2 days
40	Shorter videos accumulate more views than longer ones.	inconclusive	low	—	2 days
41	Most subscribers come from a few breakout videos, not from many videos contributing equally.	inconclusive	low	—	2 days
42	Channels with more subs per view have more engaged audiences.	agrees	low	—	2 days
43	Subscribed-state viewers have higher retention than non-subscribers.	not currently testable	low	—	2 days
44	Day-0 traffic is dominated by subscribed-state viewers.	not currently testable	low	—	2 days
45	Suggested impressions reach mostly non-subscribed viewers.	not currently testable	low	—	2 days
46	Channels accelerate after 1,000 subscribers.	inconclusive	low	—	2 days
47	Renaming a video resets its testing cycle.	inconclusive	low	—	2 days
48	Most videos peak in their first three days.	inconclusive	low	—	2 days
49	YouTube ranks by session length, not per-video metrics.	agrees	low	—	2 days
50	Saturday and Sunday uploads get less engagement than weekday uploads.	inconclusive	low	—	2 days

Have a YouTube belief you want tested? or email it directly. Useful submissions get added to the canon — credit appears in the card's source line.

Keep reading

Where the verdicts come from

Each verdict is a check against one channel's tracked data, not a comparison across many channels. These surfaces describe what the data can and can't say behind the journal.

Limits registry

What is missing, coarse, or approximate in the data — and what that means when reading a verdict.

Known weaknesses

The strongest good-faith arguments against the inferences this dashboard makes, with the guardrail in the code.