Most videos under-perform for boring reasons: no chapters, keyword in the wrong place, a thumbnail that repeats the title, an upload at 3am, no end screen. The 2026 Algorithm Checklist turns the current set of recommendation signals into 25 pass/fail items for one video, orders the failures by impact, and estimates the boost from fixing them all.
This guide explains what the 25 items cover, how to use the checklist as a pre-publish gate, and which items are worth failing on purpose.
What the 25 cover
Five groups: session-start signals (searchable topic, notification-ready timing, external link plan), Browse eligibility (thumbnail clarity, title length and keyword position, no repeated text), topical authority (fits your cluster, playlist assigned, links to your own videos), engagement velocity (pinned comment planned, question in the video, Community teaser), and satisfaction (hook under 8 seconds, chapters, no misleading claim, end screen set). Each is a yes or no.
Pre-publish gate
Run it with the draft title, description, thumbnail description and structure. Twenty or more passes means publish. Fewer: the priority order tells you which failures matter most; the top three are usually fixable in ten minutes. The estimated boost is a rough percentage improvement in first-week reach from full compliance.
Failing on purpose
Some items conflict with a deliberate choice — an off-cluster experiment, a controversial title, a video published at an odd hour for a live event. That is fine. The checklist tells you the cost so the choice is informed; it does not tell you never to make it.
Tips
- Run it on every upload for a month; the same three items fail repeatedly and become habit fixes.
- Pair with the Algorithm Score Predictor when a video matters — the checklist is fast, the predictor is deep.
- Keep the output as your upload checklist in Studio's notes.
Common mistakes
- Treating 25/25 as the goal — 22 with good content beats 25 with weak content.
- Running it after publishing — Most items are cheap before upload and expensive after.
- Ignoring the order — The first failure listed is worth more than the last five combined.
Worked example
Input: "Title: Budget Grinder Comparison; home espresso; Thursday 3pm; chapters yes; end screen not set; no pinned comment planned."
What the report returned: 19/25. Top failures: keyword not in first 3 words (title), no pinned comment, end screen unset, thumbnail text repeats title. Est. boost from full compliance: +18–25% first-week reach.
Frequently asked questions
Is this based on YouTube's actual algorithm?
It is based on the publicly documented and widely observed signals as of 2026 — session starts, satisfaction, engagement velocity, topical consistency — translated into checkable items.
How often does the checklist change?
It is reviewed as YouTube's documented signals change; the items are stable for months at a time.
Can it check a published video?
Yes by URL, though several items (pinned comment, end screen, teaser) need you to confirm.
How is this different from the SEO Score Checker?
SEO covers metadata only. This covers the full set of distribution signals including timing, engagement and structure.