YYoung and Profiting
← All frameworks
MarketingSean Cannell

People-First YouTube Algorithm

Optimise discovery, attention, and satisfaction for real viewers

Difficulty
Advanced
Time to result
~ongoing to results
Steps
6
Confidence
99%

The People-First YouTube Algorithm combines four main inputs: previous viewer behaviour, click-through response, watch time, and viewer satisfaction. YouTube first matches content to clusters of interests inferred from what people already watch, making a clear niche and relevant neighbouring channels useful. Topic, title, and thumbnail then compete for the click. Once viewers arrive, average duration, percentage viewed, retention, and continued viewing reveal whether the content holds attention. Finally, surveys, return behaviour, likes, comments, and bingeing help indicate satisfaction. No isolated metric is universal: click-through rate often falls as distribution broadens, while a long video can succeed with a lower completion percentage but high minutes watched. The durable rule is to replace “algorithm” with “people” and optimise the entire experience.

Origin

Sean Cannell organises YouTube's disclosed signals into four factors, then simplifies the system by telling creators to replace the word algorithm with people.

Core principles

  • 01The algorithm models people rather than replacing them
  • 02Viewer history creates niche recommendation clusters
  • 03Topic, title, and thumbnail earn the click
  • 04Watch time and satisfaction sustain distribution
  • 05Returning and bingeing viewers signal durable value

How to run it

  1. 1

    Map viewer behaviour

    Identify the topics, channels, and adjacent interests the intended audience already consumes.

    Pro tip Competition can enlarge the recommendation pool rather than merely crowd it.

    Watch out Do not chase an unrelated trend that confuses the channel's audience associations.

  2. 2

    Earn the click

    Select a timely, relevant topic and package it with a compelling title and thumbnail.

    Pro tip Treat topic as the foundation; packaging cannot rescue weak relevance indefinitely.

    Watch out A high click-through rate on very few impressions is not proof of broad appeal.

  3. 3

    Hold attention

    Open for a new viewer, deliver efficiently, and use retention curves to identify dips and excess length.

    Pro tip Add one extra round of revision to remove fluff.

    Watch out Do not speak only to returning fans when the goal is cold discovery.

  4. 4

    Measure minutes intelligently

    Consider both average percentage viewed and absolute average duration in the context of video length.

    Pro tip A strong long-form video can create exceptional total watch time despite modest completion.

    Watch out Avoid applying one retention threshold to every format.

  5. 5

    Deliver satisfaction

    Pay off the title and leave viewers feeling that the session was worthwhile.

    Pro tip Use comments, surveys, return viewers, and qualitative feedback together.

    Watch out Retention tricks that delay a weak payoff can produce watch time but dissatisfaction.

  6. 6

    Create continued viewing

    Connect related videos with series, end cards, and a coherent library so viewers can keep solving the same problem.

    Pro tip Track average videos viewed per viewer where available.

    Watch out Do not force a next video that is unrelated to the viewer's current goal.

In the wild

First-time marathon runner

A woman searching for marathon training begins receiving related shoe, nutrition, and coaching videos. A creator whose title earns her click, whose lesson helps her train, and whose library supports the next questions becomes a repeated recommendation for similar runners.

Discovery expands because the creator satisfies a real cluster of viewer needs rather than gaming one metric.

Common mistakes

Optimising one metric

A high click-through rate, completion rate, or view count has little meaning without impression scale, length, and satisfaction context.

Delaying a weak payoff

Keeping viewers waiting can raise watch time while making them regret the viewing session.

Copying a famous creator

Viewer-history matching rewards relevance within a niche, not imitation of the largest channel.

Is it for you?

Best for

It is best for creators using YouTube analytics to improve an educational or expertise-led channel over time.

Not ideal for

It is not ideal for anyone seeking a guaranteed threshold or shortcut independent of audience value.

From the transcript

bottom line YouTube has two major metrics click through eight in watch time

Sean Cannell · 38:00

the fourth one almost nobody knows about and it's viewer satisfaction

Sean Cannell · 46:00

replace the word algorithm with people you're not optimizing for the algorithm you're optimizing for people

Sean Cannell · 49:30

From the episode

Sean Cannell: Start a Profitable YouTube Channel in 2024

Sean Cannell