Module 1: YouTube Strategy
How the YouTube Algorithm Thinks
Many new creators imagine "the algorithm" as a mysterious judge that randomly decides which videos succeed. In practice, YouTube has explained the broad principles of its recommendation system publicly many times: it tries to match each viewer with the videos they are most likely to watch and be satisfied with. Once you understand that the system follows viewers, not creators, your strategy becomes clearer. Your job is to make videos that specific people want to click, keep watching and feel good about having watched.
Where views come from
Your videos can be discovered through several surfaces, each with slightly different signals:
| Surface | How viewers arrive | What matters most |
|---|---|---|
| Home feed | Personalised recommendations when someone opens YouTube | Viewer's history and interests; how similar viewers responded to your video |
| Suggested videos | Recommendations beside or after another video | Relevance to what the viewer just watched; topic and audience overlap |
| Search | Viewer types a query | Relevance of title, description and content to the query; how well the video satisfies searchers |
| Shorts feed | Swiping through vertical short videos | Whether viewers keep watching or swipe away; replays and engagement |
| Subscriptions, notifications, external | Subscribers, links shared on other sites and apps | Your existing audience and promotion |
YouTube Studio's analytics shows traffic sources for every video, so you can see which surfaces work for you.
The signals YouTube has described
YouTube has said its system considers signals such as:
- Clicks: whether people who see your thumbnail and title choose to watch. YouTube Studio reports this as impressions click-through rate.
- Watch time and retention: how long people watch and how much of the video they watch.
- Satisfaction: measured through surveys, likes, dislikes, shares and "not interested" feedback.
- Viewer context: what this particular viewer has watched, searched and engaged with before.
No single metric decides success. A high click rate with poor retention suggests a misleading title or thumbnail. Good retention with a low click rate suggests a strong video with weak packaging. YouTube has also said it does not simply favour long or short videos; it favours videos that viewers value.
What this means for your strategy
- Serve a defined audience. The system learns who enjoys your videos. If every video targets a completely different audience, it struggles to find the right viewers.
- Package honestly and well. A clear, compelling title and thumbnail earn the click; the video must deliver on that promise.
- Respect the viewer's time. Get to the point, structure the video clearly and cut slow sections.
- Build series and related videos so viewers who enjoy one video are recommended the next.
- Be patient and consistent. Many videos gain views gradually as the system tests them with more viewers, and search-based videos can grow over a long period.
Worked example: two cooking channels
Two creators in Nairobi start cooking channels. Channel A uploads whatever the creator feels like: a recipe one week, a phone review the next, a vlog about a wedding after that. Viewers who liked the recipe are not interested in the phone review, so each video is shown to a different, small audience, and growth stalls.
Channel B focuses on "affordable Kenyan dinners in 30 minutes". Every video has a clear title such as "3 Budget Githeri Recipes Ready in 30 Minutes", a thumbnail showing the finished dish, and a quick start with the finished meal and cost. Viewers who enjoy one video tend to enjoy others, so YouTube can confidently recommend Channel B's videos to similar viewers. Channel B's analytics show suggested videos becoming a major traffic source over time.
Common mistakes to avoid
- Chasing tricks such as uploading at a "magic time" or using hidden tags, rather than improving videos.
- Clickbait that wins the click but loses viewers in seconds, which damages satisfaction signals.
- Buying views or subscribers, which breaks YouTube's policies and fills your channel with people who will never watch.
- Judging a video after one day. Look at performance over weeks.
- Deleting and re-uploading videos repeatedly hoping for a different result.
Exercise
Choose three successful videos in the niche you want to work in. For each, write down the likely main traffic surface (search, suggested, home or Shorts), why someone would click the title and thumbnail, and how the first thirty seconds keep viewers watching. Then read YouTube's official explanation of how recommendations work (search 'How YouTube works recommendations') and note two points that change how you will plan videos.
Key points
- YouTube's recommendation system follows viewers, matching them with videos they are likely to watch and value.
- Views come from home, suggested, search, Shorts and subscriptions, each visible in YouTube Studio's traffic sources.
- Clicks, watch time, retention and satisfaction matter together; no single metric decides success.
- Serve a defined audience, package honestly, respect viewers' time and avoid shortcuts like bought views.