# What the algorithm actually “likes”

Course: [Short Video: Reels, TikTok, Shorts](https://the-app.fyi/academy/short-form-video)
Lesson 1 of 5 · 3–5 min
URL: https://the-app.fyi/academy/short-form-video/what-algorithm-likes

### The algorithm isn’t a fan — it’s a counter

The algorithm doesn’t “like” you personally. It maximizes the odds that the next viewer **won’t swipe** and will do something.

> It maximises the odds the next viewer **doesn’t swipe away** — and actually does something.

### Signals that beat the myths

1. **Retention** — how far they watched, where they dropped
2. **Rewatches**
3. **Saves / shares / comments** (not likes alone)
4. **Watching the key early seconds**
5. How similar the video is to what already worked for that person

### What usually isn’t a cheat code

- A secret hashtag
- Posting “only at 6:03 p.m.”
- The same sound as a millionaire account, with no meaning of your own
- Buying views

### Platform differences (very rough)

**TikTok** (from zero)

Often easier to break through from zero

**Reels** (living profile)

Leans harder on an already-alive profile + recommendations

**Shorts** (retention)

Its own ecosystem; retention and clickability matter

> **The main signal:** Takeaway: make a video people **finish**. Everything else is secondary.

## Check yourself

Which signal usually matters most for recommendations?

1. Number of hashtags
2. Retention and meaningful actions (watch-through, saves, comments)
3. Buying views in the first hour
4. A mandatory trending sound on everything

Explanation: Platforms push what people stick with and react to.