Why Your Faceless YouTube Shorts Get No Views (It's Retention, Not the Topic)

You picked a good topic. You wrote a real script. You published the Short. And it got 84 views and stopped.

If that’s the loop you’re stuck in, the problem almost certainly isn’t your topic. It’s retention — how much of each Short people actually watch before they swipe. That single metric decides whether YouTube pushes your video to more people or quietly buries it, and it’s the thing most faceless channels never look at.

I can say this with some confidence because we run a faceless Shorts channel that publishes every day, fully automated, and we’ve stared at a lot of retention graphs. The pattern is boringly consistent, and it’s worth understanding before you make another video.

the metric that actually matters

Open YouTube Studio and sort your Shorts by average view duration, not by views. That’s the honest scoreboard.

What you’ll usually find: the Shorts that got pushed held a higher percentage of their length, and the ones that flopped show a cliff in the first few seconds — people arriving and leaving almost immediately. On our own channel, the average view sits in the mid-teens of seconds on ~40-second Shorts. That gap between “how long it is” and “how long it’s watched” is the whole game. Close it and videos travel; leave it open and they don’t.

Views are the result. Retention is the cause. Chasing views by changing topics is treating the symptom.

why faceless Shorts leak viewers

Faceless content has a specific weakness: there’s no face to hold attention, so the words and pacing have to do all the work a presenter’s charisma normally does. Three things quietly kill retention on faceless Shorts:

• Opening with setup instead of the payoff. “Today I want to talk about…” is a swipe. The viewer decides in about two seconds whether to stay, and context-first openings spend those two seconds earning nothing.
• A flat stretch in the middle. Any second that just restates the last one is a place for a thumb to leave. Retention graphs almost always dip in the same spot: the lull after the intro, before the payoff.
• Spoken-only numbers. A figure said out loud but never shown on screen doesn’t register, so the “wow” moment lands soft.

Notice none of those are about the topic. They’re about structure.

the honest fix (the part you can do today)

Here’s the free, do-it-yourself version — enough to genuinely improve your next Short:

1. Open cold on the single most surprising claim. No intro, no channel bumper, no “in this video.” The first line should be the thing that makes someone stop scrolling — a specific number or an implied-impossible statement. If your best line is currently at second 20, move it to second 0.
2. Pay off partially inside the first 8 seconds. Don’t hold the reveal for the end. Give a real down-payment on the promise early, so the viewer has a reason to believe the rest is worth staying for.
3. Make every second carry a new fact. Read your script and delete any line that doesn’t add a concrete, new detail. Density is retention.

Do just those three things and watch your average view duration on the next few uploads. That’s the honest test — your own retention graph, not a guru’s screenshot.

You can prototype the whole thing for free, too: our stickman Shorts generator turns a topic into an animated faceless video right in your browser, so you can test a hook structure without a render pipeline.

the part that takes reps

Everything above is the diagnosis — the what and the why, and it’s the important half. The part that takes practice is the craft underneath it: finding the exact cold open that holds, knowing where to plant a mid-video re-hook so the middle doesn’t sag, and picking topics that structurally over-retain instead of ones that merely sound interesting. None of that is secret — it’s just reps, and a willingness to read your own retention graph honestly after every upload.

We learn it the same way, in public. If you want the honest numbers behind all of this — which of our own Shorts held and which died, and why — the build log of the channel that taught us this shows the zeros too.