How long does a short-form video keep earning views?

"TikToks keep earning for months, Reels die in 48 hours" is the kind of claim that circulates without evidence. Sometimes it is true for a particular account and niche. Whether it is true for your content is a question you can answer directly, and the answer changes how you plan campaigns.

The measurement

For each video, express the views it had at fixed checkpoints as a share of what it has now:

• share at 24 hours = views@24h ÷ current views
• share at 72 hours = views@72h ÷ current views
• share at 7 days = views@7d ÷ current views

Take the median across videos, not the mean — short-form distributions are heavily skewed and one runaway video drags an average anywhere.

The one rule that makes it valid

Only videos tracked from publication count.

If you added a video to tracking three weeks after it was posted, your first reading is already past every checkpoint. Including it either produces a nonsense share above 100% or, if you clamp it, silently inflates the early-share figure. Either way the curve becomes an artefact of when your team happened to add things.

Practically: filter to videos where the first snapshot is within roughly 24 hours of the publish time, and report the sample size next to the result. A curve built on 12 videos is a hint; on 200 it is a planning input.

Reading the shape

The gap between the 72-hour and 7-day shares is just as informative. A large gap means a second wave — recommendations picked the video up after the follower audience was exhausted. A tiny gap means whatever was going to happen happened on day one.

What changes once you know your curve

Settlement dates for creator payouts. If 85% of views arrive in the first 72 hours, a 30-day settlement adds a month of waiting for a few percent of the number. If only 45% arrive by then, a 14-day settlement systematically underpays creators — and they will notice.

When to amplify. Front-loaded curves mean paid support has to be triggered within hours, which requires frequent sampling and an alert path. See view velocity.

How long to keep videos on fast tracking. There is no point checking a video every hour once it is provably finished. Curve data tells you when to move each cohort to a slower cadence — which is how you keep collection costs sane at scale.

Reporting windows. A 7-day campaign report on content with a 30-day tail understates delivery. Match the window to the curve, and say which you used.

Segment before you conclude

One curve for everything hides the differences that matter. Split by:

• Platform — the three major feeds behave differently, and their view definitions differ too.
• Format — tutorials and listicles tend to have longer tails than reaction or trend-driven content.
• Creator size — large accounts front-load through their follower base; smaller accounts depend more on recommendations, which arrive later.
• Whether it was boosted — paid amplification rewrites the curve entirely. Analyse boosted content separately or the organic baseline is polluted.

The infrastructure this needs

Nothing exotic: timestamped readings from the moment of publication, kept indefinitely, with the publish time recorded. That is the same data that powers velocity alerts and defensible payout numbers — which is the argument for capturing it properly once rather than reconstructing it from screenshots later.