Shorts against long cuts, and how to read the numbers honestly

Different physics, so different maths

A short is distributed by a feed that decides within hours. A long cut is distributed by browse and search and can still be accumulating views eight months later. Their view curves do not have the same shape, and the same total on both means two entirely different things.

The consequence: any channel-level average that mixes them is dominated by whichever format you happened to publish more of that month. It will move when your mix moves and you will read that as performance changing.

Compare the curve, not the total

The useful comparison is the shape. For each format, look at what fraction of the video’s current views arrived in the first 48 hours. A short is typically almost all of it. A long cut that is also almost all of it did not find browse distribution and is effectively finished. A long cut still adding meaningful views in week three is worth studying.

You cannot get this from a lifetime total, which is why any tool that only shows totals cannot answer the question. It needs a daily series, which means sampling every video every day from the moment you start.

Divide by what it cost you

If a long cut takes eight hours and a short takes forty minutes, a long cut needs roughly twelve times the return to be the better use of the day. Write down your real edit times for a month. Most creators find the honest ratio is not what they assumed, in both directions.

Eight videos, minimum

Below about eight videos in a bucket, one outlier moves the median enough to invert the conclusion. This is the single most common way creators talk themselves into a format change they later reverse. If you have five shorts, you do not yet know how your shorts perform, and no tool that tells you otherwise is measuring anything.

Runsheet’s own format finding starts lower and makes up for it with a stricter test. It will not compare the formats until the channel has at least 4 of each, and once it has them it names a winner only when the two interquartile ranges, the middle half of each format’s videos, do not overlap. At the small end that usually means it says the two cannot be separated, and until it has the videos the box says what it is waiting for. The thresholds are published.

What shorts are actually for

The strongest case for shorts on a channel whose centre is long-form is not views, it is keeping the channel present in the weeks when a long cut is not ready. That is a consistency argument rather than a growth argument, and it is the reason to plan both formats on a single rhythm rather than as two separate pipelines.