Endings, and the four asks that cancel each other out

An ending does three jobs, and most do one

The last few seconds have to resolve the thing that was opened, land what it means, and say what happens next. Most endings do the third, badly, and skip the first two, which is why so many otherwise decent pieces finish with a logo, a website, a phone number, a QR code and three social handles arriving simultaneously over music that fades out.

Take them in order.

Resolve. Whatever question the piece opened has to close. If you started with a shopfront that no longer fits, the ending shows the sign that does. Shown, not stated: a claim at the end is the weakest position for a claim, because the viewer has just spent ninety seconds forming their own view.

Land the meaning. One line, or one image, that says what this was about beyond the events. This is the part that gets cut for time and is the reason some endings feel like they stopped rather than finished. It need not be profound. “Two panels. Same sign. Reads from the bus stop now.”

Ask. One thing. Which brings us to the real problem.

Why four asks produce fewer actions than one

A viewer at the end of a piece has a small amount of willingness to act. Present one option and they either take it or they do not. Present four and they must first choose, and choosing costs effort at exactly the moment their attention is lowest. A meaningful share resolve the effort by doing nothing, which is always available and always free.

There is a second mechanism, specific to screens: the four asks physically compete. The QR code, the URL, the handle and the phone number each take space and contrast from the others, so all four are smaller and less legible than one would have been, and at the moment where the viewer has least patience for squinting.

So: one ask, at the size the whole ending can afford. Everything else goes in the caption, the description, the bio: places that are still there after the piece has ended.

Make the ask small enough to be true

The ask has to match what the piece earned. A ninety-second film introducing an unfamiliar business has not earned “book a consultation”; it may well have earned “save this” or “call and ask what it would cost”. An ask that overreaches is not merely ignored: it retroactively reframes the piece as a sales pitch, which devalues everything before it.

A useful check: would a reasonable person who liked the piece do this within ten seconds, on the device they are holding, without standing up? If not, the ask is too large or the wrong channel.

Two mechanics worth knowing

The loop. On platforms that replay automatically, an ending that connects back to the opening makes the second viewing feel intentional rather than accidental, and watch-through figures rise. This is a real effect and a modest one; it is not a reason to contort a piece that has a proper ending.

Holding the last frame. Give the final card two to three seconds of stillness. Editors consistently cut this short because on the tenth viewing it feels interminable: a reliable illusion, since you already know what it says and the viewer is reading it for the first time. Three seconds is roughly the time to read and act on a short line. In DaVinci Resolve, Kdenlive, Shotcut or CapCut this is one drag of the clip’s edge.

Endings that do not ask for anything

Not every piece should end with a request. A film whose job is to make somebody think differently about a subject can end on the thought, and appending “follow for more” to it cheapens the whole thing: a sudden switch from talking to selling, in the last two seconds, which viewers register as a reveal of what the piece was for.

When there is no honest ask, end on the resolution and the meaning, and let the work stand. The brand can be present without a demand attached. This is a harder case to make to a client than a conversion-shaped ending, and the argument that works is about the measure agreed at the start: if success was defined as a shift in belief, then a request to buy is not the ending that was commissioned.