Short writing is where a pile of options actually earns its keep

Volume is a bad idea in long form and a good one here

For an article, asking for ten versions is close to useless: you cannot read them properly, they differ in ways that do not matter, and choosing between them is more work than writing one. The evaluation cost swamps the production cost, which is why more options make the job harder.

Short writing reverses that arithmetic. Forty headlines can be scanned in two minutes. Sixty subject lines can be sorted into a keep pile and a discard pile faster than three could be written from nothing. Where a candidate takes seconds to judge, having many of them is a straightforward advantage.

There is a second reason it works here. Short forms have a small design space, so a large batch genuinely covers it — the direct version, the question, the number, the contrarian angle, the plain descriptive one. You end up looking at the range rather than at one arbitrary point in it.

The batch is raw material for a choice, not a shortlist

The mistake is treating the output as candidates to pick from. Most of a large batch is filler by construction, and the value is rarely a line you can use unedited. What you are actually harvesting is angles: two or three framings you had not considered, which you then write properly yourself.

A practical method is to mark the ones that are interesting rather than the ones that are good, since those are different sets. An awkwardly phrased headline built on a sharp idea is more useful than a polished one built on the obvious idea, and it survives being rewritten.

Then close the list and write. The final version is usually a hybrid of the structure of one candidate and the specific detail of another, produced by you in about a minute, because you now know what the space of options looked like.

Force the range instead of hoping for it

An unconstrained request for many titles produces one idea in many outfits, because there is no pressure towards difference. Naming the categories fixes it: ask for five that state the conclusion, five that name the audience, five that use a concrete number, five that are deliberately plain, five that lead with the objection.

You can also constrain mechanically. Ask that none exceed nine words, that none use a colon, that no two share their first word. Those restrictions are trivially checkable and they push the batch outwards, because the easy formulations get eliminated by the rules rather than by your patience.

Supplying the material matters more than the count. A request that includes the actual argument of the piece, the reader it is for, and the one thing you want remembered will produce a better batch of twenty than a bare topic will produce of two hundred.

What the batch cannot tell you

It cannot tell you which line will perform. Performance depends on the audience, the surrounding page, the channel and what was sent last week, none of which is available, and confident predictions about which option will do best are not grounded in anything you should rely on.

It is also weak on collision. A name that is available as a trademark, a subject line that resembles a phrase your competitor owns, a headline someone published last month — checking those requires a search, and the check is the part that protects you rather than the generation.

And it will drift towards what already exists, which is exactly the wrong tendency in a naming exercise. If everything in the batch sounds like a category you have heard of, that is a fair signal it will not stand out, and the plain descriptive option may be a better bet than the cleverest one available.

Where this genuinely fits in the work

The natural point is after the piece is written rather than before. A headline written first tends to become a promise the article then has to keep, whereas one written afterwards can describe what the piece actually turned out to be, which is usually different from what you set out to write.

It also fits the everyday tasks nobody wants to spend time on: a subject line for a routine notice, a label for a button, a summary for a listing. These are small decisions with disproportionate effect, and being able to see twenty versions of one raises the floor without costing you an afternoon.

The limit is worth stating. This is a technique for finding angles and moving past a blank line, not a substitute for knowing what you want to say. If none of the forty options is any good, the usual cause is that the piece underneath does not have a clear point, and no batch of titles will supply one.