Bring the position with you, because a draft without one finds the middle

Balance is the default and it is not neutral

Ask for a piece on whether an organisation should adopt something, and the result will lay out considerations on both sides, acknowledge that context matters, and close by noting that the right answer depends on circumstances. Nothing in it is false. Nothing in it is a recommendation either.

This is the safe output for a question whose answer wasn’t supplied, and safety here has a specific shape: every claim qualified, every objection granted, no sentence that could be quoted back as wrong. That shape is appropriate for an overview and useless for a decision.

The mistake is thinking the balance can be edited out afterwards. A piece built to be balanced has its structure organised around the balance — the sections are the two sides — and converting it into an argument means rebuilding it, not deleting the hedges.

The conclusion is the specification

An argumentative piece is written backwards from its conclusion. Which objections need answering, what evidence has to appear, what can be left out, what order the reader will accept it in: all of that is determined by what you are trying to establish.

So supply the conclusion, and supply it as a sentence you would defend. We should not do this, because the cost falls in the first year and the benefit is speculative. That single line does more for the draft than three paragraphs of context, because it tells the writing what it is for.

If you cannot state the conclusion, that is worth noticing rather than working around. It usually means the thinking hasn’t been done, and a fluent draft produced at that point will feel like progress while quietly deferring the only difficult part of the job.

Objections are where the value is

The genuinely useful request in argumentative writing isn’t the case for your position — you already have that — but the strongest version of the case against it. That is a task with a clear target, and it uses the breadth of the material rather than asking for a judgement.

Ask for the three objections a well-informed opponent would raise, ranked by how much trouble they cause. Then answer them yourself, because your answers are the part a reader has not encountered before, and because writing an answer is how you discover whether you have one.

Expect one of the objections to be better than you thought. That is the outcome to hope for, and it is also the reason to do this before the piece exists rather than after, when the sunk cost of a finished draft makes the objection considerably less welcome.

Hedges: which to keep and which to cut

Not all qualification is padding. When the evidence is genuinely mixed, saying so is accuracy, and stripping every hedge produces the overconfident register that makes readers distrust the whole piece. There is a real difference between a qualification about evidence and a qualification about nerve.

The test is whether removing it would make a sentence false. It may be that costs rise removes nothing but risk; costs rose in the two cases we examined is doing genuine work and should stay. Applying that test one sentence at a time is quick and it is the whole of the technique.

What consistently should go is the closing paragraph that restates both sides and declines to choose. It appears with remarkable regularity, it undoes the preceding argument, and it is almost always the last thing added.

Where balance is the right output

Plenty of writing should not take a position. Explaining a contested area to someone who has to form their own view, summarising a debate, documenting options for a group that will decide — in all of these the balanced default is exactly correct and forcing a conclusion would be a distortion.

The failure is only in the mismatch. What causes trouble is a piece that was commissioned to make a case, arrives balanced, gets published, and leaves the reader with no idea what the author thought, which is a common enough outcome that readers have learned to skip to the end looking for it.

So decide which kind of document you are producing before anything is drafted, and say which it is in the instruction. That one decision determines more about the result than any amount of subsequent editing.