The search behaves differently from an ordinary job search in one respect that governs everything else: you are no longer competing with the people who can drive to the office.

The arithmetic you are up against

LinkedIn's workforce data through 2022 and 2023 repeatedly found the same imbalance — remote roles were a small minority of postings, in the region of a tenth to a seventh, and attracted around half of all applications. Whatever the exact figure in a given month, the ratio is the point: applying to a remote role is entering a queue several times longer than the one for the identical job down the road.

This makes the standard advice actively harmful. Volume is the losing strategy, because volume is what everyone else in the queue is already doing, and the screening step that stands between you and a human is tuned to reject rather than to find. Proxy work is the risk here: sending eighty applications feels like a full week and is the most legible possible way to make no progress.

What actually clears the queue

Evidence a stranger can inspect in two minutes without a conversation. Remote hiring is a trust problem before it is a skills problem — the employer is deciding whether someone they will not see in a corridor will produce work unsupervised — and anything that settles that question early does more than any phrasing on a CV.

In practice that means something with a URL. Writing that demonstrates how you think about the actual domain. A repository, a dashboard, a teardown of the company's own funnel or product, a short recorded walkthrough of how you would approach the problem in the posting. The form matters far less than that it is finished, public, and about the work rather than about you.

The asymmetry is worth stating plainly: fifty applications is roughly a week and produces nothing durable. One public artefact is roughly the same week and keeps working after the search ends — it is still there for the next role, and it is the difference between applying and being found. That is leverage, in the exact sense.

If you have no remote experience

Remote-ify where you are before you move. The credential the employer wants is evidence of asynchronous output, and that can be built inside a conventional job: run a project in writing, move a recurring meeting to a document, produce the weekly update nobody asked for. Six months of that is a demonstrable record, and it is available without changing employer.

If you are also changing field, change one thing at a time. Industry or function, not both — changing both at once means arriving with no transferable proof and accepting the salary that goes with it.

And be concrete about which roles are actually winnable remote without a track record. The ones whose output is inherently legible — writing, code, design, support, data, anything with a visible artefact — are structurally easier than roles that are mostly presence and judgement, because the evidence problem solves itself.

What the research says about the thing you are choosing

Nicholas Bloom and colleagues ran the field's defining experiment at a Chinese travel firm, published in the Quarterly Journal of Economics in 2015. Working from home raised performance by 13 per cent and cut attrition sharply. It also, in the same study, roughly halved the promotion rate of home workers conditional on performance — they produced more and were promoted less.

That finding is nearly always omitted from remote-work advocacy and it should determine how you choose. Bloom's later randomised trial of hybrid work at Trip.com, published in Nature in 2024, found no effect on performance or promotion and a third less attrition — so the penalty attaches to being fully remote and invisible, not to not being in a chair.

The operational conclusion is not to avoid remote work. It is that visibility has to be manufactured deliberately because the building is no longer doing it for you — written output that circulates, a standing relationship with someone senior, a record of what you shipped that exists independently of your manager's memory.

What this rules out

It rules out application volume as a strategy, which is the one thing most people in the queue are doing. It rules out a CV as the primary instrument, since it is the format in which you are indistinguishable from several hundred others. And it rules out treating remote as neutral on career progression — on the only randomised evidence available, fully remote carried a promotion cost and hybrid did not.

It does not rule out applying. It rules out applying first. Build the thing that answers the trust question, then apply, and expect the artefact rather than the application to be what opens the conversation.