Choosing a Remote Model: Three Real Options and One Fake One

Most organisations never decided on a remote model. They reacted to circumstances, then stopped adjusting, and now have an arrangement nobody chose and nobody can explain.

There are three real options, and each has a coherent logic. There is also a fourth thing people say, which is not a model.

Fully remote

No default office. Everyone works from wherever they are. Meeting in person is an event you organise, not a routine.

What it requires. Documentation as the primary medium, because there is no corridor. Deliberate onboarding, because nobody absorbs anything by proximity. Hiring managers who can assess people they never meet. And real budget for periodic gatherings — fully remote organisations that never bring people together tend to fragment within about two years.

What it buys. Access to people anywhere, no property cost, and a forcing function on clarity: decisions have to be written down because there is no other way to transmit them.

Where it fails. Work requiring dense, fast, ambiguous collaboration. Early-stage teams inventing something. Anything where the answer emerges from a group arguing at a whiteboard for a week.

Hybrid, defined

Specific days in the office, specified in advance, the same for a team.

The important word is defined. Hybrid works when the whole team is in on the same days, because the point of the office day is the people. It fails when everyone chooses their own days and arrives to find an empty floor and a full calendar of video calls — which is the commonest implementation and the worst possible outcome. You pay for the commute and the property and get none of the benefit.

What it requires. Coordination at team level rather than individual level, and honesty about which activities actually justify the day.

Office-first

Default in person, remote as an exception with a reason.

Unfashionable and legitimate for work that genuinely needs it: heavily regulated environments, physical processes, roles requiring supervised access to systems.

What it requires. Saying so plainly in hiring. An office-first organisation that advertises "flexibility" and then requires attendance loses people at month three and deserves to.

The fake option

"Flexible."

It means nothing operationally. Flexible about what — location, hours, notice period, which days? Decided by whom, the individual or the manager? Reviewable when?

In practice "flexible" resolves into whatever each manager decides, which produces different rules on different teams, resentment between them, and no ability to plan anything. It also loads the entire decision onto individuals who have the least information and the most to lose from getting it wrong.

If you cannot write your model in three sentences that a new hire would understand, you do not have a model.

How to choose

Start from the work, not the preference. Which activities in your organisation genuinely degrade at distance? Be specific. "Collaboration" is not an answer; "resolving ambiguous requirements with three functions in the room" is.

Count the real cost of each. Fully remote costs coordination overhead and gathering budget. Hybrid costs property plus commute plus scheduling. Office-first costs your hiring pool.

Decide at the level that can execute it. Team-level hybrid works; company-level hybrid where teams differ does not.

Write it down and date it. A model without a written statement is a rumour, and rumours drift.

Then leave it alone for a year. The single most damaging thing organisations did between 2021 and 2025 was change the model repeatedly. Every change resets trust, invalidates people's housing and childcare arrangements, and teaches everyone that the current policy is temporary.

The honest summary

All three real models work. The failure is almost never the model chosen — it is choosing none, or changing so often that no one can plan.