What an offsite can and cannot do

Why it is capable

Two or three days is enough for a real task. Not a big one, but a finite thing that two people who have never worked together could produce and that somebody would use afterwards.

Co-location removes the entire cost of initiating contact. The thing that prevents a cold pair forming remotely is that somebody has to decide to start it. In a room, starting it costs nothing: you turn round.

And it is bounded. Everybody knows it ends, which makes people willing to spend attention they would not spend on an open-ended commitment.

How that capacity usually gets spent

Three ways, in descending order of frequency.

• On presentations. A day and a half of slides from leadership, roadmap walkthroughs and quarterly reviews. Information that could have been a document is delivered to eight people who flew somewhere to receive it. This is the most common use and the worst, because it reproduces the all-hands at enormous cost.
• On facilitated exercises. Values workshops, personality instruments, team charters, retrospectives on the year. These are better than slides because people talk, and they are still one-to-many: the outputs are group artefacts, and the pairs that form are between whoever happened to be on the same table.
• On socialising. Dinner, an activity, drinks. Valuable and already covered: it lowers the cost of a future conversation and does not constitute one.

What almost never happens is the thing the format is uniquely good for: two people who have never worked together making something small together, in the room, for four hours, with somebody waiting for it.

An offsite arranged for pairs

The structure is not complicated and it is unpopular with everybody who has to organise one.

Before: run the pair count. Identify three or four pairs that do not exist and should — the ones the seam audit named. Find a small piece of genuinely needed work for each.

During: give each pair a half day and a room. Not a workshop, not a breakout with a template, not a facilitator. A piece of work, a deadline of four o'clock, and somebody who will look at it afterwards.

Keep the presentations to one session and put them on the first morning so that everything afterwards is the real content.

Keep the dinner. The dinner is the part that makes the next day's work easier and it should not be cut to make room for more sessions.

After: take the count again in six weeks. The pairs made in a room do not survive automatically; they survive if they are given a second piece of work before the first has faded, which is a thing somebody has to remember to do.

What an offsite cannot do

It cannot fix a structural problem that the work recreates. If the reason two people never work together is that their work does not touch, three days in a room will produce a pleasant pair that evaporates in a month, because nothing at home will cause them to speak again.

This is the single commonest disappointment with offsites and it gets attributed to the offsite being badly run. The event did its job. The arrangement of work undid it, as it was always going to.

So the honest sequence is: fix what you can about how work is arranged first, and use the offsite to start the pairs that the new arrangement will then sustain. An offsite is excellent at starting things and has no ability whatsoever to maintain them.

The argument you will have with the organiser

Proposing that half an offsite be spent on actual work is unpopular, and the objection is always the same: people have travelled a long way and should not spend it doing what they could have done at home.

It is a good objection and it has an answer. They could not have done it at home — not with that person, not in four hours, not without three weeks of scheduling. The thing being bought with the travel is not the hours but the removal of every obstacle to two people starting something, and that is spent by using it, not by preserving it.

The version that genuinely could have been done at home is the slide deck, and it is usually the part nobody proposes cutting.

And the cheap version

If there is no budget for an offsite — and for many teams there is not — almost everything above survives the translation. The pairs are made the same way, with a task and a deadline, and the only thing lost is the room.

Losing the room costs more than it sounds, because initiating contact goes back to being somebody's decision. The remedy is that the task has to be assigned rather than offered, which is the whole content of the last part of this site.