Six months into distributed working, the teams that are visibly struggling are not the ones with the worst tools or the least discipline about meetings. They are the ones that cannot remember anything. The symptom is easy to spot once you know to look for it. The same question is answered three times in a fortnight. A decision taken in May is reopened in August because nobody can find what was agreed or why. A new joiner cannot do anything without interrupting the two people who hold the context. A team lead spends most of the day being a search engine for their own project. That is not a communication problem and it will not be fixed by better meetings. It is a storage problem, and it is the strongest argument for writing things down — much stronger than the argument everybody actually makes, which is that asynchronous work means fewer meetings.
Writing is a storage format, not a communication style
The office concealed this for years. A verbal organisation has a functioning memory as long as everybody sits together: the person who knows is four metres away, the context is in the room, the overheard conversation does half the work. Remove the room and the memory goes with it, because the memory was never in the organisation. It was in a handful of people's heads, retrieved on demand by proximity. Useful framing: information in a distributed team sits at one of three levels of durability. Ephemeral. Chat messages, calls, quick video huddles. Findable for about two days, then effectively gone. Semi-durable. Threads, tickets, comments on a task. Retrievable if you know where to look and what it was called. Durable. A document with a title, an owner, a date and a location people can predict. The failure mode of nearly every struggling team this year is that they have only the first level and are unconsciously treating it as the third. Everything important happened in a channel. The channel scrolled. It is all technically still there and none of it is retrievable.
Four artefacts, and no more than four
Documentation programmes fail by being ambitious. Keep the required set small enough that people can hold it in their heads. The decision record. One page per significant decision: the question, the options considered, what was decided, who decided it, the date, and one line on what would cause us to revisit it. That last line is the one everybody omits and the one that stops a decision being relitigated every time somebody new disagrees with it. The open question log. Everything not yet decided, each with a named owner and a date by which it will be. Distributed teams do not usually fail on decisions taken; they fail on decisions quietly left in suspension while everyone assumes somebody else is handling it. The explainer. One page per system or process that answers "how does this actually work here", owned by a named person with a review date. Not a manual. A page. The team's working agreement. Where each type of thing lives, what response time is expected on each channel, what genuinely requires a meeting, and what hours people are contactable. Half the friction in remote teams is unstated expectations about response time, and it costs an hour to write down. Everything else is optional. If your programme requires more than these four, it will be abandoned by November.
| Record | Question it answers | Keep visible |
|---|---|---|
| Decision record | What was chosen and why? | Decision owner, date and revisit trigger |
| Open question log | What still needs a decision? | Named owner and due date |
| Explainer | How does this work here? | Page owner and review date |
| Working agreement | Where do things live? | Channels, response expectations and contact hours |
Qualitative summary of this article's source text, not a measured outcome or performance estimate.
Write before the meeting, not after
The habit that changes behaviour fastest is inverting the order. Most teams meet to discuss and then write a summary nobody reads. The alternative is to circulate a short written proposal in advance, open the meeting with five minutes of silence while everyone reads it, and then use the meeting only for the disagreement. This works for three reasons. It forces the proposer to think, because vagueness survives speech and dies on the page. It lets people who need time to process contribute properly rather than losing to whoever speaks fastest. And it produces the record as a by-product rather than as a chore assigned to the most junior person present. It also shortens meetings dramatically, which is the benefit everyone notices and the least important one.
Why the wiki is a graveyard
Every organisation already has a documentation site somebody set up with enthusiasm, and almost all of them are unusable. The causes are consistent and fixable. There are three competing places to write, so nobody knows where to look or where to put things. Pages have no owner, which means nobody is responsible when they go stale. Nothing has a review date, so a reader cannot tell whether a page is current or four years old — and an out-of-date page is worse than no page, because it is confidently wrong. And nothing is ever deleted, so the search returns eleven results of which two are authoritative and nine are archaeology. Fix all four or do not bother: one canonical location per type of content, an owner on every page, a review date on every page, and a genuine willingness to delete. An unowned page is a rumour with formatting.
Practical Guidance for Async Practices Workshop
- Adopt the decision record and nothing else for the first month. One page, six fields, including what would make us revisit it. Habits beat templates.
- Keep an open question log with owners and dates. Suspended decisions are the quiet killer of distributed delivery.
- Circulate written proposals before meetings and read them in silence at the start. It produces better decisions and generates the record automatically.
- Publish a one-page working agreement per team. Where things live, expected response times, contactable hours, what needs a meeting.
- Put an owner and a review date on every page, and delete aggressively. A confidently wrong page costs more than a missing one.
- Pick one canonical location per content type and enforce it. Three tools that overlap are worse than one tool that is imperfect.
- Reserve synchronous time for disagreement, ambiguity, relationship and bad news. Those four are where talking genuinely outperforms writing.
- Make the written record the input to onboarding and measure it there. How far a new joiner gets before needing to interrupt someone is the only honest test of your documentation.
The Regional Angle
Four considerations matter particularly here, and the first is about fairness. A written-first culture quietly advantages whoever writes English most fluently. In a typical regional team, English is the working language and the first language of almost nobody: colleagues are thinking in Arabic, Hindi, Urdu, Malayalam, Tagalog or Tamil and translating as they type. Move everything to writing without acknowledging that, and you systematically over-weight the contributions of a few confident writers while capable people who are articulate in a meeting go quiet. The mitigations are practical rather than sentimental. Say explicitly that imperfect English is fine and that nobody will be judged on grammar. Use templates and bullet fields rather than prose, because a form is far easier to complete in a second language than a blank page. Accept voice notes as an input and have someone convert them into the record. And keep writing quality out of performance conversations, where it is a proxy for fluency and not for competence. Second, be deliberate about which language the record lives in. Regional organisations run two documentary worlds: Arabic for statutory filings, government correspondence, contracts and anything that will be read by an authority, English for internal operations. Decisions about processes governed by Arabic-language regulation get recorded in English by whoever attended the meeting, the two versions drift, and eighteen months later nobody can reconcile what was agreed with what was filed. Decide per document type which language is authoritative, and where both are needed, say which one governs. Third, work with the approval culture rather than against it. In owner-led and family-run groups, significant decisions are made verbally, often quickly, and a subordinate who circulates a written decision record before the decision has been blessed can appear to be presuming. The convention that works is the confirmation note: immediately after the conversation, a short written message back to the decision maker — as agreed today, we will do X, starting Y, please correct me if I have this wrong. It is polite, it is normal business practice, and it converts a verbal instruction into a durable record without anyone having to change how they prefer to decide. Fourth, and most urgently this year: knowledge that leaves here usually leaves the country. In a normal year, the person who set up that reconciliation process is still in Dubai and can be telephoned. This year they were made redundant in May, their visa was cancelled, and they are in Kerala or Cairo or Manchester with a new number and no obligation to you whatsoever. Every regional group has lost institutional knowledge permanently in the last six months, and the departures are not finished. That makes written records an asset-protection measure rather than a productivity nicety, and it makes a documented handover a condition of final settlement rather than a courtesy.
The objection worth taking seriously
The strongest objection is that this is advice from and for a particular kind of company. Distributed documentation culture was developed by software firms staffed entirely by literate knowledge workers with individual laptops and quiet rooms. A regional group running warehouses, clinics, construction sites, retail floors and a contact centre has a workforce that does not sit at a keyboard, customers who phone rather than write, and regulators who want a stamped paper document. Telling a shift supervisor to file a decision record is not a productivity intervention; it is a way of generating non-compliance. The second objection is subtler and comes from inside the practice. Writing cultures can degenerate into performance: fifteen-page documents circulated for status, an expectation that everything is justified in prose, and a genuine slowdown in decision-making as the written form becomes a hurdle rather than a record. The organisations that overcorrected this year have people spending mornings documenting work they have not started. Both objections argue for the same discipline: document decisions and open questions, and almost nothing else. Do not document processes that change every week. Do not require writing from people whose work is physical. Keep the artefacts to four, keep them short, and remember what you are actually buying — not fewer meetings, but an organisation that can still answer in October why it chose something in May, when the person who chose it has left the country.
Common Questions
Does this mean we should have fewer meetings?
Probably, but that is a side effect. The point is a retrievable record; the meeting reduction follows from doing the writing beforehand.
Which tool should we use?
The one people already open. Tool choice explains very little of the variance between teams that manage this and teams that do not; ownership and review dates explain most of it.
How do we get people to actually write?
Make it required for exactly one thing — decisions — and make leaders do it visibly. Programmes that mandate documentation everywhere achieve it nowhere.
What should we expect over the next twelve months?
Expect written-first practices to persist in the teams that adopted them properly and to evaporate everywhere else the moment offices reopen, because habits without artefacts do not survive a change of setting. Expect documentation and knowledge tools to be marketed hard as the answer, and expect most of those purchases to produce another graveyard. And expect the real test to arrive with the next wave of departures, when organisations discover exactly how much of their operating knowledge was stored in people who are no longer in the country.
Async Practices Workshop — we install the four artefacts that give a distributed team a memory, and the review discipline that keeps them worth reading.