Section 1
Where the friction actually sits
Four specific places, in roughly descending order of cost. Idiom and reference. Sporting metaphors, military phrasing, jokes that depend on a shared television schedule: all of these are invisible to the person using them and opaque to everyone else. Directness norms. In some working cultures, a flat this will not work is professional candour, and in others it is a public humiliation, so the same feedback produces cooperation in one office and silence in another. Meeting behaviour, where a pause that reads as thoughtful consideration in one place reads as disengagement in another, and the loudest time zone wins by default. And status signalling, where the way seniority and disagreement are expressed varies enough that a junior person's carefully coded objection is heard as agreement. None of these are matters of translation. They are matters of what a listener is entitled to assume, and the fix is to reduce how much has to be assumed.
Section 2
Build a shared reference set from your own work
The workable approach is not to teach everyone one another's cultures. It is to build a small set of references the team made together, which then does the job that idiom used to do. A named incident everyone worked through. The customer situation that explains why a rule exists. A decision that was reversed and why. These accumulate quickly once you write them down, and they belong to nobody's home culture, so they carry equal weight for all participants. Say we are treating this like the migration last March, and the whole team has the same picture. Say we are not going to drop the ball on this one, and you have given several colleagues a puzzle instead of an instruction.
Section 3
Auditing what has to be assumed
The comparison that follows takes your recurring artefacts, the specification, the all-hands, the performance conversation, and asks what each one currently requires the reader to assume. Rows where the assumption is cultural rather than technical are your rewrite list.
Section 4
Write first, then meet
Distributed teams that work well tend to be written-first, and the reason is not documentation discipline. It is that written text can be reread, translated, and answered by someone who needs an hour to compose a careful reply in their second language. Live meetings advantage native speakers and extroverts, and they do it silently. So put the decision, the reasoning and the alternatives in writing before the meeting, and use the meeting for what only a meeting can do. Then check comprehension by asking someone to restate the decision in their own words rather than asking whether it is clear. Yes is the polite answer in most places on earth, and it carries almost no information. Founders exporting a story built for one market into several also underestimate how much of it was local, a trap visible in [From Garage to Global: Inspiring Startup Stories](/blog/from-garage-to-global-inspiring-startup-stories).
Section 5
What stories cannot fix
Shared narrative can paper over disagreement that should be surfaced. A team that has learned to tell the same story about a project may simply have learned which objections are unwelcome, and the quiet version of that is the dangerous one, because it looks like alignment. The second problem is asymmetry. On most global teams the stories come from headquarters, the examples come from the largest market, and the retelling always flows in one direction. That teaches everyone else that their context is the exception. Deliberately sourcing incidents from the smaller offices corrects it faster than any statement of inclusion. Practically, much of the daily friction is scheduling and handoff rather than meaning, and [Using Zapier and Other Workflow Automation Tools With AI](/blog/using-zapier-and-other-workflow-automation-tools-with-ai) covers the mechanical half. Investor communication carries its own version of the assumption problem, treated in [Telling Your Story to Investors: What Works, What Doesn't](/blog/telling-your-story-to-investors-what-works-what-doesn-t).