In a company where everyone sits on one floor, processes have a safety net nobody designed: someone gets up and asks. The missing paperwork turns up because the person from admin leans round a door. The doubt is settled in thirty seconds in a corridor.

That net doesn't exist when the operation is split across two islands, and a video call doesn't replace it. What replaces it is redesigning the process so it doesn't need the net.

I run internal digital support and automation for a company that works this way every day. Here's what I've watched break, and why.

1. The problem isn't distance, it's lag

What kills a distributed process isn't the kilometres: it's that information arrives late and out of order.

In one building, an incident is known the moment it happens, even if nobody writes it down. Across fifty kilometres of water, that same incident is known when somebody decides to mention it — which is usually once it has already become a problem.

In one location, the process finds out by itself. In two, the process only finds out what somebody bothered to write down.

The practical consequence: in a distributed process, recording has to be faster than telling. If logging the incident costs more effort than picking up the phone, nobody logs it, and at month end there's no way to know what happened.

2. What's a "reminder" in one place is "state" in two

A notification on its own fixes nothing. The question that matters isn't "was it sent?" but "who has it now?".

In a single-site company, when something stalls, somebody notices by eye. In a two-site one, an item can spend three days in no-man's-land with everyone convinced the other side has it. The classic exchange — "I thought you were handling it" — isn't a failure of people: it's a process with no explicit owner at each moment.

Every step of a distributed circuit has to answer two questions without phoning anyone: what state is it in and who does it depend on right now.

3. The hours aren't the same, even when the clock is

This one is subtle and it costs money. Two sites in the same time zone can run on completely different rhythms: a warehouse taking deliveries first thing and an office that works afternoons share an hour, but they don't share a moment.

If an order notification goes out at 17:00, in the office that's "before I leave" and in the warehouse it's "tomorrow". That lag, repeated, is what stretches deadlines while nobody is doing anything wrong.

The fix isn't technical, it's design: notifications fire when they're useful to whoever receives them, not when the previous step finishes. A first-thing digest for the warehouse and an immediate alert for the office beats the same email to both.

4. Documentation stops being bureaucracy and becomes infrastructure

On one site, documentation is a luxury: if someone's away, you ask the person next to you. Distributed, it's all that's left.

When a process lives in two places, the person who knows how something is done isn't available at the moment somebody needs to know it. And I don't mean hundred-page manuals: I mean every step saying, in one line, what it does and what's expected of it.

It's the same reason I write circuit documentation with the exact identifiers of each list and each field, not with descriptions. The person on the other island can't guess what I meant.

What this means if your company works like this

If you operate across two or more points — two islands, two units, an office and a site — three things pay off before any new tool:

Make recording easier than telling. A three-field form on a phone beats any complete system that requires opening a laptop.

Give every item a visible owner. It isn't enough for it to be assigned somewhere: it has to be visible without asking.

Make notifications respect the recipient's routine. A badly timed alert gets ignored, and an ignored alert trains people to ignore the next one.

None of the three is expensive. All three require having looked at how people actually work, which is the part nobody wants to pay for and the only part that makes the rest work.


If your operation is split and processes keep falling through the joints, tell me about it. I work from Tenerife, and this is what I do for companies here. If you'd like to see the diagnosis format first, the example is published.