UTSUROI
日本語Early access

How to build a team where requests don't get left hanging

Dropped requests are a process problem, not an attention problem. Remove the copy-over step and the odds change by an order of magnitude.

“Hey, what happened with that thing I asked for?”

Teams where this question flies around several times a week are usually conscientious. Nobody is slacking. Things still get dropped.

They get dropped because the process is built to drop them.

Every step has a small chance of failing

Here is a typical request flow, with probabilities attached. The numbers are assumptions to show the idea, not measured values.

Step Chance of dropping
A asks in Slack B doesn’t read it: 5%
B copies it into the task tool Forgets to add it: 5%. Forgets to set a due date: 15%
B looks at their own tasks Misses it: 10%
B does the work —

Each number is small. They multiply anyway. And the “forgot the due date, 15%” case does not look like a drop at the time. It shows up two weeks later.

What most teams do here is remind people: “please add it properly,” “set reminders.” That nudges each step’s probability down a little. The number of steps stays the same.

Remove one step

What works is not lowering the probability but removing the step.

The most fragile one is the copy-over. The request starts in one place (chat) and the task is managed in another (the tool). A person carries it by hand between the two. Remove that.

A reaction turns the request into a task. That alone changes the table:

Step Chance of dropping
A adds the reaction (the request becomes a task) Unread no longer matters
Copy-over The step no longer exists
Once set up, the due date is notified in a public place Someone else catches a miss
B does the work —

The drop in “Missed: 10%” is not an improvement in attention. With due-date notifications turned on, the due date is out in the open, so if B misses it, someone else notices. The structure no longer depends on one person’s memory. That is all it is.

Don’t blame: a design principle

There is a flip side to this view. Don’t treat a dropped item as one person’s fault.

As long as a drop is “an individual’s responsibility,” it does not get reported when it happens. If it is not reported, it cannot be detected, and the process never gets fixed. It is a vicious circle.

Turn drops into “something the process should detect and handle,” and the opposite happens. A drop becomes visible as a matter of course, you can see which step it fell through, and you can discuss removing that step.

A team can move to asynchronous work with confidence only when that confidence comes first. If people can’t trust that the system will catch what they miss, they go back to synchronous out of anxiety. More meetings, more check-ins, and the synchronous share goes up.

Where to start

You don’t have to change everything at once. These three alone make a noticeable difference.

  1. Don’t finish a request in speech or a DM. Write it somewhere that lasts.
  2. Give every request an owner and a due date. “Someone this week” is not a request.
  3. Keep “waiting for confirmation” as a state. If you can’t tell “not started” from “sent back, waiting for a reply,” the waiting stays invisible.

UTSUROI is a tool for running these three on Slack and the board.


For the concrete steps, see the recipe.

← Back to all posts