Handoffs are where companies leak

Ask a company where its time goes and you'll get answers about work: this team is under-resourced, that project is complicated, we're waiting on the vendor.

Measure it and you'll find something different. Most of the elapsed time in a typical business process isn't spent doing anything. It's spent waiting at a seam — sitting in a queue between one team finishing and another starting.

I've measured this in enough places to stop being surprised. In an eleven-step onboarding process we mapped last year, the actual work took about four hours. The elapsed time was nine days. Everything else was handoff.

Why seams are invisible

A step has an owner. The seam between two steps has none.

Each team measures its own segment, honestly, and reports good numbers. Sales closes fast. Onboarding executes in two hours. Both are true. Neither of them is measuring the four days a deal sits between "closed" and "someone in onboarding noticed."

This is a structural blind spot, not a people problem. Every incentive points inward at the segment; nothing points at the space between segments. So the space between segments is where all the slack accumulates, and it accumulates invisibly because no dashboard has a tile for it.

The four questions

At every seam in your process, ask:

1. Does work arrive, or does someone have to go find it? Pull-based seams — where the downstream team checks a queue, a shared inbox, a folder — are slower than push-based ones by a factor that scales with how busy the downstream team is. And they degrade silently: nothing breaks, things just take longer, and the delay is invisible unless you're timing the seam specifically.

2. Does everything needed arrive together? If the receiving team routinely has to go back upstream for one more detail, the upstream Out and the downstream In don't match. Every unit of work through that seam pays the round trip. Fixing the mismatch is usually an afternoon and pays forever.

3. Would anyone notice if the queue stopped? If work quietly stopped arriving, how long until someone asked? In most companies the honest answer is "when a customer complains." A seam with no alarm is a seam that can fail completely without generating a signal.

4. Who is accountable for the seam itself? Not the sending team, not the receiving team. Someone whose number is the total time across the boundary. Without this, both teams optimize their own side and the seam belongs to no one.

The cheapest fix in operations

Instrument the seams before you do anything else.

For each handoff, record two timestamps: when the upstream step completed, and when the downstream step started. Publish the median and the P90 of the gap.

That's it. No new tooling, no process redesign, no reorg. Two timestamps and a weekly number.

I have never once seen this done for the first time without producing a genuine surprise — a seam eating days that everyone assumed took hours. And the fix, once the number is visible, is almost always small: an automated notification, a shared definition of "ready," one person made accountable for the boundary.

The reason it stays broken is not that it's hard. It's that nobody could see it.

The deeper version

A seam that leaks is a fitting that doesn't fit. The law of flow says segments connect because their inlets and outlets are defined; a leaking seam means one of those definitions is missing or wrong.

Which gives you the permanent fix, as opposed to the patch. Write down what the upstream segment produces and what the downstream segment requires. Compare them. The gap is your leak, in writing, before you've spent a minute on process improvement.

Most companies try to fix handoffs by asking people to communicate better. Communication is what you need when the interface is undefined. Define the interface and the communication becomes unnecessary — which is the point, because "remember to tell them" is not a system, it's a hope.