The handover is where the margin goes.

Ask each department head whether their part of the process runs well and most will say yes, with evidence. Sales closes deals. Operations delivers them. Finance invoices them. Every step is measured, and every step performs. Yet the end-to-end takes three weeks longer than anyone thinks and costs more than the sum of its parts. The gap is not in the steps. It is between them.

Why joins go unowned

Organisational structure follows function, and so does measurement. A department is accountable for what happens inside its own boundary, which means the moment work crosses a boundary it belongs to nobody until it is picked up on the other side.

That interval is invisible in every departmental report, because each side records only its own elapsed time. Sales logs the day it closed. Operations logs the day it started. Nobody logs the four days in between, and nobody is asked to.

What it looks like from inside

Handover cost rarely appears as a line item. It appears as four familiar complaints, all of which are usually blamed on the wrong thing:

• Rework. The receiving team asks for information the sending team already had, in a different format. This is normally diagnosed as carelessness. It is almost always a specification problem: nobody agreed what “complete” means at the boundary.
• Chasing. Somebody spends a material part of their week finding out where things are. Chasing is not a personality trait; it is what a person does when a process has no status.
• Re-keying. Data leaves one system and is typed into another. Time paid for twice, errors introduced free.
• Expediting. Individual jobs get pulled forward because somebody escalated. Each expedite is reasonable and collectively they are the process.

Finding yours

You do not need a mapping exercise to locate the worst handover. You need one number, and most businesses can get it in an afternoon.

Take twenty recent orders, cases or jobs. For each, record the date it entered the business and the date it was invoiced. That is total elapsed time. Then record the time actually spent working on it. Ask the people who did the work, whose estimate is good enough.

The difference between the two is waiting. In most mid-market businesses the working time is a small fraction of the elapsed time, and the surprise is not that there is waiting but where it sits. It is rarely where the leadership team assumes.

Twenty jobs is enough. The pattern is usually so pronounced that a larger sample only makes people more certain of something they already saw.

Fixing a join

Three things fix most handovers, and none of them requires software.

1. Define “complete” at the boundary. Written down, agreed by both sides, and short: a list of what must be present for work to be accepted. Most rework disappears here, because most rework is caused by work being passed on before it was finished, by a definition nobody had written.
2. Give the boundary an owner. One named person accountable for the transition itself, not for either side of it. This is the least intuitive of the three and usually the most effective.
3. Make the queue visible. Whatever is waiting at a boundary should be visible to both teams without anybody asking. A shared list is enough. The point is not the tool; it is that waiting becomes something you can see rather than something you discover.

Automate second, not first

The instinct on finding a bad handover is to integrate the two systems. Sometimes that is right. Frequently it is expensive and it automates the current arrangement, faults included.

Fix the definition and the ownership first, then run the improved process manually for a quarter. What you learn in that quarter is what the integration should actually do, and often that the manual version is good enough and the money is better spent elsewhere. An integration built to replicate an unexamined process is the most expensive way to keep a problem.

Prove it in the accounts

Handover improvements are easy to claim and easy to lose. The benefit shows up as shorter elapsed time, less overtime, fewer credit notes and faster cash conversion, all of which are visible in the management accounts if somebody is looking.

Agree before you start which two or three lines should move, and check them at ninety days. If nothing moved, the change did not work, whatever the process map says. That discipline is the difference between a transformation programme and a transformation exercise.

This is the first thing we look for on performance and transformation engagements, and it usually explains more of the margin gap than cost lines do. It is general guidance rather than advice on your operation.