Scaffold or Straightjacket?

Douglas Adams' classic sci-fi comedy novel “The Hitchhiker’s Guide to the Galaxy” opens with a bulldozer approach Arthur Dent’s house. Since Arthur is still inside the house, he is naturally concerned.

When Arthur confronts the foreman of the demolition crew, he is informed that his house is to be destroyed to make way for a highway bypass. When discussing the public notice that the local planning office had posted, they have this conversation:

“But the plans were on display…”

“On display? I eventually had to go down to the cellar to find them.”

“That’s the display department.”

“With a flashlight.”

“Ah, well, the lights had probably gone.”

“So had the stairs.”

“But look, you found the notice, didn’t you?”

“Yes,” said Arthur, “yes I did. It was on display in the bottom of a locked filing cabinet stuck in a disused lavatory with a sign on the door saying ‘Beware of the Leopard’.”

For the record, there was no leopard.

This is a funny moment for both the absurdity and the familiarity. Anyone who has interacted with a bureaucracy can recognize their experience in Arthur’s. Unfortunately, we tend to attach the name process to both that kind of experience and a very different one.

Process as (Accidental) Constraint

Some processes are deliberately designed to limit or constrain the “consumer” of the process. These are the exception though.

Most of the time, a group or department manager will create a process for how that particular group does its work. Where the trouble arises is that an Arthur Dent doesn’t just interact with that one group. Instead, he has to deal with several groups that each have their own processes. Each group knows their own process, but probably has no view into the processes of the other groups. They can point Arthur from their own department to another (to go get a signed form of some kind of another.)

Each group acted reasonably, but the experience from Arthur’s point of view is absurd.

As a personal anecdote, my wife is a US citizen who was born to two US parents in a US Army field hospital in Bangkok, Thailand. Consequently she had dual citizenship until age 18. At that point, the US State Department contacted her to declare which citizenship she intended to keep. She had to send a form back to that department, including a consular certificate of natural birth. Where would that certificate come from? The US State Department. In other words, she had to send a request to one office in the State Department to get a document to send back to another office in the State Department. For years we have joked that those offices are probably across the hall from each other.

We have here what the outside observer experiences as one “process.” But no-one can tell them the entire process because there is no global designer. It is a piecemeal of constantly-changing internal departmental processes. Thus the whole picture is shrouded (the lights had gone) and there are leopards.

Taiichi Ohno’s Kind of Process

Taiichi Ohno created what we now call the Toyota Production System. It has inspired decades of study in quality and rapid improvement. From TPS we have gained vocabulary like “kanban”, “andon cord”, and “kaizen.” You could get a Master’s level course in process design by studying just Toyota and Waffle House. (The American restaurant chain.)

One of the most eye-opening things about TPS is how they approached processes. Every work station had the work process printed and posted right where the work is done. Every process was updated almost every day. In fact, Ohno would walk the factory floor looking for process documents that looked aged: stained paper, yellowing, tears, etc. He would then ask the worker why they had not learned anything in such a long time. He would then ask that worker’s manager why they had not learned anything. In TPS, kaizen is not a big enterprise initiative… it happens a thousand times a day in small groups of workers and their managers solving problems together.

On a previous project, we had an XP lab full of pairing stations. We wanted all the pairing stations to be identical: same OS, same configuration, same IDE. That way any pair could take any station on any day and be productive. To make this work, we had a wiki page with setup instructions. Every time we needed to do a fresh setup, we would walk through the instructions. I wrote the initial instructions, but even so I walked through the instructions each time to make sure I incorporated improvements that other people had made. If we found errors, we updated the process. If we found ways to improve efficiency, we updated the process. In fact, the most common kind of problem came because we didn’t do the process often enough so today we would probably reimage one station every day to make sure we kept pressure on improving that process. The consistency of the stations paid dividends every day because we never had contention for “the QA machine” or “the machine with the memory card reader.”

Scaffolding versus Straightjackets

It’s an unfortunate collision in the English language that both of these experiences have the word “process.”

Taiichi Ohno’s kind of process is a scaffolding. It supports the work and lifts up the worker to perform at a higher level of quality. It captures the best of what we’ve learned about how to do the work so that everyone can benefit.

Because the process is written and posted right with the work, it means that changing the process document actually changes the process. (As opposed to changing the document then holding training sessions, sending work in progress back to square one, and having stragglers following the old process for months.) In other words, they write the process down exactly so it can be changed!

Ohno’s processes allow the worker to improve his or her own work. The Arthur Dent style of process is defined by the worker for other people to follow. The difference is immense.