Skip to main content

Why should we Limit WIP

August 31, 2023

A worker drowning in workHave you ever heard the saying, "The fastest way to get something done is to give it to someone busy." We all know it's silly, but we also all know it works. How does it work? The busy person already has too much work to manage, so they park the other work, get this thing done as quickly as they can, then go back to what they were doing before. So it works, but at the cost of all of the other work being delayed.  

There is another way. What if instead of saying "Yes" to another thing, we said "No" instead, or to be more honest, "Not yet"? This is limiting Work In Progress (WIP) - not starting something new until the last thing is finished. My Mum always said "Nothing ever has 4 screws in, in my house." What she meant was that my Dad would always have 25 things on the go, so where a thing needed screwing to the wall with 4 screws, he would put in 2 for now, thinking he would come back to it, but of course he never got back to it. I think there is a lot of that in many of us...

What happens when we limit WIP? We do one thing at a time until it is done. We finish one thing, then finish the next, then finish the next, and so on. Finish is a verb - a thing you can do. When we do this, we need to prioritise or order our work to do the most important / valuable / risky thing first, and then work on it until completion. Measured in elapsed time, the most important thing gets finished as quickly as it can be finished. That means we have something done really quickly, that delivers value, but also may give us learning about how important the next thing is, meaning we waste less time working on the less valuable things.

This elapsed time from start to done, is a named measure in Kanban, we call it Cycle Time (or sometimes Lead Time). So, when we limit WIP, we shorten cycle time. (That is technically only true until you hit the minimum possible cycle time for a thing, but that rarely happens in knowledge work, so most of us don't need to worry about that constraint most of the time.)

So what, we have shorter cycle time, what does that mean to us?  In 2018 we knew that it took around 15±5 years to develop a new vaccine, then 2020 happened. We developed half a dozen vaccines in a few months and so we learned that vaccines can be developed in months rather than decades.

Why does it normally take so long? Most of the time taken up by waiting. Waiting for funding rounds, waiting for access to the right people, getting time in the right laboratory on the right machine, waiting for analysis, and so on. With a Kanban lens, we can see that around 14.5 years is spent in queues, not actively doing any work! The workers are busy working on other things instead. They are wasting huge amounts of time task-switching. In 2020 we had a new governing constraint, the pandemic caused us to need vaccines as soon as possible, meaning that they no longer waited in queues for funding or people or labs or machines. COVID-19 vaccines were the single most important thing being worked on by the pharma companies. Vaccine work was finished as quickly as it could be done. We got value in 3 months rather than 15 years. Some of those vaccines were even new technology - the very first RNA vaccines to ever be developed and deployed, opening a whole new branch of tools for avoiding other illnesses. Using Kanban language, we Limited WIP, and drastically reduced Cycle Time. 

But here's the weird thing. By limiting WIP, we shortened cycle time which reduced the amount of work that we needed to do. Let me say this again. By shortening the cycle time on a piece of valuable knowledge work, we took less effort to deliver it. That's a bold statement, I know. But think about it.

How many budgeting meetings weren't held for the vaccines debating on whether to fund them or something else? How much management time wasn't taken up deciding what we work on when, who works on it, and what we should be doing in that lab right now. How much time didn't we spend having to read up on the work done last year when we last worked on this thing? How much time did we not waste working out what the last person did when they were working on it before they left the company or got promoted out of the job. Relearning is a massive overhead - having to relearn something we already learned a while back, just long ago enough that we forgot the detail and have to go through it again? A classic example is parents having to relearn how to do fractions again to help their kids with their maths homework, even though they were pretty good at fractions when we were their kids age. 

Lets bring this to a typical office scenario. Ever had to work on a big, important document for your company? It probably had to go through technical approval, legal approval, brand approval, and perhaps other approvals. How long did it take to get the document done? How many times did it need to be updated before it got approved? Cycle time in for the document was probably measured in months. Imagine you are the person in Legal, having to go through this big technical document with a fine toothed comb looking for legal mistakes. Not exactly a fun job, but a few days later you send back a list of changes - just like all of the other approvers. Now the document writer has to incorporate all of those changes that finally come in from all of the reviewers. It took days, weeks or even months just to get the reviews back, usually just long enough that new work has flushed all the detail from your mind, meaning you have a relearning burden before you make the changes then send it back out for review round 2. It's probably a month or so later now, and so the legal reviewer (if you are lucky enough that it's the same person) has to read and review the whole thing again, while meticulously checking their changes got applied. So do all of the other approvers. How much work just for a single document by each worker?  How much cycle time? 

What could we do differently? How about you get the writer and all of the reviewers into a room for a day and write the document from start to finish together. It might take a day of everyone's time and need some coordination, but there are loads of added benefits:

Sticky note with "Not Yet" written on it

1) The work is finished as quickly as it can be;

2) We have to work out loud, using the more intelligent parts of our brains (see my post "Have You Ever Been Stuck?");

3) We share learning, the whole team learns a bit about legal and brand etc, so the next document is even easier;

4) No-one has to relearn the document months after they last looked at it, no one has to rewrite it over and over. We do it once together and it is approved by everyone. In other words, we build a truly cross-functional team to do the work (sound familiar?);

5) It takes less work, the cycle time is reduced from months to a day or 2, the value is gained months earlier;

6) The work is far more engaging than reviewing a document someone else wrote, so it enriches the workers jobs. 

By limiting WIP, we shorten Cycle Time, which reduces the effort needed to do each thing, meaning you get more things done. By delivering the most important thing as quickly as it can be delivered (we never compromise quality in Scrum or Kanban) we get value and learning sooner, which means we spend less time working on the wrong things, meaning we spend more time delivering valuable things.  

And it all starts from saying "No" - or "Not yet" anyway. 

What did you think about this post?