Fanning the flames on Work in Process limits

I’ve been thinking about different analogies to help teams understand the importance of limiting work in process (WiP) – the value is in the stuff we finish, not the stuff we start.

Having too much WiP ultimately reduces throughput, but this is counter-intuitive so I need a simplier way of communicating the idea.

I was struggling to light the bbq over the weekend & the thought came to me – can lighting & maintaining a fire be like managing WiP?

Lets find out….

I’m going to use a log burning stove instead of a bbq as there’s more constraints & control. I’ll be using the language from this post on how to build a fire.

The problem i’m trying to solve with the fire in my log burner is that i’m cold & the heat from the fire will warm me up (heat is the value)

Log burnerProduct Development
HeatValue
Air / Oxygen (fuel)Work in Process (WiP)
Wood (fuel)Work in Process (WiP)
Damper (top vent)Byproduct of delivered value leaving the system
Front ventWork entering the system
Draft (flow of air / oxygen)Flow of work

Whats in a fire?

In a nutshell, we need to balance the amount wood & oxygen to generate the most amount of heat (heat is what we value).

When we’re starting a fire, we open both the front vent & the damper in order to create a draft which pulls the air (oxygen) from the room & up the chimney.

We add wood (fuel) to the burner & set light to it.

When the vents are open, the fire is roaring. This looks great, but delivers little value – the burner isn’t giving out much heat.

We close the vents to reduce the draft, reducing the oxygen, the fire settles down & starts generating the heat that we value.

In order to maintain the fire, we need to add more wood. We may also need to tweak the vents to increase / decrease the draft.

How does it go wrong?

Building a fire sounds simple doesn’t it, but of course there’s nuances to keeping it lit & generating heat.

However, here are some examples of how we can over-tweak (aside from not having enough fuel or oxygen!)

We leave the vents open = too much WiP

If we leave the vents open, too much oxygen will be in the burner. Not only we get a weak fire, we’ll also ‘burn’ through wood at an accelerated rate.

So we’ll still be cold & have no wood to show for it.

Leaving the vents open is like taking too much work into the system – it’ll look fantastic in the stats & metrics, but the value actually generated, the stuff we deliver, will not be as high as it could be.

Having too many work items in process results in many distractions for the teams. These distractions cause the teams to lose focus & context switch which has an impact on throughput.

We add too much wood = too much WiP

If we add too much wood, we risk smothering the fire as it can’t get enough oxygen to keep it alight & it gradually loses heat.

As with too much oxygen, if we overload the teams with too much work (over-utilisation), the level of work will be unsustainable & value will drop off as the team members ‘burn out’

We add the wrong wood at the wrong time = Doing things in the wrong order

Different stages of a fire’s lifecycle require different types of wood (fuel).

Starting a fire requires smaller pieces of wood (or kindling) which burn quickly.

Once up & running, we need bigger pieces of wood which burn slower & give off more heat.

If we put the wrong wood in the fire at the wrong time, the fire won’t start or it’ll go out.

Similarly with product development, if we attempt to solve problems at the wrong time, or even the wrong problems, we’re not going to deliver value for our customers.

And so to the ashes…

To summarise this analogy for now…

In order to get a fire which is generating a lot of heat, we need the correct balance of fuel & oxygen. Too little or too much of either will result in a fire which is not as hot as it could be.

It’s the same for product development – we need a balance between the work coming in & going out of the system in order to keep it generating value as effeciently as possible.

If there’s too much Work in Process, you’ll observe a decrease in the amount of value delivered.

What are your thoughts on this analogy? Do you disagree? Do you have a corollary?

Thanks for stopping by,

Duncs

Further research on WiP:

https://www.boost.co.nz/blog/2018/11/manage-project-risk-limiting-work-progress

https://livebook.manning.com/book/kanban-in-action/chapter-5/19

https://www.investopedia.com/ask/answers/051315/what-difference-between-work-progress-and-work-process.asp