“Here’s the problem: We’ve got a line of kids on a hike in the woods. In the middle of the line, we’ve got Herbie. We’ve already taken the pack off Herbie’s back to help him go faster, but he’s still the slowest. Everybody wants to go faster than Herbie. But if that happens, the line will spread out and some of the kids will get lost. For one reason or another, we can’t move Herbie from the middle of the line. Now, how do we keep the line from spreading?”
~ Alex Rogo in The Goal
WHERE DOES DBR COME FROM?
This is how Goldratt introduces us to the concept of Drum-Buffer-Rope (or DBR) in his book, The Goal. Alex Rogo, manager of a fictional factory, has just come back from a scout trip where he had trouble keeping the troop together because some were faster than others. Herbie, “the fat kid” (as Goldratt describes him), got particular attention because wherever Rogo put him in the line of children, he caused problems: either the kids were walking into each other, or some were disappearing into the distance.
Rogo’s children came up with suggestions: “A drummer” was the first. “You have Herbie beat the drum”. “Tie ropes to everyone” was the other suggestion: “You tie everyone together at the waist with one long rope. So, that way, no one could get left behind, and nobody could speed up without everybody speeding up.” The book then continues to explain how metaphoric drums and ropes were used in the factory to keep inventory to a minimum, maximise output and give predictable delivery dates.
HOW IT WORKS
The drum is the slowest person (such Herbie) or process in a system. It is the limited resource, bottleneck or constraint. It is setting the pace. The drum tells the upstream system when to add more items to the system, stopping too many things being in progress at once, or unnecessary inventory as Goldratt would describe it. This rhythm of the drum also gives us the maximum delivery rate because the system can only go as fast as its slowest part. The first step in DBR is to identify the drum. Once you’ve done that, the next thing to do is make sure that the drum is beating as fast as possible, making sure that it is never idle (Goldratt uses the term ‘exploit the constraint’). The example in the scout story is to take Herbie’s backpack off him and share its contents amongst the other scouts so that Herbie can walk as fast as he can.
The rope is what keeps the faster scouts from disappearing ahead of the rest of the troop. It is choking the faster parts of the system to stop them going too fast and building up unnecessary inventory. As with the scout troop, having part of the order completed early doesn’t satisfy the customer; it’s only when all of the order (or scouts) get to the end that matters.
So why are buffers needed? Why isn’t it just Drum-Rope? In an ideal world, the drum would keep on beating, everything feeding into the drum would be delivering a steady stream of materials, and everything downstream would be receiving the drum’s output and getting it to the customer. The drum would be working at capacity and would never be idle. Making sure your drum is working at a sustainable capacity would be your only concern. But, in reality, input to the drum might be disrupted (meaning the drum wouldn’t have anything to work on) and delivery would be impacted downstream. So we add a buffer. Buffers differ depending on who is writing about them. Most articles on Goldratt’s Theory of Constraints refer to fixed time buffers being used to ensure that the shipping schedule isn’t affected. In knowledge work, we may add a buffer before the drum to ensure that the drum is kept busy, but we may also add a buffer after the drum to give some flexibility to the downstream processes. Think of it like one of the scouts stopping to tie a shoe lace: you wouldn’t want everyone suddenly crashing into him or the scouts ahead suddenly being yanked backwards because of a small delay. In effect, a buffer stops the process being painfully dependent on the other parts and gives a bit of flexibility.
HOW DOES IT APPLY TO YOUR WORK?
Think of your roadmap being like the path walked by the scouts. Imagine each product backlog item being like a paving slab, and they are all laid ahead of you with the top priority being the nearest slab. The passing of your team over a slab represents them working on and completing that work item. You do not want some of the team to be miles ahead of the pack, choosing a route that they think is best. If they have walked ahead and taken the wrong route, they’ll need to turn around and walk back. Keeping everyone together (using a rope) means the team can proceed in unison and take the best route options based on the latest knowledge they have at the latest opportunity.
Your equivalent of the drum is the slowest part of your system. Having low work-in-progress limits and monitoring cycle times within your system will help you find where your drum is. The rhythm that is beaten on your drum is your delivery rate: how much work comes out of your system each day (or other unit of time determined by you). There is no point of pulling work in at rate faster than the drum beat, because it will just clog up your system.
The need to employ buffers depends on how smooth your flow is. A cumulative flow chart will give you some idea of whether the sections of your process are smooth or jagged – implying whether you need no/small buffers or large ones (the more variation you have in your flow, the bigger buffers you will need to compensate). Monitoring the consistency of cycle times, using charts such as a scatterplot, will also give you useful insights.
Limiting work-in-progress is an effective rope as it stops parts of the team shooting ahead. The shorter rope length being a metaphor for a smaller WIP limit.
I’m not proposing that the DBR concept is a fail-proof method within knowledge work, but it is a good concept to make teams think about ways of working. It is important to remember that your system is a living thing and may change from time to time: the drum might get moved to another place, the ropes might need to be shortened or lengthened, and buffers might need to shrink or grow to stop your “walk” becoming a bore.
Enjoyed this post? You can now sign up to our mailing list to receive a monthly round-up of our posts, features and competitions.
Please note that we will only use your email address to send you a summary of our posts, features competitions; we will not spam you or pass on your details to any third party.