Explaining Herbie…

I have been asked by a friend to write an article explaining the concepts of Herbie from the book “The Goal” in a more accessible way. The book itself is amazing but in my opinion the language used in the 5-focusing steps has tried too hard to be scientific and precise that it makes it harder to understand. My annotations simplify the terms but likely lose some of the precision of the wording, I offer this as an introduction to the topic but encourage reading the book to get the full impact, I cannot do it justice here.

Herbie is a character from the book “The Goal” he was used in the story as a metaphor for a system constraint and the impact it has on your delivery. I’d highly encourage you to read the book but hopefully it is not essential to be able to follow the explanation, the purpose of the metaphor was to make the thought processes relate-able, hopefully this helps.

The 5 focusing steps

1. Identify the system constraint
2. Exploit the constraint
   (Make the best use of what you currently have)
3. Subordinate everything to the constraint
   (Make decisions based on the constraint, the system should run at the pace of the constraint)
4. Elevate the constraint
   (Make improvements to the constraint)
5. Repeat the steps

Step 1: Identify the constraint

We need to figure out which part of the system is holding us back the most – limiting our ability.  We should know that only an improvement at the constraint makes a difference.

Herbie: In the example of Herbie, he was observed as holding everyone back he was slow and the rest of the group had to regularly stop to wait for him to catch up or were racing ahead.  In the case of Herbie the constraint was evident through observation.

Reality:  There are different ways to identify a system constraint sometimes observation and experience is sufficient, or metrics can help.  The simplest way is to look at your work is it piling up somewhere? Do you have a lot of stories in Dev Done, or waiting on PO approval or needing wire-frames? Piles of work are a sign of a constraint.

Another common constraint are blockers, if you are consistently waiting on another team, another person or another resource this is a sign of a constraint, we may use blocker clustering or other root-cause technique to see if there is a common theme to delays.

The third most common technique is to review cycle-time for stories, look for stories that are taking longer than others see if there is a pattern, common aspects, this could be a constraint that may be improved.

Step 2: Exploit the constraint (Optimize)

Before we consider adding capacity, we should make the best use of the capacity we already have. We ‘Exploit’ in this sense means doing everything possible to use the constraint to its fullest capacity. This is also the easiest and generally cheapest improvement you can make to your system.

Herbie: In order to maximise the performance of Herbie, the contents of his backpack were shared out so he could concentrate on the most important task of walking.  Making the others slower (less efficient) by carrying his stuff was unimportant as Herbie sets the pace for the entire group.

We could also have ensured that he got his lunch first so could start sooner, and others could have covered his chores so he was more rested – our goal is to get the best out of Herbie on his most important task – walking.

Reality: In our systems we need to identify what the most important task is, typically this is getting a story to done and the software in the hands of the user.  So when you identify your constraint look for anything they are doing that is not helping to move stories or for anything they are doing that someone else could do. Are they filling in paperwork that someone else could do? does this task really require their expertise? Could you make them a coffee so they could focus. (This is not encouragement to over-work, breaks socialising and generally a healthy working environment always takes priority).

Step 3: Subordinate the non-constraints

The job of all non-constraints is to subordinate their decisions to the constraint’s needs.

That sounds extreme wording but essentially it means that no mater how fast you go, if you are dependent on someone slower you are wasting your efforts, so work at the pace of the constraint, you have reserves so you can catch-up if necessary.  Use the time saved to help optimize the rest of the system.

We do this as step 3 because it is more disruptive and more costly than step 2.

Herbie: In the story Herbie was put at the front and the rest of the group were told not to overtake him. When Herbie was able to go faster there was no one in front to delay him so we got the best from him and as the others were able to go faster overall they were able to make up any delays. The goal was only achieved when everyone was done so there is no benefit in getting ahead.

Reality: Don’t let work pile up, if we limit Work in progress we focus on flow rather than utilization, put effort where it has most impact on the whole system.  In the case of dependencies on 3rd parties, could you give them advance notice so they are better prepared and can respond sooner. If you are dependent on a 3rd party could you see if there is a time/day they can respond more quickly, could you tailor your process to be more convenient for them? Tailor your process to maximise the effectiveness of the constraint, and this is the tough bit – even if that means making the process worse for others.

Tailor your process to maximise the effectiveness of the constraint,  even if that means making the process worse for others.

Step 4: Elevate the constraint

Only once we’ve completed the previous steps does it make sense to add more constraint capacity, and thereby increase system performance. Because adding capacity is tremendously expensive in terms of time and money, we do it as a last resort, not a first step.

Herbie: In the case of Herbie, we could get him doing fitness training so he is quicker, or buy him a bicycle.

Reality:  This is typically where you spend money, hire more people, get better/more equipment. Invest in training or education. Remember improvements here improve your entire system so the investment must be calculated based on system income not local costs. Maybe you ship parts by air, maybe you fly in an expensive expert, maybe you lose money on this aspect of a project so that the project as a whole makes more money. Keep the focus on the gain to the entire system and not get distracted by the cost at one stage.

Step 5: Return to step 1

The inevitable result of the first four steps, and the reason this is a “continuous” improvement method, is that the constraint will inevitably move somewhere else.

This step insists that you start back at the beginning, and don’t let inertia become the constraint.

A second example of the steps being applied…

Identify

Access to overworked Doctors in an emergency – constraint is a shortage of Doctors:

Exploit: Make the best use of what you currently have

• Ensure Doctor is not doing any activity that someone else could do. All admin work is removed, prep work done by nurses or other staff.
• Vacations/time off is deferred,
• Doctors may work overtime so long as it doesn’t risk the patient or cause the Doctor to quit

Subordinate: Make decisions based on the constraint, the system should run at the pace of the constraint

• Any non-urgent treatment is stopped until emergency is over.
• Cases are prioritized so Doctors are only working on the most urgent cases.
• Patients are only ‘prepped’ according to the capacity of the Doctors.
• Patients are prepped in advance so that the Doctor is never waiting.
• It may be quicker for a Doctor to move between patients rather than them coming to the Doctor.

Elevate: Make improvements to the constraint

Now we look to increase the number of Doctors or their effectiveness

• transferring Doctors from other areas withing the organisation
• recruiting more Doctors even at inflated rates or short term contracts,
• any increase is an increase so we can hire someone that is semi-retired or part-time.
• We may automate activities or buy tools to improve the efficiency and effectiveness of the doctor.
• We could train Nurse to perform some of the Doctor’s activities

Repeat

We look and see whether the Doctors are still the constraint or whether there is a new constraint.

Summary:

There may be a little ambiguity with some of these improvements, e.g. A Doctor moving between patients may be considered elevating or subordinating, but it doesn’t matter it is an improvement, it is a thinking process and some changes don’t neatly fit into one category, don’t let that distract you from the thinking process itself.

---

The Goal by Eli Goldratt

This book uses a production plant as a metaphor for delivering value to your customers. The emphasis is on the manufacturing equivalent of delivering working software, doing it regularly, and that delivery is not about development, OR production OR deployment it is about everything – essentially encouraging Dev Ops.  It also has a lot to say about how the only Done that matters is to have it in your customers’ hands.

Read it!  This is hands down the best book on Agility I have ever read.  There are many others by this author all of which are great but this is my favourite.