Tempered Lean, Undesirable Effects (UDE's) Part #2

Previously I talked about Experiments and UDE's (Undesirable Effects) in processes, but JUST the interdependent ones.

We looked at experiments and briefly the PICK process for choosing a good experiment, and also trying to surface all undesirable effects (UDE's) prior to conducting an experiment.

Now we will look at options as to how you apply these experiments in a production environment (you could also have this in other processes like offices, IT departments, construction sites, and even in sales).

The Normal solutions

There are two basic solutions which I have seen used; the first solution is to simply rework the entire area or process every time there is a significant change - doing it offline to tie-in later . This obviously has a number of benefits. Firstly, it avoids having to make a lot of small changes - you just design from scratch using first principles of process design. This stops quality control being missed at each step change with the intention of reduced waste. Secondly, it avoids downtime which costs time and money and can be optimized completely offline from the primary process.

The second solution which I have seen is to change the machines and processes which produce the products incrementally. These changes are a lot more tricky to interweave into the current state of the process and require a lot of time and resources but it more closely fits with the Lean concept of "ongoing" improvement. It may take a bit longer as many experiments may need to be done over a longer time, but, if requirements change you haven't sunk a huge capital cost in a completely new production line and you can remain flexible and agile and most importantly, should it take you away from throughput, it's still easily reversible!

The Usual problems (UDE's)

Both experiment approaches/solutions have problems. New parts and processes can be quite expensive to make. They often require new tools to create the parts, training of personnel, and/or rewriting procedures. They also have to be manufactured and checked so they are often slow to become available to the market. If you take too long, you may actually miss the market or potential to be first mover, second mover, or worse.

Changing the machines is also expensive. It may involve retooling or buying new machines. These new machines may also have to be serviced on a regular basis and this can be expensive, especially if it's less familiar. Another reason is that as machines become more complicated, they are increasingly becoming closed ecosystems where you may not have expertise or to even be allowed to work on core functions or future upgrades in-house.

The right-to-repair movement is fixing some of these issues in certain industries, albeit, slowly. Much more slowly than a company may wish to move to keep in step or ahead of their market.

The Real alternatives

The main problem with changing the machines is that they can take a long time to change - and lead times seem to be getting worse - "Yes, you can totally have that machine delivered in September.....2024". The technology necessary to change the machines may not be widely available. Changing of major parts also requires downtime. This can affect production and results in costs being higher.

What about incremental experiments that have a higher chance of success? That could work! The benefits of re-working your own internal processes is twofold:

1.) First, you get to truly know your process and can immediately see if the changes gave you the expected results.

2.) Second, you have accounted for many of the potential undesirable effects in your process and thus have a higher chance that it will be a successful experiment.

3.) Third (bonus) benefit, you will have invested time in developing your people in how to make good decisions and how to move towards the goal while improving morale.

So, as a conclusion, see that easily reversible experiments where you have accounted for as many UDE's as is foreseeable can be a great strategy in Lean and the process of ongoing improvements that take you towards your goal as well as improve thinking processes and morale.

What Experiments have you tried?

What do you see as roadblocks to changing your processes?

What do you use to pick which experiments are of the highest benefit?

I'm interested to know.