Robust Ready

Picture of Adam Bahret
Adam Bahret
Screenshot 2024 05 22 at 5.33.08 PM 1

The Lightning Fast Reliability method has a motto, “Robust Ready and Risk Guidance.”

“Robust Ready” means we create a robust design first. We don’t measure reliability first, why? Because the reliability of your prototype is not good! Prototypes have crappy reliability so why are you measuring it? But we always do this as our next step after seeing our prototype sputter to life and run for a little bit. The reason is this. After a quick congratulations, leadership wants to know “Is it reliable?”

Your annual review may be influenced by how few issues you report each week after that question. Here is what you do. The next week report this, “We know what the reliability is for the design. It’s crap.” Then add “ We just saved the company $100K in testing and six weeks of program time, you’re welcome.” Please don’t do that, I don’t want to be responsible for any sudden career redirections.

Guess what else? Your next iteration is going to have crappy reliability as well. Remember good reliability isn’t running it for ten days and having no failures. Good reliability is in the range of 95% and 99.99% in adverse use cases over many units over many years. A few of you are arguing my point as you read this “We got 10,000 hours on the bench right out of the gate with no issues.” Well, I’m not convinced you have a good design. Do you know why things appear ok early in some programs? Because we are human and the term “Don’t let the fox watch the hen house.” is a common saying for a reason.

If we start measuring reliability early at the request of leaders, eager to keep the program moving quickly, something miraculous happens. Our not-so-good (from a reliability perspective) design starts performing amazingly. All we had to do was make the test easier, open up our acceptance specs, or explain away failures as “one off’s” and “outliers.” I can see you smiling right now because you have seen this. It’s human nature. When being judged in a hopeless situation you want to at least survive long enough to get it right later.

So we established that all of you make crappy designs, are liars, and probably have an anxiety disorder. Well, you’re in good company because “ Hi I’m Adam and I have a hard time with creating robust designs in freight train fast product development programs too.” (in unison) “Hi Adam, welcome to the group.” So let’s change the game together. Let’s make it fair and a process that produces the right product. The right product is one with a precise blend of reliability, new technology, cost point, and time to market, that grabs the market share we want.

How do we start? We learn how the design doesn’t work first. We have shown that the design works in nominal (ideal) conditions when we turned on our prototype. This is critical to do first. We can’t work on reliability if our thing can’t do the thing it is supposed to do. This is an important checkpoint. Believe it or not, I have seen designs go into reliability trials even though they couldn’t satisfy design requirements. That freight train stops for no one. We don’t need 10,000 hours of run time at exactly 25C sitting on cotton balls in a room where people aren’t permitted to speak loudly. That data has no value. We already know it works, let’s move on and learn how to make it better.

Remember the point of this product program is to get thousands of these products in the hands of customers. Those people who don’t read manuals, knock things off of counters, leave stuff in the rain, and think the first step to fixing any issues is to punch it Fonzy style. What we need now is to see it not work. It is so critical that this is our next step. Without knowing more about how it doesn’t work we are starved for the input we need to improve the design.

Once we see it break, we want to see it break again and again and again. This is going to be difficult when leadership just wants weekly reports where success is measured in not having issues. They have to be on board and flip how they think as well. They need to be disappointed when you just keep coming back each week showing them a perfectly running system. This should be very suspicious because new designs have crappy reliability, remember. None of us are that good, Edison wasn’t “I have not failed, I have just found 10,000 ways that won’t work.”

The bottom line is that the mindset, top to bottom of the organization, has to be that your product has thousands of failure modes. They are there dormant, waiting for the right combination of conditions to be evident. The only question is if you find them or let your customer find them. Because on a long enough timeline, they all emerge. So unless you plan to hold onto your product forever, and only run it in the lab, the customer gets almost all of that timeline. Your mission is to find as many of those failure modes as you can while you have the chance. That is the Robust Ready principal.

Check out  the second part of the method, “Risk Guidance”


Share this post