Book Indexing and Software for Indexes
Optimism feeds into the results of engineering planning and execution often resulting in massive delays and overspends. Many defence projects and also major civil engineering projects display these characteristics. Closer to home, many do-it-yourself and hobby projects take longer than planned.
Exceedingly successful projects where minor failures in the overall scheme of things are perceived as disasters. These characteristics are observed with volume car manufacturers and aircraft enhancements. There are social services, contracts and trade deals that go for years with out a problem then suddenly everything appears to collapse and the headlines are full of it.
What's Going On
When updating the design of a 19 inch rack which was part of a larger piece of test equipment I never understood why the development and manufacture of the subsystem was taking as long as the rest of the test equipment. I'd hold regular meetings and would be updated on progress. Everyone reported on their knowledge of what was going on and no-one was consciously trying to mislead. The next month there would be new updates and the new delivery date would move by about a month. The time to completion became a constant.
Many of you who have either been involved in large contracts or have observed such
things as the development of nuclear fusion will be familiar with the symptom.
How can it be understood
I wanted to understand what might be going on. Using well known principles from the quality world I tried to develop a simple formula which could provide some sort of explanation.
AQL (Assessed Quality Level) and Six Sigma are aimed at designing to precise levels of accuracy and repeatability. The measured level of quality normally represents a corporate quality level resulting from long term investment in corporate knowledge, capital investment and individual training. A company operating at a true Six Sigma level is very special and a single failure makes headline news.
A six sigma success rate is 99.9999998027 percent. The failure rate is approximately 0.2 per million. Produce 5 million cars and one will fail on test. Write 5 million lines of computer code and one will contain an error.
The quality level can be written as (number of failures)/(Total items).
Maturity (You won't find this in text books!)
A maturity level can be written by inverting the formula and using logs (base 10)
Product maturity = Log10((total items)/(number of failures))
The six sigma example above will have a product maturity of 6.7.
So, how do you get to this level of product maturity. The answer is by following a process. Normally the process is iterative and it may take many years for a car to reach the maturity we expect today.
The process maturity can be characterised using the same formula as product maturity.
Process maturity = Log10((total items)/(number of failures)) and needs to be understood as the level of success in a single design cycle.
The magic of this approach is that it is possible to write down the number of iterations that may be required using a known process maturity to reach the required product maturity.
Iterations = Product maturity (target) / Process Maturity
Observation suggests that the normal procees maturity within an organisation is about 1. Yes you read it right, only one.
To reach the required six sigma product maturity is likely to require seven iterations!
It doesn't mean complete iterations but some part of the product or process needs to be changed resulting in collateral change, program delay and cost.
Light Bulb Moment
Once I'd got this far I understood why so many colleagues were still reworking designs they may have been involved with for may years.
You can understand why Tesla is having so much difficulty getting production right.
Also Diesel emissions, Boeing 767 Max, Rolls Royce Trent and many others.
The principles here are only a mental guide and can't be used for planning but can help understand what might be going on. It may a also help with understanding what may happen when a well developed process undergoes a major disruption. For example the impact of Brexit or even a single trade agreement.
Decide how to measure maturity. Count documents, components, words, miles of railway track, houses, it could be anything.
Record the number of defects in a period. It could be a year, a program phase, anything. Calculate product maturity. Calculate process maturity.
Define metrics based on these measurements. Monitor and improve.
External events can have a major impact. Obsolescence, trade conditions etc.