I have spent the last two days working with a large engineering company, sharing the lessons from one of their large mega-projects.
One thing that became quickly apparent was that many of these lessons were not new. They had been identified many times before.
One of the participants described them as "20 year lessons". As he described it - "we have known about these things for 20 years, yet we keep on making the same mistakes time and again".
This was a case of the organisation knowing what to do and how to do it, but somehow that knowledge not reaching the individual decision-makers in the individual projects. So people ended up making decisions and choosing solutions which were known to be faulty.
The solution, the meeting decided, was to standardise equipment and processes wherever possible. The best knowledge could be embedded into the standards, and if people followed the standards, then faulty solutions would not recur. Then perhaps we will see the end of the 20-year lessons.