The Data Studio

"Best Practice" - Say It Only If You Have Evidence

In Words You Should Never Use I trashed the term "Best Practice". I did this because the term "Best Practice" is widely misused in the computer business as a selling tool rather than something to improve the effectiveness of the way we work and the results we produce.

As usual, I think you'll find it's a bit more complicated than that.

Suppose you were looking at how to organise a prison system. You could look at prisons around the world and study their practices and outcomes. Practices you might be interested in could relate to:

Outcomes could include:

Suppose you then analysed the results across all the prisons studied in all the countries. You might be able to come up with consistent guidelines about what practices corresponded with what outcomes. Your study might be more likely to reveal correlations than causes, but you could try moving to those practices that have been associated with better outcomes and continue the study to see if they do actually have the effects you predicted from the correlations.

After doing all this you would be justified in stating what the "Best Practices" were, and encouraging more prisons to follow them. You should ensure that the studies continue, in case other changes in the world affect your conclusions or in case a correlation you found turned out to be less than consistent when monitored in a larger number of prisons and countries. You might also find new pactices to measure and identify new outcomes that have become interesting. As a result, the "Best Practices" would probably change over time.

The term "Best Practice", in my opinion, depends on thorough, diligent research and a strong base in evidence.

We should still try new things, absolutely! But we should not use the term "Best Practice" for what we are doing if it is just somebody's idea.

When I was a junior consultant, with expertise in quite a narrow field, I was asked to write "Best Practices" for the processes around the tools supplied by the company I was working for at the time. I felt a bit uncomfortable about it but I did as I was told and my document became the "Best Practice" for the company. That was completely unjustified, since the document was simply my opinion, with the tiny evidence base that I had found in my own work at best, and pure speculation at worst. If it had been called "Some Ideas You May Find Helpful" that would have been OK, but to call it the company's "Best Practice" was dishonest. I hope it didn't do any damage.

Often, companies produce "Best Practice" documents because it is thought to give them credibility and authority. Sometimes the "Best Practices" are designed to favour the company over its competition. I have seen misuse of the term "Best Practice" many times, and that is why it is a term I prefer to avoid.

This looks like an important book on the subject: Building State Capability. I plan to read it and may then add it to Quite Interesting.