The story I wrote for our SQLskills Insider newsletter on Monday last week resonated so well with people that I thought I’d publish it here on my blog too (which I only do rarely). I try to mix things up in the newsletter so it’s not all about SQL Server technical matters; people seem to really like career topics too. You can sign up for the bi-weekly newsletter here (and get all the 100 past issues too) – more than 12,000 people get it now!
Here’s the story, from the newsletter section I call Paul’s Ponderings. Enjoy!
Last week Jonathan pointed me at an interesting story about a psychology experiment involving monkeys and bananas and the reinforcement of constrained, negative thinking. The experiment actually never happened, but the story is quite illustrative. You can see a graphic of the story here: http://i.stack.imgur.com/MyQki.jpg (totally safe for work), and I’ll paraphrase quickly below:
- A group of monkeys are trained to know that if any one of them attempts to climb a stepladder with a banana at the top, cold water is sprayed on all of them (i.e. temptation is prevented by the spraying of cold water, so the group prevents any individual from trying to get the banana).
- When a new monkey is introduced, and tries to get the banana, all the other monkeys prevent it as they don’t want to be sprayed with water. The new monkey learns that it can’t get the banana.
- Eventually all the monkeys are replaced, and they all know not to try to get the banana, but none of them know why, as they’d only been told not to get the banana by the others.
It’s a really interesting story of how conditioning can become groupthink (even if the experiment didn’t actually happen).
There are obvious parallels in human society, and specifically in the work environments of many of you reading this newsletter. Let me explain:
- A new person (A) joins a team. The new person has a great idea for how to do something differently/better, and everyone else on the team prevents them from expressing their opinion because that won’t be allowed, it won’t work, and/or they could get into trouble (e.g. from an intransigent and influential boss/architect/senior developer).
- Eventually all the original people leave the team and it’s only A and the people who joined the team after A that are left. They all have the ingrained knowledge that they can’t try anything new, or try a specific technology because they won’t be allowed to etc.
In other words, being constrained to incumbent technologies and methodologies becomes “the way it’s done around here, because it’s always been that way.” It doesn’t matter if the wider world knows that’s maybe not the best way to do it, the historical groupthink wins out.
We see this with new clients over and over, and it can be really difficult to educate the client team as to different (usually better) ways of approaching a problem. Sometimes it’s just one person who stymies all thoughts of innovation or change in the team, and sometimes it’s the collective groupthink that does it.
In fact, there was a client (years ago), that Kimberly actually “fired” because of this. One meeting was enough but she told her contact (“the management”) that she’d try again and meet again for a second session. In her post-mortem with management, her main quote was – your problems are more than technical (specifically person “X”) and until that’s resolved, I can’t help you.
Here are some simple examples of ingrained behavior we’ve seen and corrected:
- Always use NOLOCK to work around blocking and deadlocking problems (and then it’s used everywhere, and no one even realizes the negatives)
- Only ever rebuild an index to remove fragmentation, without considering FILLFACTORs or trying to reorganize rather than rebuild
- Always use a heap and never use a clustered index because heaps are faster
- When creating a primary key, always use a clustered index as that’s the default so Microsoft must know that’s always the right way to do it
- For each column that’s used in a WHERE clause, create a single-column nonclustered index
- Always use OPTION (RECOMPILE) to prevent plans being put into the plan cache and taking precious memory away from the buffer pool
- Always use sp_executesql to put a plan into cache for ad hoc statements so that their plans get reused
- Always create one file per processor core for tempdb, because that was the Microsoft guidance for SQL Server 2000
As I hope you can see, these are all very blinkered (and in some cases downright wrong) views on a variety of SQL Server topics. There are cases when some of these are the right thing to do, but not ALL the time just because ‘that’s the way we do it here’.
Does this remind you of anywhere? Or of anyone?
It’s a dangerous way for a team to work because it can lead to all kinds of problems (such as poor performance, failed disaster recovery, good people leaving the team, business failure) that can be extremely hard to fix unless someone takes a stand or someone from outside the team helps to break the impasse. And even then, it’s a delicate process to provide education and slowly change the thinking of the team, or of that one person who dominates the team’s thinking.
Call to action: Consider the environment you work in and whether the situation described above is how your team operates. Are you the person to take the stand to try to change the groupthink? Do you do it or do you move on? Or are you the person who’s unreasonably dominating the team’s thinking? Do you allow yourself to be changed or continue in your intransigence? It’s a difficult situation to recognize you’re in, whatever your role is, and a hard choice to decide what to do about it.