What are you really measuring?


File this under the “Lies, damn lies and statistics” section.

As I walked into the office the other day I looked at a monitor used to display various bits of news and information. One of the things it displayed was the number of service desk calls over the last period. I’m sure my reaction was meant to be, “Wow, the service desk are smashing it by answering all those questions!”, but my actual reaction was, “Wow, our services suck!”

The issue here being you can look at those statistics in a number of ways. If our systems were better and more intuitive we would get less calls, so a statistic showing a large number of answered calls reads as a bad thing to me. We should be aiming to reduce the total number of calls by building better systems. The fact people need to interact with the service desk should be seen as a negative IMHO. That’s not to downplay the work our service desk are doing, but we should be aiming to make it as unnecessary as possible.

So coming back to the title of the post, when people post an assortment of KPIs, it’s worth taking a step back and asking what they mean, or could mean. What is really being measured by these KPIs? Maybe that number you display with pride is actually a sad indictment on your systems. Maybe not. I know what I think. ūüôā

If we were discussing a database we might say with pride, “My system processes 10 gazzillion I/O requests per second!”, but if that is because the database design is bad, and the application is terrible, all this proves is you have created a crap solution. There is nothing to be proud of here.

It’s also funny when people post numbers without context. If I say I’m going to invest ¬£1 million into something you might think that’s a lot in isolation. If I were talking about the NHS, which costs over ¬£140 billion per year to finance, then all of a sudden ¬£1 million seems rather trivial. Context matters!

We get these sort of out of context numbers all the time in tech. They might be benchmark results or manufacturer figures. This may sounds rather odd from someone with a science background, but rather than focusing on the numbers, we should probably focus on the feel. Does this feel fast/slow. Does this feel better/worse. Does this service feel good/bad. I’m not talking about throwing science out of the window completely. I’m talking about putting things into context, rather than just falling for the, “A bigger number means we are better!”, mentality.

Just a thought…



Author: Tim...

DBA, Developer, Author, Trainer.