There's lots of great material out there to familiarize you with the kinds of things PA measures: KPIs, Metrics, Measures, etc, etc. Smarter folks than me can teach you about these things.
But as my customers continue to expose me to their PA deployments, and ask my advice, I'm struck by a common weakness. So many treat PA like they've treated most reporting: "give me a bunch of data, grouped different ways, and I'll spend effort looking for important insights". That's a passive, accidental strategy.
Its also how we wind up with PA dashboards that split Incidents by priority over time and age in 4 different visualizations that nobody *REALLY* knows how to react to.
Come to the dashboard. Stare at the visualizations. Go to the coffee machine. Hope the caffeine rush takes the emptiness away. Wish for a fire drill, if only the shrieking wail will remind you that you're alive.
There's a better way my friends. Stop being sanitized. Stop being "objective", in the sense that your emotions take a second seat. Embrace the *humanity* that we suppress in reports.
HACK: BRING YOUR EMOTIONS
For those of us passionate about our jobs, what typically happens when we socialize with work peers? Someone, probably after a drink, leans forwards and says "You know what's BS?!"
And *THAT* my friends, is the secret to Performance Analytics. Being judgmental. Black and white. Opinionated. There's a whole world of decisions we've already made about performance. You know what's BS? Waiting a full day for a response.
You know what's BS? Spending money on "process improvements" only to have the process take longer.
You know what's BS? That this object has degraded in the same way each month, for a year, and its still not fixed.
You know what's BS? That there's work out there supposedly "needs" doing that hasn't been *TOUCHED* in 30 days!
You know what's BS? That this thing hits 3 different groups before finding the right home.
You know what's BS? Being told this system is mission critical, but its backlog of work grows by 1/5 an FTE every week.
Want people to take notice of your PA build? Think only in terms of bad and good, not "possibly interesting". Use those freaking threshold and target records. Draw a line in the sand. Where? IN THE SAND!! It doesn't matter if your target is good (yet). What matters is if you're heading towards it. 1 week of watching progress and you find out if your target is sane or not, then adjust it.
Performance Analysis is measuring the movement towards good or bad. It *requires* emotional investment. We don't need another gentle undulating cosine wave form. We don't need another volume over time breakdown. We want measures that drive reaction/reward: THAT is bad / THAT is good.
Build for humans. Build for emotional reaction. Do so, and your PA efforts will be celebrated in the halls of Valhalla.
5 Comments
