Sunday, November 11, 2007

From Metrics to KPIs

Asking the question "So what?" should be on the top of everybody's mind when either reading a report, or developing a report for consumption. I've been involved in many reporting and analytic product meetings where glossy eyed product managers want to hear all the buzz words like "role based dashboards", and graphical reports that show "trends over time". (my personal favorite is "should we look into using OLAP?") What they fail to see is that if they don't focus on delivering true information to the user, the system will just be a pretty operational reporting solution that won't be able to tell the user anything intelligent.

Lets say you have a report that shows you have accumulated $500.00 of sales in a day. In my eyes, that report relays a metric. Metric tells you what "the number" is. But what does that really mean? It doesn't tell the user if they're doing well or poorly. Are they doing well against yesterday's sales? A week ago? Against a forecast? Nope - it only says you've made $500 bucks in sales in a day.

So what?!

A very simple way to make metrics more meaningful is to come up with some way to define targets to compare them against. In the example above, if the system had a defined target for $600 of sales per day, a report could now show that you haven't met your target by $(-100.00). Seeing that number tells the user a hell of a lot more information, and screams back "THATS WHAT!". By my definition, that would be a Key Performance Indicator (KPI). It gives more then just the number; it tells you how you are performing.

Doing something as simple as defining targets to measure the metrics against will take your BI solution from its basic state to a state where you actually start providing answers to your users.

Jim on Technorati

1 comment:

Ozzie Goldschmied said...

This essay connects me back to something I learned a few years back, which I will discuss in more length shortly. The short message is this - data, information, knowledge, wisdom.

You start with the data which is $500 in sales. You next can compare the $500 to a value and elevate to information, something has been learned or can be learned. Being able to see a trend, predict results given relative accuracy, or having experienced much of the same before yields knowledge. Combining the previous three with an innate sense of what will be is what is named wisdom - knowing things in a wizard like manner.