Did Somebody Say Strategy?

Howard Dresner

My friend and mentor Howard Dresner has launched Dresner Advisory Services: http://www.howarddresner.com .  Here’s a good interview with him from yesterday’s Investor’s Business Daily.

You can also pre-order his excellent book, “The Performance Management Revolution” on Amazon.com. It comes out in mid November.

And read more about the book on Wiley’s site here.

xPM and Incentive Plans

Executing on your corporate goals depends on the behavior of your people.  You want your incentive plans to drive that behavior.  You want different behaviors in different parts of the organization (we think of it in 8 distinct functions, in 3 different layers – strategic, operational, and tactical).  AND the drivers of behavior at all 24 of those intersections (8 functions x 3 layers) MUST align with your corporate objectives.

Paralyzed by Feuds?

There may be at least three ‘management system’ camps in your organization: the ERP camp, the Data warehouse/BI camp, and the Performance Management camp.  Some are run by I.T., some by Finance, and still some by Operations (or key Business Units).

Each camp thinks they have the solution to the enterprise information systems strategy, yet each could be contributing to building silos, splitting environments, and wasting resources.

This can be a paralyzing situation when each group starts to compete for senior management attention and funding.

Strategy starts at the top

A project manager asked me if there should be a corporate strategy defined, communicated and translated into department strategies/operational terms. Or did I think the bottom-up approach can work? Can departments work out their strategies independently of each other and then consolidate that  to the top?

Among other things, I gave him a quick illustration of how bottom-up can go wrong:

IT Portfolio metrics (in the Financial Serivces Industry)

A VP at a large financial services institute asked me if metrics-based approaches to evaluating financial services IT value is a valid exercise.

He was interested in the successful use of metrics to evaluate the status of an IT portfolio supporting financial services firms/banks/brokerages, etc. He wanted to know what types of metrics (if any) have been the most useful, and if it was more helpful to have a handful of measures, or were hundreds required in order to be truly meaningful. He also wanted to know what of confidence organizations have in their collection.

My answer was that he should be looking at the business metrics (not pure IT metrics such as up-time, call closure, project budget, etc.) and that there are some common financial services metrics for different parts of the organization (eg: Assets Under Management, Non Interest Income & Expense, Cost of Interest, Credit Quality, etc.).

The trick is to show a direct correlation between the IT portfolio and the business performance metric. If you can show the correlation, then it is an extremely valid exercise to show IT value. For example, if you have an IT project in the portfolio to implement a management reporting system (say a Credit Risk dashboard) that will show branch managers and associates credit risk vs. household profitability, and they see that it costs your company money to serve below a certain FICO, and they "fire" their customers (take action) and improve branch profitability, then the ROI for the initiative is positive. Once you have a win with actuals data, you can start to plan for branch profitability targets and monitor variance (another IT initiative). Maybe a branch is consistently underperforming, so it gets closed (more positive ROI). And so on.

Gartner Group says that the right number of metrics to juggle for a given role is about 9 (Gartner 2007 BI Symposium, Chicago). You would want to identify those 9 for unique roles (the branch managers' 9 is different from the CFO's 9) - but still interrelate them.

One way to improve confidence in the numbers is to start reporting them (and even tie variable compensation to them), then people pay attention to them and begin to fix problems such as data quality and master data issue. You can even start to benchmark performance internally and externally.