In my last column, I provided a brief outline of what I see as the basic goals of Corporate Performance Management (CPM) and briefly discussed the pre-requisites for successful CPM. To summarize, there is a  virtuous cycle where strategy leads to execution and then reporting and measurement tell you how well you are executing on the strategy and leads to changes in strategy and the cycle starts all over again.

The requirements of a good system, among others, includes strong support from senior management as well as broad company participation and an excellent internal reporting system.

We will talk more about the pre-requisites in a later post, but the next topic I would like to cover is what exactly should we measure. There is an old saying that "you get what you measure", and the corollary is that you should be very, very careful about what you measure. I'm sure you can all give me some examples of measuring things that resulted in unintended consequences such as measuring billed revenue and finding out later the sales guys brought in the billed revenue but because of side deals or other intricacies of SOP 97-2, you found out that you couldn't record this as revenue for external reporting purposes (if you have some particularly egregious examples, feel free to send to me at Craig.Bruya@microsoft.com).

A good example of unintended consequences, is something I picked out of recent Wall Street Journal article. During Carly Fiorina's tenure at Hewlett-Packard, she wanted to have HP be regarded as a technology leader and wanted to demonstrate that through patenting of new technology. Here is an excerpt from that article.

Given the poor performance of the stock during her tenure, Fiorina takes credit for fixing the company in other ways-such as by turning the company from an innovation laggard into the "number three innovator in the world in 2005." But by what measure? Fiorina is basing her claim on a rise in the number of patents.

But what Fiorina doesn't mention is why the number of patents skyrocketed. Much of it had to do with a program put in place in 1999 to get HP into the top 10 patent producers. It relied on paying engineers for each new possible filing. At the time, it was $175 for a basic "invention disclosure," $1,750 if it became a patent application, and another chunk of cash and a plaque for an actual patent.  One engineer, Shell Simpson, nearly tripled his salary by working weekends in the first year by filing 120 disclosures and 70 patent applications-at one point taking two weeks off to work on patents full-time.

As you can see, she got what she measured. I'm just not sure that is what she wanted Shell Simpson to be doing.

Here is an example that relates to my own family history. My very first relative to come to the New World in the 17^th century migrated from central France to Quebec after fighting in the French and Indian Wars. My father has done some extensive genealogy studies of the family back to that time.

Coincident to reading the family history, I was studying North American history in school and I learned that the French government was having a very hard time getting people to settle, stay and populate their territories in the New World. To incent this activity they paid people to have children and the more children you had the more money you made.

If you have gotten this far you probably won't be surprised that the first Bruya (spelled "Brouillet" at the time) in the Americas

had 13 children and almost 100 grandchildren. J I generally tend to not believe in government incentives as a political practice, but it is a bid odd to think that I might not exist except for the generosity of the French government over 300 years ago.

So the most important lesson here, is be very careful about what you measure and also critically review your measurements to make sure you aren't inadvertently encouraging bad behavior.

In searching for the right measurements to achieve a balanced scorecard, here are some considerations I try to follow in getting the right mix of things on the scorecard. You can't do all of these things, but it's unlikely the most important ones don't show up in here somewhere.

1. Long term investment returns vs. Short term results. All businesses must perform in both the short and the long term. Your measurement system needs to be attuned to this. You not only have to measure the things that are material to current results, but also the leading indicators that you are being successful in setting the groundwork for growth 3 and 5 years from now.
2. Health of your channels of distribution. Unless you sell your products direct, you need to make sure that you have healthy distribution channels. Are your partners healthy and profitable, is the number of retail outlets increasing, are your franchise operations successful, etc.
3. Products. What products are doing well? Which ones are growing or shrinking, or have good margins or bad margins or are your targeted investments for future growth?
4. Customer segments. Whether you focus on enterprise, mid-market, small business, consumers or all of the above, it is important to segment both your revenues and expenses by these segments so that you can understand whether it makes sense to go after all customer sets. Enterprise customers are where the most revenue is, but it also takes a significant investment in support to keep those customers happy.
5. Market definitions and opportunities. What market are you going after, how big is it, what is your market share? Coke, for example, measures "share of stomach" or how much liquid consumers consume. Microsoft, not surprisingly measures where the PC's and Servers are to determine their market share. It's hard to sell software to someone who doesn't have a PC.
6. Key Objectives. What are the most important things you are focusing on this year? Are you trying to launch a new product, or open a new store or starting in a new geography or entering a new line of business? If you have set up a key objective and communicated to the team that it is important, then you better measure whether or not you are successful. Recently I asked one of our Microsoft Dynamics partners what was new. He told me he was expanding from Microsoft Dynamics GP into selling Microsoft Dynamics CRM. So I asked him if he was separately measuring CRM customer adds or investments in training for CRM and he said he was not because it wasn't material. After a lengthy discussion, I think I convinced him that if it was important for him to get into the CRM business it was also important to measure his progress even if the amounts are currently not material.
7. Your feedback loops. Don't forget to look broadly across all your constituencies. Are they happy? Do they have an opportunity to give you feedback? Do you have a mechanism for capturing and acting on this feedback?  These constituencies include, but are not limited to, the following:

a. Employees
b. Partners
c. Customers
d. Business Decision makers
e. Community

A final consideration is to determine what is measurable and what is not and how much it might cost to measure. Surveying your customers is time consuming and expensive. Maybe you can only do it on a rotating basis or once a year or maybe only your largest customers. It is important to think through the cost/benefit equation for these things. Measuring your revenue by product or customer or geography is probably straight forward (assuming you met my original criteria of having an excellent internal reporting system). But what if you want to measure revenue by each customer segment? That might require an investment in 3^rd party databases in order to segment the customers. Is it worth the cost? Only you can decide.

Next time, we will spend some time talking about some best practices in measurement systems. In the interim feel free to contact me with corrections, disagreements or suggestions about other subjects of interest.

Craig Bruya
Chief Financial Officer of Microsoft Business Solutions