In selecting the initial 22 measures shown in Figure 1, due heed was paid to the advice of users of existing scorecards as well as consultants. In their view, the metrics should be:
 | objective and unbiased; |
| statistically reliable, with small margin of error; |
| unobtrusive, and not disruptive of work or trust; |
| inexpensive to collect; |
| qualitative as well as quantitative in nature; |
| robust, measuring the things that count most heavily in the foundation's performance; |
| quantifiable, lending themselves to aggregation, calculation, and comparison; |
| efficient, as a group enabling the drawing of many conclusions out of a small data set; |
| comprehensive, addressing all the significant features of the foundation; |
| discriminating, with the result that small changes in them are meaningful; and |
| impervious to gaming by management or staff.(12) |
For each measure we specified long-range goals and, whenever feasible, identified benchmark comparisons. When possible and appropriate, goals are based in considerable part on peer benchmarks. The proposed goals are geared to ensure that the Fund will at least sustain its current level of performance, and have a strong probability of exceeding it.
While Fund operations already routinely generated data for 10 of the performance scorecard metrics, and new internal tracking systems exist for another three measures, confidential audience, grantee, and staff surveys are necessary for the production of the remaining nine metrics on a regular basis. The 2002 Harris Interactive Commonwealth Fund Grantee Survey and the 2003 Harris Interactive Commonwealth Fund Audience Survey, along with the 2005 Center for Effective Philanthropy Foundation Staff Survey, provided baseline data for most of these measures. To generate 2006 data for all of the measures not generated internally, we commissioned Mathew Greenwald Associates, Inc., to survey Fund audiences and grantees confidentially in January 2006 and the Center for Effective Philanthropy to survey Fund staff confidentially again in February 2006.
(13)
