An organization’s assigned portfolio defines where relationship managers’ time and energy will be focused. Thus, it pays off to critically evaluate and optimize the makeup of your organization’s gift portfolio. From a fundraising perspective, spending time on prospective donors that don’t have the capacity to make a meaningful gift is an inefficient use of valuable and limited resources.
You may already know that predictive modeling can unearth new major gift prospects. But how can it help us improve a major gift portfolio that is already assigned?
In essence, predictive modeling can help evaluate two factors that are key to fundraising success: the portfolio’s quality (having the right donors assigned) and the portfolio’s performance (capturing the available fundraising potential). The following steps provide a high-level overview of how CCS uses predictive modeling to help our clients optimize their major gift portfolios.
Step 1: Evaluating Quality
A quality major gift portfolio contains prospects who have both capacity and affinity. This means that the prospects not only have the financial means to donate to charity, but also have an interest in giving to your organization specifically.
A nonprofit’s CRM data can reveal a lot about affinity. CCS builds predictive models for our clients using this internal data, leaving us with a score that estimates how likely that donor is to give to your organization.
To estimate capacity, we conduct external wealth screening research. Each wealth screening service takes a slightly different approach to produce gift capacity estimates. In general, these services use assets, real estate, income, and public giving data to generate a value that estimates how much money a particular household can donate to all charities over a five-year period.
To visualize the quality of a portfolio, CCS often puts these affinity and capacity data points together to build a quadrant chart like the one below. Prospects that both score well on the predictive models and who exhibit meaningful gift capacity fall in the upper right quadrant of the chart.
Illustrating the quality of a portfolio in this way brings up key questions, such as:
- What proportion of our portfolio is in the upper right quadrant? In a perfect world, the entire portfolio would fall there. In the real world, portfolios with at least 50% of their prospects in the upper right quadrant are considered to be high quality.
- What prospects are of lower quality than we thought? Often CCS recommends that some prospects be considered for removal from the portfolio for optimal use of gift officers’ time. Laying out capacity and affinity together can pinpoint who these donors are.
Step 2: Evaluating Performance
In addition to assessing your portfolio’s quality, you also want to assess your portfolio’s performance. A portfolio can contain the most ideal prospects and still not live up to its potential in terms of actual dollars raised.
We can estimate how well a portfolio is performing by using a metric called the capacity capture rate. Capacity capture rate is the ratio of a prospect’s last five years of giving to the minimum value of their estimated gift capacity range. The table below groups prospects into minimum estimated capacity ranges and computes the average (mean) and median capacity capture rate for each of these groups.
In the case of the client example above, we know there are outliers who unduly influence the average. When the average capacity capture rate is meaningfully different than the median capture rate, there is a good chance that there are some outliers. Removing the outliers can help us get a better sense of the underlying capture rate for the portfolio as a whole.
CCS typically sees an average capacity capture rate of 10% and a median capture rate of 1%. Values greater than that indicate a portfolio that is performing well. However, even if the overall numbers indicate good portfolio performance, digging a bit deeper can uncover more valuable information.
The chart above illustrates the capture rate ranges. This chart shows some key pieces of data that can help inform segmentation and prioritization of the portfolio, including:
- The percentage of the portfolio that hasn’t made a gift in the last five years
- The percentage of the portfolio that is under-performing in the context of a donor’s gift capacity
- The number of prospects who are giving more than 100% of their estimated gift capacity, indicating that the wealth screening vendor has underestimated the capacity
Step 3: Bringing Quality and Performance Data Together
The chart below shows a segmentation we recommended to a recent client who was considering their portfolio’s quality and performance data.
Understanding the quality and performance of your gift portfolio enables making data-driven decisions to optimize the portfolio. It will allow you to identify some prospects that should be removed, others that need to be further researched or need new engagement strategies, and others where you will want to stay the course since the current strategy is working.
By optimizing major gift portfolios, fundraising teams can ensure that their limited time and resources are being expended to yield the best results. To learn more about how predictive modeling can help your organization reach its fundraising potential, explore our Data Analytics service offerings or contact firstname.lastname@example.org.
About the Author
John Sammis, Senior Data Scientist, brings more than 30 years of experience with statistical analysis and predictive modeling. He has devoted more than 15 years of his career to charities, universities, hospitals, and other nonprofit institutions, helping them produce models and use the results to achieve their fundraising goals. At CCS, John helps philanthropy and fundraising professionals apply leading-edge data analytics tools to address specific organizational objectives. As a cornerstone member of the CCS Analytics team, John is the primary producer of customized predictive models for CCS clients. He is an expert in the areas of exploratory data analysis, interactive model building, model diagnostics, and data vetting and cleaning.