Annual Giving Analytics (Part 2)

The previous post in this Annual Giving Analytics series focused on how analytics can be used to further the goals of Annual Giving, but how exactly is it done? This post will document and describe a few methods and approaches that can get you going on furthering your use of the analytics for annual giving at your institution.

Annual giving is primarily interested in three variables that can influence the dependent variable of how much a particular constituent is likely to give: Segment, Method, and Timing. Segment is the demographic segment that a donor is categorized as being a part of, method is the channel or means by which an individual donor makes gifts, and timing is exactly when they are asked, either during the calendar year or relative to specific institutional events.   

Building a predictive model to assess the optimum conditions for annual giving can help your program to understand a number of things, like which donors are the most likely to give and therefore who should be prioritized by the program, where ask amounts should be set for the maximum contribution, and how each solicitation channel like email, mail, and phone should be utilized. Good places to start when building a predictive model are who is likely to renew, at what levels should people be solicited, and who is likely to upgrade to leadership annual giving.

Let's go through the creation of these three models step-by-step:

Giving/Renewal Likelihood Model

-          Flag donors in your database that have a renewal history.

-          Gather data points such as demographics, activities, interests, and geography related to these donors.

-          Use logistic regression or decision trees to rank donors based on probabilities to renew.

Ask Amount Model

-          Use a linear regression model to predict the largest or most recent gift given.

-          If the resulting “predicted” dollar amount is higher than the most recent gift amount then increase the ask.

-          If the “predicted” amount is lower then leave the ask amount the same.    

Leadership Giving Likelihood Model

-          Derive a binary dependent variable for existing donors at the leadership level.

-          Use a binary prediction method like logistic regression or decision trees to compare independent factors.

-          Output will be predicted probability that can be ranked.

Models like these can be constructed for any number of donor behaviors that might be important to your institution like sustainer/recurring, catastrophe response, preferred giving channel, or cause or interest specific appeals. But of course, all of these models are best used together in order to maximize efficiency and effectiveness.   Results for the programs for which the models are used to boost efficiency should be considered holistically as one strategy and results should be captured not only by appeal but also by demographic segment. Tracking, testing, and reporting mechanisms must be in place prior to launch in order to accurately measure performance improvements.  

Annual Giving Analytics (Part 1)

This is the first entry in a two-part series dedicated to Annual Giving analytics. These posts are derived from a webinar that was given by Bentz Whaley Flessner partner Joshua Birkholz and BWF’s Annual Giving expert Heather Greig. They are available to be viewed on the BWF youtube channel.

This post will explore the relationship that business and data analytics can have with Annual Giving operations and the next post will be more in-depth with regards to the techniques and methodologies that can be used.

Analytics generally fits into the business process of a philanthropic fundraising organization in the following sequence:

- Initial segmentation & prospecting if an organization's constituency. 

- Portfolio optimization for major gift solicitation and performance management of operations.  

- Advanced forecasting and simulations to better understand how much money can be raised.     

Most readers understand segmentation but there are many ways to segment an annual giving constituency, including:

1)      Philanthropic Interests

2)      Engagement and touch points: events attended, publications received, services provided, etc.

3)      Giving History such and length of giving relationship or amount given

4)      Channels by which gifts are made: mail, phone, or online.

  

Analytics generally supports prospect research by deriving a quantitative scoring methodology that measures how close or “warm” an individual is to the organization. Scores can then be used to prioritize which individuals should have a more complete dossier completed. 

Performance enhancement is the category of analytics usage that best describes its relationship with Annual Giving. Annual Giving is usually considered “low gifts at high volume” and the cost to raise each dollar is relatively high. But the true value of Annual Giving is not necessarily in how much money is raised at the lowest possible cost, but instead to cultivate a culture of philanthropy and engagement at an institution. All of these factors, maybe somewhat misaligned, make Annual Giving operations ripe for quantitative analytics. Our next post will explore specific techniques and methodologies for integrating analytics into the Annual Giving business process.     

2012 is OVER! (So let's talk about it)

Happy New Year Everyone!

2012 was an exciting year at DonorCast. We strengthened our team by adding two new positions and expanded our market to include clients in Continental Europe and the United Kingdom. But not everything we did this year was new, we also freshened up our Fundraising Analytics industry-wide survey. Let's go over the results.

Ninety-one individuals responded to the survey, 31 more than the 60 that responded in 2010. This larger sample size gives us a clearer picture of how analytics is being used to improve nonprofit fundraising. Higher Education continues to be the largest category of institutions participating, with research universities the largest sub-category. This probably shows that higher education organizations have the most developed and sophisticated fundraising operations. However, even though research oriented institutions made up 50% of the respondents, both staff sizes and the amount of money that respondents raise annually varied greatly, with the average staff size being about 25 fundraisers raising approximately $45 million a year.

The analytical techniques that are applied to the multiple aspects of fundraising have remained stable with most respondents using descriptive analysis, predictive modeling, and point-based scoring to identify prospects, analyse performance, and segment donors to create more donor-centric appeals. Interestingly, more organizations have stated that they use analytics to assist in the production of reports, which might mean that analytics is becoming more ingrained in the culture of their organizations. This is a good sign and should lead to higher productivity across all aspects of their fundraising operations. Of the software tools used to complete analysis, both the open-source program R and the statistical package Minitab are more widely used now than they were two years ago. SPSS is still the most widely used software program but it's share has fallen by about 20%.

Staffing, and staff compensation, have remained stable since 2010. Ninety-one percent of the respondents have dedicated prospect research and analytics staff with analysts reporting an average annual salary of $45,000 to $50,000 and Directors earning roughly $80,000 per year. Two directors reported earning over $100,000 in annual salary.

While there weren't any "eye poppers" in this year's survey, that might be a good thing. Since the fundraising analytics industry has grown substantially in the past five to ten years, two years of relative calm might show that our institutions have embraced analytics for the long haul and we are fully integrated into our institutions' operations. The increased use of anayltics in reporting and the integration of R and Minitab into our shops shows that people in our industry are continually experimenting with analytics – both its uses and the products used to create powerful analysis. I think this bodes well for future innovation in our space.    

Analytics and Research in Fundraising

Last time, I wrote about how research and analytics functions are viewed in other industries in order to gain some perspective on how the two disciplines fit into a fundraising shop.
In most institution's fundraising development programs there are two categories of activity and responsibilities, front-line fundraising and back-of-the-house operations. Front-line fundraising is pretty self explanatory; it is working with donors face-to-face on a day-to-day basis and is what our prospect and relationship managers do. Sometimes we refer to this work as acting as a liaison between philanthropic donors and an organization's mission. Back-of-the-house operations encompass a lot of activities from technology management, data maintenance, gift processing and…and…and…prospect research and analytics. This goes to show that before front-line fundraisers can talk to donors, or an annual giving director can send out their mailing, quite a bit of work needs to be done to understand an individual donor’s ability and propensity to give. This work is done by analysts and prospect researchers.  

The first step in any prospect development process is analytics. Analytics staff, using advanced data mining and modeling methodologies as well as wealth screenings conducted by outside vendors, identifies an organization’s best potential prospects. These prospects are then placed into refined pools and assigned to the limited number of managers that all institutions have. In addition to identifying potential prospects, analytics staff will also produce technical reports that help management to best understand the performance of the entire development program. These activities are programmatic and on-going. Identifying potential prospects and measuring performance never ends and is an important part of stewarding an ongoing concern like a university or medical center.

Prospect research provides the specific information that is required to best solicit and steward those prospects that are merely identified by analytics. This research uses multiple techniques – surveys distributed amongst constituents, peer screening collected by research staff, and discovery visits – to determine what it takes to turn a prospect into a donor. Unlike an on-going analytics program, these research activities are specific to each prospect at that point in time. Examples of specific prospect research deliverables include donor engagement reports, donor profiles, and targeted research for unique to individuals. Each of these deliverables are individual projects that once completed lead to better insight about already identified prospects. They answer the research question: “Is this identified prospect a potential donor? And if so, under what conditions and strategy.”  

So even though analytics is a relatively new function in fundraising development, it does not replace traditional prospect research. It enhances it, and provides accountability to the entire development program through measurement. Analytics is an aid, a technical tool that improves our philanthropic work by increasing return-on-investment and better targeting very large files of prospects.          

Analytics and Research Functions in Industry

Analytics is a growing function across all sectors and industries of the economy. DonorCast and Bentz Whaley Flessner are committed to harnessing and implementing the best analytics practices for the betterment of philanthropy and charitable giving. As many of our readers already know, we find that using analytics as a part of an organization's fundraising development program greatly increases not only the total amount of money that an organization can raise, but also the efficiency of the resources required to raise those dollars.

As proponents of analytics, we must explain what analytics is and where it fits as the third leg of a fundraising operation stool alongside front-line fundraising and prospect research. This post is the first in a two-part series regarding how we should think about analytics in fundraising and especially its relationship to prospect research.


I think that it will be helpful to look at how many other industries distinguish between "research" and "analytics". In consumer markets like packaged goods, financial services, and retail as well as the medical fields of toxicology, oncology, and epidemiology, research is the process of discovery, and is associated with proving a hypothesis. Research deals with specific cases, and generally, smaller, more manageable amounts of data that can be controlled and isolated in order to identify key relationships between independent and dependent variables. In terms of management, research is more project based, and when conclusions are found, or insights arrived at, then a research project is concluded.

Analytics, also referred to as Business Intelligence, is slightly different and is most accurately thought of as an ongoing management tool. Analytics is a program that constantly collects and measures data that can then be used by management to make decisions about, and adjustments to, strategy. Analytics is much more aligned with an organization's overall business needs – the allocation of resources, return on investment, and definition of success – than research. Many practitioners will say that as a phenomenon becomes better understood, it moves from being the subject of research questions to being a component of management analytics. Instead of understanding what something is or how it works, analytics focuses on the incremental changes that are needed to make a process as efficient and effective as possible. So while many of the tools that research and analytics use can be the same, such as quantitative methods and predictive modeling, their role in an on-going, enterprise system are slightly different.

In conclusion, analytics and research are both extremely valuable to all kinds of organizations but they play different roles and should be thought of as separate, essential functions. The next post, will discuss analytics and research in fundraising specifically and how they work together to accomplish institutional goals.                   

Medalball - Statistics is a Science

Okay, real quick. Before we are all fully recovered from our Olympic hangovers, I thought that it would be helpful for me to bring to every one's attention Nate Silver's article from a few days ago about using data to identify which sports a poor nation should focus their resources on if they want to medal at the Olympics: Medalball.

As many of you may know, Josh and Alex like to use Michael Lewis’s famous book-turned-movie "Moneyball" as an analogy for the work that we do at DonorCast helping nonprofits to identify the ways to get the best fundraising returns given their usually limited financial resources. Nate's article uses the analogy as a way to give direction to small, poor nations that might want winning a medal at the Olympics to give their countries a PR boost. But what I like most about this article is it's structure, and how it sheds light on the logic and thought that is needed to give context and underlying reasoning to the use of statistics and quantitative methods. 
Too often in our careers we encounter clients, decisions makers, or individuals that just want to know "what the data says" or to be "given the stats". They will then use this measured, quantitative information to make a decision. But really, it's not that easy, and actually such an approach can be very dangerous. 

What many people often forget, especially non-quantitative professionals, is that statistics is a science, the science of uncertainty, variability, and decision making. The scientific method requires the testing of a hypothesis, or proposed explanation of a phenomena. This means the stats don't tell a story by themselves, stats are only tools that we use to paint a clearer picture of phenomenon that we already believe that we have a good idea of how it looks. Statistics and data are properly placed at the end of a decision making process, not at the beginning.

Colin Mallows, the one time President of the American Statistical Association, once stated that "statisticians should give more attention to the questions that arise at the beginning of a problem or an issue:

1. Consider what data are relevant to the problem,
2. Consider how relevant data can be obtained,
3. Explain the basis of all assumptions, 
4. Lay out all sides of an argument,
5. Formulate questions that can be addressed by statistical methods."  

This is what Nate Silver does in Medalball. He begins with considering what data is relevant to solving the problem of "which sports a small country with limited resources should direct their energy to in order to maximize their medal count". He already had a string of logic that he would wanted to follow in order to solve the problem that he then used statistical methods, in this case descriptive analysis, to verify and test his hypothesis. 

Statistics and data are not replacements for thought and contemplation, instead they are merely powerful aids for better understanding our already formed ideas and presumptions about the world.

Morgan Zehner Introduction

My name is Morgan Zehner and I am the newest member of the DonorCast Team. I want to introduce myself to the awesome community that we serve and I thought that a blog entry would be a good place to start. I will be the primary analyst for all of DonorCast’s program analysis projects. I am really looking forward to providing continued innovation to the DonorCast forecasting methodology and designing nonprofit sector metrics to better understand development productivity and performance. And, like Josh did, I am going to write a book. 

So, a little about ME (visit my web page and watch my bio-video). I began my career in nonprofit leadership as the Executive Director of Dupont Circle Main Streets. Dupont Circle is an internationally recognized commercial district in Washington, DC. I first joined the nonprofit world because I was passionate about a specific mission – urban revitalization – but I quickly realized that I had a knack for fundraising. While in Dupont, I was able to quadruple the organization’s budget by executing a portfolio of fundraising methods – from annual giving and grant writing to special events and securing a $90K major gift. After leading the organization through this growth period, I started my own consulting business where I provided small businesses and nonprofit organizations with counsel in the areas of market research, strategic planning, and fundraising development.

While the first part of my career has been incredible and full of accomplishments, I have always wanted to do technical, analytical work. As an undergraduate student at McGill University, I was drawn to quantitative social research and wanted to use those skills, or as Alex says “flex those muscles”, in my career. With that as my calling, I attended the full-time MBA program at the Carlson School of Management at the University of Minnesota where I focused my studies on market research and strategic management. It is this combination of experience and education that has made Bentz Whaley Flessner, and DonorCast in particular, a perfect fit. BWF is an entrepreneurial and innovative small consulting shop where I use my technical skills to serve an industry I respect. Nothing could be better and I am greatly looking forward to working with all of you.  

If you have any questions or would like to introduce yourself to me (please do!) drop me a line or follow me on Twitter

mzehner@bwf.com

@morganmz

Thanks! and looking forward to working with you,

Morgan 

Predictive Modeling for Direct Mail at Children's Mercy Kansas City

The direct mail program at Children’s Mercy Hospitals and Clinics in Kansas City has been raising more money and is costing the hospital less to implement. Why? Because David Logan‘s annual fund team has been paying attention to the numbers and using predictive modeling to better target the mail they send out. Check out the video below for a more complete explanation of David and team’s work…