The Hot Seat is a series in which prominent industry experts answer grueling questions stemming from prospect research to consulting to analytics. How will they do under pressure? Read to find out!

David Lawson is a leader in bringing actionable insights to the fundraising community. He is the CEO and Co-founder of NewSci, LLC and author of the notable book, Big Good: Philanthropy in the Age of Big Data & Cognitive Computing. Additional career highlights include but are not limited to - In 1997, he founded Prospect Information Network (P!N), which became the largest wealth screening company before being purchased in 2004. In 2014, he was among the early application developers approved to use IBM Watson. And, David is the recipient of the Apra Distinguished Service Award and the CASE Crystal Apple Award for Teaching Excellence. If you have any questions for David, you can reach him at David@NewSci.ai

1. True or false: Data is a renewable resource – vulnerable and valuable. Please explain why with your answer. 
   
   The blessing and the curse of data is it never stops being created, so while some call data the new oil, it is more like the sun, constantly generating new data and combining data to create even more for us to consume and make sense of. So, data is renewable in the truest sense of the word.
   
   As to vulnerability, data presents special challenges because it can so easily be transferred between entities and individuals. That ease puts a premium on an organization's data governance capabilities. Too often data is merely stored instead of managed and that leads to data breaches and data leaks. Data is also vulnerable to becoming out-of-date or irrelevant. Two big vulnerabilities are the quality of the data and missing data. Ensuring data quality is a full-time job, and identifying missing data, including why it is missing, is too often overlooked and leads to erroneous information and insights.
   
   The value of data is closely linked to the ability or inability for it to be turned into actionable insights. Without the insight, it is just a collection of facts and data points, and without actions, the insights are rarely turned into something of high value.
   
   My experience over the years with wealth screening has been that the organizations who combine the data obtained from a screening vendor with internal information in order to go beyond just wealth indicators, and then integrate the insights into their fundraising processes, are the ones who realize the highest return on their investment. 
   
   Truly appreciating the value of data requires an organization to see data acquisition and analysis not as a cost-center, but rather as a profit-center. To make your case look at the value of the insights you provide such as what is the mean major gift from newly discovered prospects. Multiply that by the number of new prospects and you have a potential Return on Insight formula, and you can of course use the actuals to show the value already realized. For lower level donors, look at direct mail and marketing. You need a correct address in order to potentially receive a gift, so it is fair to highlight the amount of donations received through the mail from people who had a corrected address (or a correct email) in the last year. If you really want to show executives how valuable keeping track of people is, calculate the life-time value of your donors and then use that as you multiplier.
   
   Machine learning, natural language processing, and deep learning, known collectively as AI, are making data exponentially more valuable. At the same time, this technology has the potential to dramatically increase the vulnerability of data as it requires the aggregation of more and more sources to create the algorithms used to glean insights. Monitoring the quality of all the data going into an AI-powered platform requires new levels of data governance especially around data provenance to ensure you know from where data came. Privacy regulations such as GDPR are making it mandatory for organizations to be able to explain how their algorithms work which means you have to know what data was used, how it was used, and its origin. We can expect this type of data regulation to become more prevalent as algorithms are more deeply integrated into our everyday lives making the consequences of data bias and misuse even more consequential
   
   
2. What is our argument for the ethical use of data analytics in the nonprofit sector?

   
   Being data-driven has become a cliche, but it became one because it is critical to an organization’s success. For a nonprofit organization this is even more true because they do not have resources to waste being data-blind. Before analytics, organizations ran on intuition and gut instincts. In a small nonprofit you might get away with this approach much as a startup company relies heavily on the passion and knowledge of its founders. Once an organization matures it is frankly unethical to keep operating without a deeper understanding of what is happening within the organization and to use that understanding to make better decisions.
   
   

   This does not, however, give a nonprofit free reign to do whatever data collecting and analysis they want to do. Just because one is doing good with data does not give one a pass on not doing that good as ethically as possible. Keep in mind data ethics isn’t just about you and your organization. It is also about the sources of data you acquire and the companies you use to manage your data. They too must adhere to the same standards you set for your organization. 
   
   

   Data ethics begins with establishing clear values and principles for how your organization sources, collects, stores, analyzes, transmits, and uses data. With these in hand you can build an ethics-first data governance program to support your organization’s actionable insights needs.
   
   
3. What are some strategies for effective data collection? 
   
   

   Start with cataloging all of your current data. This is a daunting task, but easier now that there are platforms specifically designed to help you create one. The data catalog needs to be across your entire organization, not just in a departmental silo. You will be amazed, and at times shocked, by where data resides and who has access to it.
   
   

   With the data catalog as the foundation, you can begin a deeper analysis of the quality of your data as well as identifying missing data. This is never a fun exercise as you will surely find serious quality issues and critical data completely absent. Rather than fear what you will find, recognize the first step in data governance is accepting the inherent imperfections of data. What is important is to then establish processes to increase the quality of your data and minimize instances of missing data.
   
   

   With the catalog and data analysis completed, you can focus on data provenance, also known as data lineage. This provides a clear picture of the sources of data, enabling you to accurately trace issues and errors back to their source and, if a breach occurs, where it happened and what data was impacted. Whether it is an internal source such as your gift processing team or an external data provider, knowing where, when, and who your data came from can no longer be ignored if your goal is to have an ethical, effective, and efficient data collection operation.

   
   Given the speed at which data is created, collected, and analyzed, it is also worth considering an anomaly detection system. How you monitor for anomalies will depend on the frequency and importance of the data being monitored. You want to find a balance between being alerted to anomalies in time to take action, and not having so many alerts you start to ignore them. To help make the case for anomaly detection, take a data point such as the amount of a donation and trace it through your reports, dashboards, analytics, and predictive models dependent on the data to show, for example, the negative impact of a $1,000 gift being entered as $100,000 or meetings with campaign prospects not being recorded. Anomaly detection was something we did a lot of in my wealth screening company, Prospect Information Network (P!N). What we found was data providers during an update would not always send everything we were expecting. Because of the large volume of data, it would have been easy to miss, for instance, one county's real estate records, and only through anomaly detection were we able to find missing data and fix it!