A case study from the Canadian Red Cross partnership with Walmart
I was meeting with a friend who works for a Hospital Foundation recently, and in conversation she asked about our database and our analytics capabilities. A vendor approaching her had made quite a show of demonstrating what was described as their ‘analytics’ capabilities.
That got us talking. One thing that quickly emerged is that there is some confusion in the sector over terminology and what people mean by ‘analytics’. Traditionally, analytics was dominated by mathematicians, statisticians, or computer programmers. While still retaining its roots, marketers and other business owners are now much more involved. This is due to operational imperatives, the explosion in the amount of data available and its accessibility due to better visualization techniques.
Reorganizing how we use analytics
A couple of years ago, we reorganized the Marketing & Philanthropy team at the Canadian Red Cross into marketing ‘channels’ and business support ‘hubs’. The idea was to move from a generalist “Jack of all Trades” model that duplicated effort and missed out on the benefits of functional alignment, to one where we’ve created centres of excellence where specialization and depth of expertise is valued and developed.
One of the marketing hubs created was the Analytics Hub – initially three self-identified ‘data nerds’. The challenge: come up with answers to questions we didn’t know we had yet. With 1.2 Million active donor records and an organizational commitment to invest in our IT systems architecture, they could barely contain their glee…
A logical place to start was to clarify what kind of analytics work we were already doing. A useful and common framework that we have been working with is to break the area into four broad avenues of enquiry:
- Descriptive analytics
- Diagnostic analytics
- Predictive analytics
- Prescriptive analytics
Descriptive analytics is something that many charity marketers have been doing for years. In essence, one is looking backwards, analyzing an individual, campaign or channel performance against a number of predefined metrics. Dashboards can be built to display results. Benchmarks can be established. Business decisions are supported and success or failure evaluated.
A practical example is our annual campaign with Walmart Canada. Walmart Canada and their customers have been incredibly generous supporters of the Canadian Red Cross for many years. Together, we have raised more than $35 Million to support families here in Canada be prepared for and recover from disasters. These are often the types of disasters that you don’t see on the news – the ‘personal disasters’ that occur because of a house-fire, for example. Walmart associates prompt their customers to give to the Canadian Red Cross and the company matches gifts as well.
Using descriptive analytics has allowed us to look at the daily trends, store performance and even individual management performance in previous years to establish benchmarks for this year’s campaign. Results were broken down to a daily target to allow us to track comparisons against the previous year.
Walmart 2017 Daily Update
What about Diagnostic analytics? One variable we started to notice over the years, and wanted to enquire into, was individual store performance referenced against an individual store manager’s history. If we could identify the best performing store managers, then we could concentrate support on them and see what the impact was.
So we dug into the top twenty performing store managers (the managers who consistently caused an upswing in donations irrespective of the store they were managing). Armed with this information from the Analytics Hub, the Community Engagement team were able to reach out through our partner to those individuals, and offer a level of support and encouragement to those store managers. This ensured their commitment to the campaign and also enlisted them as internal champions whose behaviour could be modelled.
As the name would suggest, diagnostic analytics takes place when there is a focussed and in-depth enquiry following a particular line of questioning, digging behind the surface numbers to uncover a deeper pattern. Why has a campaign has underperformed? Why are certain audience segments performing better than others? How can we test the parameters of a question to determine the key variables?
The next level
Then things started to get really fun. With so many people and so many stores involved in the campaign, optimizing the day-to-day management of the campaign becomes a key question. Historical performance is only one dimension of determining a store’s likely performance.
So the team developed an algorithm taking multiple success determinants into account, and created a predicted income level for each store. This we dubbed the store ‘pace’, and combined with our daily tracking, it was an invaluable management tool for both ourselves and Walmart, allowing us to see which stores needed attention on a daily basis.
This is a simple yet effective example of Predictive analytics. It is an area that we are actively exploring, playing catch up with the commercial marketing world. This form of analytics becomes particularly important with regard to operational planning. In times of disaster, our operational colleagues need to juggle individual and community needs with the amount of resources they have at their disposal. The likely final amount of funds donated by the public is a critical piece of information for the disaster teams in those crucial first few days of planning.
To meet the needs of our operational colleagues, we put together a different algorithm, based on various factors – past fundraising appeal performance, media exposure, pattern of early online giving, etc. – to create what we affectionately call the ‘crystal ball’.
In order to get the most use out of this particular tool, it has taken time to train colleagues in the dark arts of probability ranges. Operations want certainty – but no disaster appeal fundraising campaign can guarantee that. So instead we have communicated probability ranges – a low and high end of predicted income – that allows at least some degree of certainty for operational planning. As more data is fed into the algorithm, it tightens up and within a week, all other things being equal, we can predict with relative accuracy where a two to three month public fundraising effort will land.
Probability Forecast of Appeal Total
The journey from here
Prescriptive analytics is a level of analysis that we are currently working on at the Canadian Red Cross. In this case, the learnings from the other levels are combined to ‘dictate’ the approach we take with a donor: the number, type and frequency of touch points, etc. Rather than charity marketing being a case of throwing out the widest net and seeing who one ‘catches’, it is moving us towards a highly tailored approach. Historically, this was not possible due to the nature of the data charities held: mostly transactional and very little about lifestyle, socio-demographics, or other attributes and preferences. The advent of Big Data and the explosion in the amount of publicly available data is changing this.
Our current work involves finding and developing donor segments with clear patterns of behaviour (or journeys) that take in not just transactional data, but also other data points available to us. We are now developing and testing a choice architecture that assigns probability scores to each choice point in the donor’s journey, so as to better predict what particular channel or type of communication they will value most. The prize: bringing greater intimacy and precision to how we relate to our supporters, which will deepen trust and confidence in the work, and ultimately secure their support for the future.
About the Author
Ronan Ryan is Chief Marketing and Development Officer of the Canadian Red Cross. He is a fundraising and marketing professional with 17 years in the nonprofit sector and has served as a C-level executive in a number of organizations. With a responsibility to lead fundraising efforts, Ronan has developed a new Philanthropy structure to align staff into various centers of fundraising excellence. He currently serves on the Board of Imagine Canada.
Guest contributions represent the personal opinions and insights of the authors and may not reflect the views or opinions of Imagine Canada.