From data point to donor interaction: An evolution of not-for-profit artificial intelligence tools
3 September 2020 at 7:00 am
Dr Annabelle Chuinard and her team at Keela have created accurate, predictive tools to help not for profits streamline their work and enhance their ability to attract, cultivate, and retain donors in the long run.
What do elementary particles and donor interactions have in common?
A lot, according to PhD physicist Dr Annabelle Chuinard.
Dr Chuinard spent her doctorate working at a particle accelerator in Switzerland, analysing the interactions between particles. As she moved towards a career in the field of generosity, she noticed a striking similarity between how those minute entities of matter and energy interact with each other and how not-for-profit organisations interact with their donors.
This brainwave led Dr Chuinard to her current line of work at Keela: building sector-specific artificial intelligence tools to help fundraisers raise more money and do more good.
Generally, AI is any technology (think software) that can demonstrate some of the behaviours we consider to be associated with human intelligence.
Whether that be problem-solving, planning, or predicting, this software – aka the brains of a computer – are taught how to “learn” to do this better and better the more they “practice”. Simply put the more data the machines process, the better they get at their specific task.
This technology is finally filtering down to our sector and into the capable hands of data scientists like Dr Chuinard – and at such a crucial moment in history.
In not-for-profit applications, you can use AI-powered tools to sift through donor demographics, previous donation amounts, event attendance records, and volunteer hours in order to get a true understanding of your supporters’ giving history, interests, and impact
For example, if you want to predict when an existing donor is ready to give again, you would feed your currently available data through an algorithm built specifically for that purpose. The computer then analyses giving patterns within this dataset and holds onto the information for future use. Data scientists call this “model building”. The more data fed into it over time, the more accurate it becomes. That model can be drawn upon again and again to analyse and predict when a newly-added contact will be likely to give next.
These tools enhance an organisation’s ability to attract, cultivate, and retain donors in the long run.
If you believe such systems could help your own organisation, you would be in the majority as 89 per cent of not for profits believe AI has the potential to make their organisations faster, smarter, and more efficient. Strikingly, just 23 per cent have embraced these types of solutions.
For many, AI carries the stigma of being a complicated, futuristic concept employed only by large, for-profit tech companies with exhorbitant budgets. Dr Chuinard and her colleagues are working to reframe AI as the helping hand to administration and workflow efficiencies that not for profits of any size or cause area can call upon at any time.
Having all hands and every resource available is especially important as we continue to adjust to the social and economic fallout of COVID-19.
There are 1.3 million people currently employed in the sector. And unless we find a way to switch comprehensively and effectively to online fundraising, Australian charities could face 200,000 job losses and lose up to 20 per cent of revenue according to a new report released by the Centre for Social Impact.
With our workforces holed up inside, fundraisers are spending less time in the field, making it not just convenient, but necessary, to adopt donor-centric AI tools.
For not for profits who survive on their ability to recruit donors and nurture both new and existing relationships, these types of tools can provide an operational edge that can withstand job loss, gala cancellations, and an increased administrative burden to capacity-strapped staffers.
“You’re not going to go over thousands of emails to see whether or not a donor is engaged. You’d want to have a way of accessing that information quickly and be able to respond in the most effective way,” Dr Chuinard said.
AI in the not-for-profit sector generally boils down into two main categories:
The first is creating donor profiles by organising them into clusters based on similarities. This is used to identify what kind of audience you’re talking to and tailoring your communications.
The second is by providing predictions such as when a donor is likely to give, how much to ask for, and whether they’re likely to become a major donor in the future.
Not for profit data is a rich source of information on the physics of giving that helps build accurate, predictive tools to streamline your work.
Dr Chuinard and the intelligence team at Keela have built several tools to do this:
Smart Ask: The science behind asking for a donation is evolving. We know that presenting the right amount for a specific donor increases the likelihood for this person to give and reduces the chance to underestimate their ability to give.
It was also demonstrated that if you have an array of donation suggestions and you highlight the middle option, your revenue will increase. It is not only about having choice but more so about priming the “most common” amount within an array of choices that drives people to give.
With this logic in mind, the team built a Smart Ask feature to automatically provide a recommended donation amount based on previous giving history as well as suggestions of donation amounts to be used in targeted campaigns.
Donor Readiness: Donor Readiness tells you how likely your donor is to donate within the next two weeks. It is calculated through a machine learning algorithm that analyses contact interactions, giving history, and even variables like contact location and the weather. After all, individuals are more likely to give on sunny days.
There is no common ground when it comes to defining when is the best time to reach out to all donors. Still, good timing greatly maximises the chances of giving. Combine Donation Readiness with Smart Ask, and fundraisers will know how much to ask for and when.
Campaign Recommendations: Dr Chuinard’s team analysed which donors are likely to give to an organisation’s campaigns. They also looked at the similarities between campaigns so when a new campaign is created this feature can suggest a list of likely donors for fundraisers to reach out to.
See how artificial intelligence can change the way you fundraise. Test out the AI-driven fundraising tools Dr Chuinard is working on free for 15 days.