How the Not-for-Profit Sector can make the most of Machine Learning
By Ian Woodgate, Head of Presales
At Pythagoras, we’re constantly striving to bring new, exciting, but most importantly value adding technologies to our customers. Machine learning is a game changing technology and in this post I’m going to share some ideas about how it can be used in the Not-for-Profit sector.
Fundamentally, machine learning is about using advanced techniques based on artificial intelligence to extract useful information from data. We take a sample of data and use that to train our machine learning algorithms, and then when it has learnt from the training data, we can use our algorithms to answer questions.
However, there are a few caveats. We have to have enough data, and it has to be the right kind of data. Furthermore, we have to ask the right kind of questions if we are to get meaningful and useful results. Finally, we have to make sure we use the right algorithms in the right way. On any engagement these areas are where Pythagoras’ expertise comes in.
So how can machine learning be useful for Not-for-Profit organisations? If you have a reasonable sized database containing information on donors, then here are some of the questions you could consider asking:
Which of my donors has the highest risk of cancelling their regular donation?
We can answer this question by looking at things like donor profile and engagement history. Needless to say, if you can identify donors at risk of cancelling and take preventative action then the impact could be considerable. What if you could reduce the cancellation rate by 10%, 20%, or more?
What medium should I use to approach a given potential donor?
Based on the profile of the potential donor, we may be able to determine which of email, post, call or SMS has the highest chance of success. Drilling down still further we may even be able to identify the time of day, the day of the week, or the day of the month to establish contact.
How much should I request from a potential donor?
Always a difficult decision – ask for too much and risk putting them off, but ask for too little and you could be missing out. Using donor profile information and machine learning, we can suggest recommended amounts to request on an individual basis.
Which of my donors are donating an inappropriate amount? - either more or less than they can afford?
If these can be identified, then remedial action can be taken proactively.
Of course, the above is just to highlight some of the possibilities. Machine learning is really a solution looking for problems, so if you think it might be of relevance to your organisation, and would like to find out more, then please do contact us or come along to our NfP event in London on 3rd November 2016 and see machine learning in action with NFP data.
http://www.pythagoras.co.uk/wp-content/uploads/2017/07/User-Experience-Enhancements-for-Customer-Engagement-in-Microsoft-Dynamics-365.jpg450800Claire Pearcehttp://www.pythagoras.co.uk/wp-content/uploads/2015/07/Pythagoras-Logo-RGB-Blue-Horizontal.svgClaire Pearce2017-07-31 10:33:152017-07-31 11:31:50User Experience Enhancements for Customer Engagement in Microsoft Dynamics 365