Machine Learning and P2P Fundraising
- September 14, 2020
- 3 min read
BetterNow has initiated a major project to develop a machine learning algorithm that will serve personalised recommendations to peer-to-peer fundraisers.
Since February of this year, BetterNow has been investigating the potential for applying machine learning to P2P fundraising, and after some initial data analysis, we found that there was great potential in building a recommendation engine for fundraisers.
The engine will help each individual fundraiser achieve the best possible results, by highlighting the actions they can take that have the highest potential.
One of our motivations for starting the project was the realisation that we probably had one of the best datasets there is on P2P fundraising. Not only is the number of observations very high, the dataset is incredibly rich.
In June we were awarded a grant from Innovation Fund Denmark for developing and implementing this project, and we are super excited to make this project a reality.
BetterNow has partnered up with the enterprise AI company 2021.ai who will help us with advanced data analysis and developing the recommendation engine. This work is well underway and the initial results and analysis are very promising.
The next step after this will be to implement this engine, both technically and practically, and we hope to start testing with some of our clients starting in the spring of 2021.
How will machine learning help charities?
The machine learning algorithm will be implemented in two key ways. The first is through a new improved fundraiser creation flow, where fundraisers will receive better guidance on setting up their fundraiser.
Secondly, the machine learning algorithm will generate actions for each fundraiser based on what have worked for previous fundraisers. The fundraisers will be presented throughout their journey with the actions that have the biggest potential impact for them.
What have you found from your data analysis?
As soon as we have finished the initial data model, we will start sharing some of the results from the data analysis here on this blog.
Some of the areas we look forward to sharing our results on are
- How much does a fundraiser need to donate themselves before it has an effect on how much others donate?
- What image types make the best profile image?
- What should a fundraiser write in their fundraiser description?
- How high should the fundraiser set their goal, and when should they raise it?
- How have our current email journeys impacted the results?
- How does all of the above vary across causes and fundraiser types?
- What impact does it have when fundraisers write thank you messages?
We look forward to sharing these results.
Do you want to be part of the project?
Are you an existing BetterNow customer and want to be part of testing this? Send us an email and we will get back to you with more information.
Not a customer yet, but think this sounds interesting? Don’t hesitate to contact us.
Posted by Jesper Juul Jensen in July 24, 2017
Posted by Jesper Juul Jensen in June 11, 2020