Predictive Modeling uses “big data” to help predict the future. It’s the data-driven, more reliable version of a crystal ball. Last year, I worked with 2017 npExpert Katrina VanHuss at TurnKey P2P to build a model for walking events with no registration fee to give visibility into the future success of your walk based on the event’s current progress to date.
This predictive model can help you with understanding the real-time health of your event to ensure that you will hit your goals by event day. It also gives you earlier visibility into your numbers, so you can take corrective action and set executive expectations.
For traditional events like walks, event participation and total funds raised follow an exponential growth curve when tracked over time. This growth pattern is relatively similar event to event, regardless of size, which makes it possible for you to use the model to track the real-time health of your event based on the number of days out from your event, number of people registered and total fundraising (excluding sponsorships.)
Predictive Modeling for Fundraising Events
Predictive Model for Funds Raised & # of Participants based on # of days out from the event
|Days until the Event||% of participants||% of funds raised|
Additionally, you can use this same predictive model to forecast how much money you’ll have in the bank AND how many people will show up when event day gets here.
Here’s the equation:
- Total # of participants = today’s participant count / % of participants based on # of days out
- Total expected fundraising = today’s fundraising / % from chart based on # of days out
Example of the predictive model for a peer to peer event:
I’m organizing a walk that takes place in the next 30 days. Today there are 2,000 people registered and $100,000 in fundraising in the bank. On event day, I want to have 5,000 people at the event and fundraising a total of $250,000.
Based on the chart above at the 30 days out mark, I should have 50% of my total people registered and 30% of my fundraising in the bank.
- Total # of participants: 2,000 / 50% = 4,000 participants
- Total expected fundraising: $250,000 / 30% = $300,000 in fundraising
This means that I can expect that by the time event day gets here, I may fall short on my participation goals, but I can expect to have extra fundraising in the bank. I should focus on my recruitment strategies. Having this visibility can help you decide where you want to prioritize your limited resources.
This probably goes without saying, but predictive models are not an exact science. We used Blackbaud data to pull together this model specifically for walks without a registration fee and it does help set a good stage for comparison, especially for new events without historical data.
The most accurate predictor of your event is your own historical event data.
I’d encourage you to take several years of data and plot them out week of week leading up to event day. If you are interested in working with one of the Blackbaud’s P2P experts to assist you with building customized predictive models for your own program, please contact your customer success manager or sales person.