The Future of Fundraising: Optimizing Impact with Large Language Models

read time - icon

0 min read

May 30, 2024

It’s not news that peer-to-peer fundraising is a huge opportunity for charities. We’re more likely to respond (and be more generous) when a friend asks us to donate as opposed to a random stranger. Plus, peer endorsements promote trust in a complex system of requests for our money and attention. 

It’s also not news that asking friends, family, and acquaintances to donate is hard. While charities try to reduce the burden on fundraisers by providing copy-paste templates, these don’t necessarily promote effective fundraising.

A selection of Movember’s fundraising resources for fundraisers’ social media campaigns

The trade-off here is that making fundraising easy is important to recruit fundraisers, but making fundraising look hard is important for actually raising the funds. Not only are we more willing to donate to someone that we see doing something difficult1 (think ALS’s viral ice bucket challenge), but we are also more likely to donate to someone that we see making a sustained effort with their fundraising such as regular and thoughtful posts.2

Thus, we face the age-old dilemma of harnessing the power of peer-to-peer fundraising: how can we make fundraising easy while making it look hard? The good news is, we do have a new tool available. Enter: Large Language Models (LLMs). Trained on immense pools of text data, these models understand, analyze, and generate human-like text and are the magic behind widely used artificial intelligence applications like ChatGPT.

Building off of what historically works for peer-to-peer fundraising, today we’ll break down the ways in which LLMs are an indispensable tool for charities seeking to tap into their golden network.

Transforming a supporter into a fundraiser

The first step to any peer-to-peer fundraising is, of course, recruiting the people to do it. 

The literature offers a number of barriers to this step. 

  • We tend to underestimate others’ willingness to help and therefore underestimate the potential impact of any fundraising we may do.3
  • While raising money is visible, we don’t see, and therefore don’t mentally account for, the impact that fundraising has beyond dollars raised. Raising awareness for a cause, or simply making caring about others a visible social norm, are worthy outcomes in their own right.
  • We worry about the reputational impacts that sharing our good deeds may have, lest our well-intentioned posts be seen as ‘self-righteous’ or ‘braggy.’4 According to Tan, Yan, and Pedraza-Martinez (2020), this phenomenon, known as Braggart’s Dilemma, “has prohibited social media from reaching its full potential in promoting valuable causes, spreading worthy volunteering opportunities, and most importantly, communicating the norm of helping others.” 5

Thus, we tend to overestimate the costs and underestimate the benefits when it comes to peer-to-peer fundraising. Fortunately, researchers have identified that we can help to recalibrate potential fundraisers and promote their willingness to share by drawing attention to the social benefits of fundraising, such as raising awareness and starting a chain reaction. This is an effect that we at The Decision Lab have seen in our own work with charitable organizations. 

LLMs can help to level up this effect with personalization. Using information that charities already have, LLMs can create messages that target areas of the charity’s work that are particularly close to the fundraiser’s heart and highlight benefits that are adapted for their demographic profile. 

Making the potential impact of fundraising more personal and more salient is not only important for recruiting fundraisers, but it is also likely to have benefits when it comes to the recruited fundraiser’s effectiveness by strengthening their connection to the cause.

Helping fundraisers do better by signaling more commitment

A study by Chapman, Masser & Louis (2018), delving into the factors that predict successful peer-to-peer fundraising, found that it is not the cause itself, but the actions of the individual fundraiser that matter most.2

They found that making more of an effort with fundraising (for example, by personalizing their fundraising page, sharing their reasons for supporting the charity, or posting more regularly) signaled more investment in the outcome of the fundraising and were 8 to 20 times more predictive of fundraising success than the efficacy of the charity.

The unprecedented level of personalization offered by LLMs offers huge advantages when it comes to leveraging these success factors. Let’s explore two below.

Showing commitment through personalized content

LLMs can help fundraisers personalize their fundraising pages and social media posts with unique, individualized content that speaks to their particular reasons for supporting the charity. Since the effort for fundraisers lies in crafting messages, passing off some of the work to LLMs to create personalized content allows charities to strike a balance between effective fundraising and the ease of generic templates.

For instance, imagine a chatbot that takes answers to a few questions and provides background information on the fundraiser, transforming it into a unique communication plan for a fundraiser’s campaign. With access to previous social media posts, LLMs can even adopt a fundraiser’s unique tone to create something that feels more natural to the fundraiser.

Given a brief about the fundraiser’s age, gender, personality, length of campaign, and posting frequency, GPT-3 generated these social media posts for a Breast Cancer Foundation fundraiser

Consistent outreach: the key to fundraising success

In the fast-paced world of social media, sharing regularly shows more commitment than a one-off post. Peer-to-peer fundraising campaigns often happen via individual fundraising pages (such as the Terry Fox Run). LLMs can harvest information from these pages, such as page views, donation amounts, and even supporters’ messages to create personalized encouragement to fundraisers. This consistency can help remind fundraisers of the impact they are having (and the positive reinforcement they have received from real members of their networks).

An example of an individual fundraising page for the Terry Fox Foundation

From challenge to opportunity

Through personalization and conversation, LLMs can eliminate the effort-outcome trade-off, making fundraising easier and even more impactful. 

LLMs can help turn supporters into fundraisers by personalizing recruitment efforts, enhancing fundraisers’ connection to the cause, and helping them maximize their effectiveness. In this way, LLMs offer charities an opportunity to raise more money and offer a better experience to their fundraisers whose greater impact provides a greater sense of satisfaction and achievement.

Through TDL’s work with organizations operating in a peer-to-peer fundraising environment, we’ve come to a deeper understanding of the drivers and barriers facing fundraisers, the features of a successful fundraising platform, and the challenges facing charitable organizations. If you’re looking to supercharge fundraising for your organization, whether it’s through the power of LLMs or otherwise, get in touch with us today.

References

  1. Olivola, C. Y., & Shafir, E. (2018). Blood, sweat, and cheers: The martyrdom effect increases willingness to sponsor others’ painful and effortful prosocial acts. Available at SSRN 3101447. Retrieved from: https://papers.ssrn.com/sol3/papers.cfm?abstract_id=3101447 
  2. Chapman, C. M., Masser, B. M., & Louis, W. R. (2018). The Champion Effect in Peer-to-Peer Giving: Successful Campaigns Highlight Fundraisers More Than Causes. Nonprofit and Voluntary Sector Quarterly, 089976401880519. doi:10.1177/0899764018805196  Retrieved from:  https://journals.sagepub.com/doi/abs/10.1177/0899764018805196 
  3. Flynn, F. J., & Lake, V. K. (2008). If you need help, just ask: Underestimating compliance with direct requests for help. Journal of personality and social psychology, 95(1), 128. (2008). If you need help, just ask: Underestimating compliance with direct requests for help. Journal of personality and social psychology, 95(1), 128. Retrieved from:  https://pubmed.ncbi.nlm.nih.gov/18605856/ 
  4. Silver, I., & Small, D. A. (2023). Put your mouth where your money is: A field experiment encouraging donors to share about charity. Marketing Science. Retrieved from: https://www.researchgate.net/publication/373001447_Put_Your_Mouth_Where_Your_Money_Is_A_Field_Experiment_Encouraging_Donors_to_Share_About_Charity 
  5. Tan, J., Yan, L., & Pedraza-Martinez, A. (2020). How to share prosocial behavior without being considered a braggart? In Proceedings of the 53rd Hawaii International Conference on System Sciences. https://doi.org/10.24251/HICSS.2020.482 Retrieved from: https://scholarspace.manoa.hawaii.edu/items/c60475e6-c6ff-4b64-b3a3-7dfb2ecc90ed 

About the Author

Caitlin Spence

Caitlin Spence is a Senior Associate at The Decision Lab. Before joining The Decision Lab she worked in Aotearoa New Zealand’s justice sector as part of a team using behavioural science to create more accessible and culturally aware systems. Caitlin is interested in using data and experimental design to understand how systems can be designed or changed to favour positive and equitable outcomes. She holds a Bachelor of Arts, majoring in Statistics, from the University of Auckland.

Read Next

blue photo of romans
Insight

Taking a Hard Look at Democracy

Tom Spiegler, Co-Founder and Managing Director at The Decision Lab, joins Nathan Collett to talk about what behavioral science can tell us about the 2020 US election and the state of democracy more generally.

Notes illustration

Eager to learn about how behavioral science can help your organization?