AI and its subsets machine learning and natural language processing help fundraising in a number of ways. Here are some examples, from targeted audio ads to auditing campaign assets, and improving targeting.
Targeted audio ads
Say It Now uses Natural Language Processing (NLP) to develop technology that lets computers talk to people in the same way that people talk with each other, with a specific focus on voice transactions. Last year saw Global, Europe’s largest radio company, partner with Say It Now on an Innovate UK funded research project to investigate the efficacy of donations to charity facilitated by voice assistants on smart speakers.
The research project ran advertising campaigns between September 2020 and April 2021 with NSPCC, WWF and Comic Relief. Targeted audio ads ran specifically on devices that support a voice assistant, with listeners given the option of finding out more about the individual charity or donating straight away with a simple verbal instruction, providing the voice assistant is directly linked to a payment service.
Say It Now believes this is the largest dataset ever gathered using targeted radio adverts to drive to a transactional outcome on voice assistants.
Charlie Cadbury, CEO, Say It Now, says:
“We learned that although this is a nascent channel to reach an audience there is a growing number of people who are happy to make a donation using simple voice commands. There are several new charity campaigns being run this year and Say It Now and Global will be releasing a joint White Paper in June, which will go into the learnings and benchmarks we learned in this initial phase.
“Actionable Audio Advertising is growing at an incredible rate due to the increased adoption of voice assistants and the immediate insight and metrics that are able to be delivered from these kinds of campaigns.”
Voice-based giving solutions
WPNC Digital created a voice-based giving solution for the NSPCC. In association with the charity and Amazon, the agency developed an Alexa skill that helps parents explain to their children how to stay safe from sexual abuse. The campaign features songs and games to make the experience informal while sharing serious information with youngsters.
Users are also able to make a donation by asking Alexa using Amazon Pay. This is also the genesis of WPNC’s new platform, goDonate Voice, which extends options for donors to give via the voice channel.
Vicky Reeves, Managing Director Digital and Deputy CEO, WPNC Digital says:
“We absolutely believe voice AI will be a mainstream part of our daily lives as the technology and user experience evolves. Charities should consider this as part of their digital innovation strategy, if they aren’t already.”
AI-auditing brand assets for campaigns
NCS’s (National Citizen Service) yearly campaign focuses on raising awareness and signing up teens to participate in programmes that build their confidence and unleash their potential.
Automated Creative optimises NCS ads on social and display platforms, running hundreds of ads across social media channels to test different themes, messages and visuals and discover which creative resonates best with the two NCS audiences of teens and parents/guardians.
Amy Wright, Director of Client Strategy at Automated Creative explains:
“The way those two audiences behave are very different, so we treat them differently. We would run a few 100 ads at a time, testing specific hypotheses in different sprints. For one sprint, we might focus on visual messaging – so what are the visual cues within an ad that drive performance? Then we might move on to specific messages or value propositions, then things like language styles, design formats, is it a video that drives performance, or is it a static ad – all of those different types of things as well.”
This is achieved through its CREATOR tool, which AI audits brand assets and message docs and also builds the ads at scale with automation.
As well as delivering improved performance (31% improvement on conversion benchmark); it also delivered back insight into what messaging and visual themes within the ads the audience responded to most. That insight can be used in other areas – such as CRM or programme development.
Adam Boita, Marketing Director at NCS says:
“The performance and ability to make quick learnings and changes combined with the insights that come out of this rapid sprint learning allows us to ensure we are getting the best ROI.”
Using machine learning to improve targeting
In late 2020, Parkinson’s UK and Prostate Cancer UK both teamed up with AI firm Dataro to find out how machine learning techniques normally used in large scale marketing campaigns could be applied to direct mail appeal fundraising.
Both organisations found AI-driven donor scoring increased revenue and response rates while reducing mailing volumes in their autumn appeals, compared to their normal data selection methods. Parkinson’s UK received 411 additional gifts worth over £15,000 that otherwise would have been missed. Prostate Cancer UK received 440 additional gifts worth over £11,000.
Dataro uses machine learning algorithms to analyse patterns in the charity’s entire history of fundraising, including transactions, engagement and communications data, to paint a much more detailed picture of giving. Each donor is then given a predictive score reflecting their estimated probability of giving. These scores are updated weekly and loaded directly into the charity’s database helping fundraisers choose who to include in the campaign, as well as the best list size and ask strategy.
In this case, Parkinson’s UK and Prostate Cancer UK ran experiments where they generated their own lists using their usual methods. They also generated a recommended list using Dataro’s machine learning.
At the end of the campaign, both lists were compared:The AI generated list resulted in more gifts that otherwise would have been missedThe AI generated list resulted in a higher response rateThe AI generated campaign raised more money while actually costing less to send, saving thousands pounds in print and postage as well as environmental cost
A number of charities have used chatbots for fundraising, including charity:water’s (partnered with Lokai and AKQA) Walk with Yeshi (above), which took people on a two and a half hour walk to collect water with Yeshi, a girl in Ethiopia, and the Children’s Society, which, with digital agency Puzzle, tested chatbots by using one to communicate with people interested in taking part in its fundraising events, sharing useful information such as how to set up a fundraising page.
A well-known example saw WaterAid worked with Facebook-approved developer, The Bot Platform, and Sellu, a farmer and fisherman from the remote village of Tombohuaun in Sierra Leone’s eastern province, to create the chatbot Talk to Sellu.
Users were introduced to Sellu – as well as members of the wider community – and, once subscribed to the bot, taken on a unique journey to the heart of village life in Tombohuaun, the focus community for WaterAid‘s winter fundraising appeal, Untapped.
Through a series of messages, photos and video content, Sellu took users on a guided tour of his home, introduces them to his wife Fatu and daughter Nancy and even teaches them how to say ‘hello’ in the local language, ‘Mende’. There are also updates on the new water and sanitation infrastructures being built with help from WaterAid supporters. Users were ultimately be asked if they would like to donate towards WaterAid’s Untapped appeal to help bring clean water, decent toilets good hygiene to communities like Sellu’s.