Mallory Erickson  01:30

Welcome, everyone, I am so excited to be here today with Shawn Olds. Shawn, Welcome to What the fundraising. 

Shawn Olds  01:37

Really excited to be here. Mallory. Thank you. 

Mallory Erickson  01:39

I have loved getting to know you better over these last few months. Why don’t we start with you just telling everyone a little bit about you? And what brings you to the work that you’re doing today? 

Shawn Olds  01:48

Sure. So I am the co founder and CEO of boodleAI.  boodle uses artificial intelligence and machine learning to help nonprofits find their best donors and acquire new ones as well as help them steward the ones they have along to keep them going in perpetuity finding lifetime value in those donors. I’ve been an entrepreneur after getting out of the army for a number of years. And the one thing I did consistently even while I was in the army, spent 20 plus years serving on nonprofit boards. And as you know, many times nonprofit fundraising can be very ineffective and inefficient. And wanted to find a way I jokingly say the genesis of Moodle is that my co founder and I were basically lazy. He and I both served on nonprofit boards for 20 years, we knew we were going to do it for the next 20. And we just wanted to make our lives easier. And so we wanted to find a very easily consumable way to use AI and machine learning to help both acquisition and stewardship of donors. 

Mallory Erickson  02:43

Okay, can we demystify AI a little bit? And can you tell us maybe for a nonprofit leader who doesn’t really understand the inner workings of AI, or machine learning? What can it do that the human mind can’t do? Like? Why is it an important lens through which we understand our donors? 

Shawn Olds  03:08

I think one of the most important things that I usually if I ever get asked to talk on AI, I always ask the audience, how many of you have used AI in the past seven days, and up until the launch of chat GPT, a few months ago, 50% of the audience would not raise their hand. And what’s amazing is people don’t realize they’re using AI in their daily lives. If they watch Netflix, if they use Google Maps, if they check weather.com, like you can on the laundry list, they’re using AI, they just don’t realize it. So the reality is, AI is just very pervasive in our lives. And it makes our lives a heck of a lot easier than they ever used to be. The second thing a lot of nonprofit professionals fear is oh my god, AI is going to take my job. There’s a steely head with red eyes in it, the Terminator is coming after my job. And we always tell people, the most powerful AI team is the human machine team. And when you ask what is it that the machine is going to do the human can’t? It’s not that the human can’t do it. It’s just human takes a lot longer. And so what the machine does is that campaign work we do ahead of time, you know, where we spend weeks if not months, sifting through data and sifting through spreadsheets, and trying to come up with who’s going to be the right donor. A machine can do that in hours. And so now those weeks or months we missed on potentially fundraising, we can now spend more time fundraising. What the machine can’t do is the machine doesn’t have the empathy we have. If we come up with a list of the best major gift donors to call, and Mallory Erickson is at the top of the list. And the machine calls her and says Hey, Mallory, how are you doing? And he goes, Well, my husband just had a car accident. Our daughter just got in a fight at school and my mom is sick with COVID the machines gonna say, Oh, I’m sorry to hear that Mallory, would you like to give a gift of $1,000 today, which we all know that a good development officer would number one be asking if there’s anything they could do. They would be taking some notes to call you in A week to follow up with you not to ask for a gift. But just to follow up and see how you’re doing. And somewhere two months from now, they’ll come back and ask you for a gift. And so letting the machine take a lot of that workload off of the development team and empower them to then get out there sooner and faster, but use their empathy to be able to work with and cultivate donors. 

Mallory Erickson  05:20

I love that answer. It’s interesting when I was thinking about what can a machine do that a person can’t? I was thinking, and I’m curious what you think about this. But I was thinking that one of the things that AI helps us do is get out of our own biases of things. And so oftentimes, what I see nonprofit fundraisers do is downgrade their donors or I always see people asking, Where can I find major donors? And I’m like in your email list. Right. But there’s like this out there like it’s out there. And so for me, AI helps us get curious about the data we already have, but that perhaps we can’t see clearly through our own limiting beliefs. 

Shawn Olds  06:11

I’ll give you two wonderful examples. One, and and I’ve been frustrated with for 20 years with this is development teams who focus on wealth, yes, we need to look at wealth. Yes, major gift donors are wonderful. But if that’s all you focus on, you’re losing that whole mid level donor base, that can be there for years on end. And one major gift donor that disappears, causes a huge chasm, whereas one mid level donor disappears doesn’t cause a big problem. So we had an organization using our platform that then in their 850 major gift donors, they built a model from that I should say the machine built a model for them, and then applied it to 2000 people who had donated less than $100. And the machine percolated up about a dozen names, the first person on the list that they called made a $20,000 donation. And what they were really upset about was that person had been donating $100, a year for eight years to them had just never been asked in their bias was, well, he’s been through a bunch of wealth screenings, and he doesn’t come up as wealthy. So we shouldn’t ask them. And what they didn’t realize was not only was he a high net worth person, he was an ultra high net worth person. So he had all of his assets and unnamed trust. And because he was so rich, he was retired. So he had no income. So in a while screen, he didn’t look good. But he bought all the affinity markings of all the other major gift donors. The other thing that we saw that was a great tip towards your biasness is we worked with an organization 30 years old $30 million a year eight person development team. And we we walked in, we ask them, Well, what is your best donor? How do you define it? Not what do they look like? Not Who are they? But what defines the best donors at major gifts? Is it monthly recurring? And they said oh no. For five years now we’ve been very focused on somebody who gives at least three times. And every time they increase, we said, okay, great, great way to define your best donor. What does that person look like? And they all almost in unison? Oh, we already know, it’s a married white man between the ages of 45 and 55. And he likes to be communicated with on email or phone. Okay, to your point, we fed all their data into the machine. And in less than an hour, the machine told them that they’re best donor, by their definition, was a single woman in business between the ages of 25 and 35, who’d like to be communicated with that text. Now for three years, they had been spending marketing dollars on a white married man who wanted to be communicated with on email or phone. And when we asked them, Why is it that you think it’s this white Mary, man, they all shrug their shoulders, they all said, Well, that’s who we talk to every day. And what came out is it was the businessman who had a little extra time on his hands had time to actually pick up the phone and call them as it came out wanted to be thanked, and is one of the women on the team eloquently put it wanted to mansplain to them how his money should be spent in the nonprofit. What was interesting is when they redid all their creative the next month for that 25 to 35 year old single woman, and they sent out text appeal instead of anything else. They had over a 35% increase a month a month revenue. So for three years that bias had wasted marketing dollars. And so you’re absolutely right. The machine can help us take away what is just natural human bias that we live with every day and help us see things we may not see otherwise. 

Mallory Erickson  09:21

Hmm. One of the things I love about boodle is that piece around how do people like to be communicated with both because it increases the effectiveness of those communications but I also find that it increases the confidence of the fundraiser because it’s so easy to be like they don’t want to get a text because I don’t want to get a text or they don’t want to get a call because I don’t want to get a call. And I think when you can have some support through AI to say actually these people do like getting phone calls. You know, we are not our own ideal donors. And so we get in our way there I love that. I’m curious when you’re looking at kind of AI across the sector right now, what are you the most excited to see happening? 

Shawn Olds  10:08

I’m excited to see the different branches of AI taking hold within the nonprofit community. I mean, I’m excited to see in the commercial community. But when we started the company six years ago, there were only four companies that claim to be doing AI. One of them was not actually doing a it was doing deep math. But there was no Venn diagram amongst us. We were all doing that same thing I got asked at a conference recently, is your company does it have the best AI, we have the best AI for what we do. But there’s natural language processing AI, what you see with chat GPT now, that helps you write your first draft of your letter you’re gonna send out, there’s all sorts of AI out there. And what people are realizing is there’s different ways to use it, harness it, and then be able to leverage it to really help them and their teams and what they’re doing. And a lot of people are like, Oh, isn’t Chat GPT scaring people? I think it’s the first time people are seeking out AI and getting a favorable result from it and not the steelhead with red eyes coming up, though. And that, to me is exciting. 

Mallory Erickson  11:08

Yeah, it has been so interesting to watch it’s explosion, especially when I feel like there have been tools like copy AI, and magic write inside Canva that nonprofits were perhaps even closer to, but maybe because of the quality that chat GPT started with, and just sort of the mass explosion of interest and free, it has gotten a lot of people’s attention, and I think, lowered the bar around how scary it is like how possible it is to just get a response and see how it works. For the folks who are still really creeped out by AI. I hear this with automation, too. Sometimes in the nonprofit sector, like our job is deep relationship building, how on earth could we use AI? And it makes them there’s almost this like identity conflict for them or something and using it, what would you say to some of that pushback, 

Shawn Olds  12:04

You actually gave a great example. I mean, the first step is just getting comfortable with technology. You know, five years ago, I don’t know a single major gift officer that would reach out to any of their major gift donors on a text. But if that’s how your major gift donors like to be communicated with, it may not be how you do your ask. But if you need to steward them through the year, sending them direct mail, they’re never going to open is not going to help you when you have to go back and talk to them. If they live on text, then why not give them tax? If they live on Facebook? Why not meet them where they are. And as you send out your stewardship, not your ass, but your stewardship, why not meet them where they are. So when you do pick up the phone and make that ask, they feel like they’ve been with you the whole time. And not I talked to you a year ago, and now you’re coming back a year later asking for another major gift. So one is just getting comfortable with technology and that you made a great point, we are not our own best donors, our best donors don’t look like us. And so let’s not impose our beliefs onto our donors. Let’s take our donors beliefs and use them and meet them where they are. Once you get through that it’s the realizing there’s no way as a human being, I can possibly assimilate all this data. If I want to be able to get there if I want to be able to understand and really give my donors what they deserve. I need to have every tool out there I can. And one of the great things you asked a little bit earlier about the things that excite me about AI. One of the other great things is to see AI that’s being used not just in fundraising. So there’s AI being used in the nonprofit sector to make their nonprofits more effective. If you look at like the teen tech suicide hotline, they’re using AI to understand when texts come in, which ones are the most problematic. You go back to your question on bias, there was an assumption that if a kid mentioned the word suicide, that they were going to commit suicide. And what they found over almost, I think it was six or seven years of data and texts, that a kid mentioning ibuprofen, was actually more likely to take action. And so they’re able to fast forward those calls to human being and get the service needed to a kid before they take action. If you look at United refugees, they used to literally go over with a big book of pictures, and just thumb through them and maybe make a couple dozen matches. Now they send one person over with an iPad, and they just face scan and using facial recognition, which is a branch of AI. They’re not making a dozen. They’re making hundreds of matches and reuniting people with their families. And so those type of things where people can get comfortable, I can do my job better. I can service my donors better in the way they want to be service. That’s where people have to get through, they realize they’re actually neglecting their donors if they’re not using these tools. 

Mallory Erickson  14:42

Yeah. And you made this point earlier about the fact that these tools actually give us the time and capacity for deeper human connection, you know, and so and we can actually save time we’re sending, trying to get people to respond to a survey about how they want to be community hated with, we can send them something way more valuable and use AI to determine some of those things. Also, because we’re such bad predictors of how we actually will behave, right saying what I like this versus how we behave in that moment is so different. And so I love the way that AI helps us cut through some of that noise. Are there any concerns that you have about AI, that how it could be misused in our sector, or what we should watch out for as folks are adopting different types of tools in this space? 

Shawn Olds  15:29

Absolutely. I mean, everyone gets concerned with bias and bias, as people start to hear about AI has become a very negative words, there is actually very unintended bias that happens in algorithms. As a great example, we worked with a university, and their models just came out horrible. And it wasn’t until several hours of kind of delving in that we realized that this university, unlike other ones we’d worked with was a very local university, meaning everybody kind of applied from the area. And then they went to jobs in the area, which meant that location was a bias. And when we pulled location as a data point out, all the models work wonderfully. Now the problem is people can be malicious with it, they can feed the data in that they want to get the result that they want the machine and it can only do, it can only make, you know, insights and responses based on the data that’s fed into it. So if we feed it, bad data, bad results out. And so we definitely have to make sure that as we’re using models, as we leverage models that we understand, we don’t need to actually see the data, we just need to understand where the data came from. And be confident that that data is unbiased data that’s being fed into it. 

Mallory Erickson  16:35

Wow. Yeah, that’s interesting. What’s really interesting to me about the use of AI and nonprofit in particular, is that leaders are stretched so thin, right. And so I just see it as such a like time saving tool. And even someone posted recently on LinkedIn about everyone’s obsessed with chat, GPT, but hear all these other AI tools that you should actually be watching. And I was like, Oh, my gosh, there’s that and that and that and that. And, you know, I’m not the person who’s the most up to date on all of the tech necessarily, but it really made me realize like, teams could be so supported with a variety of different tools. If AI is being used by certain individuals on a team or to power certain verticals in a organization, do you have any recommendations around how the team as a whole handles data or shares data, or make sure that their data is what they’re pulling, that they’re sort of doing some of that analysis around where there might be bias getting stuck in their models, and you kind of best practices there. 

Shawn Olds  17:46

Yeah, one important thing is as nonprofits look to work with providers, who they’re going to hand their data over to, they should scrutinize the terms of service agreement, and make sure that their data can’t be shared. There’s a lot of people who to save money, they want to go with the free service, free is only free so far, because a lot of those free services are taking the data and reselling it. That’s how they make their money. Companies, if it’s a for profit company, it’s got to make money somehow. So most organizations like a CRM, and the strong AI platforms will tell you in their terms of service, we cannot use your data for anything but what you guide us to do. So there’s no selling or sharing or doing anything with your data except what you tell us to do. And that’s the beautiful thing about AI is once you feed your data in, there’s no PII when the models are built, there’s no way to backwards trace if Mallory was a donor, once that model is there, there’s no way to tell that Mallory fit into that model. And so it makes it a very easy way to amongst nonprofits that they want to eventually to be able to share models and leverage models amongst themselves without ever giving up their donor data. So first thing is make sure that your data is not being kind of stolen out from under you, and sold off. The second thing is just making sure teams talk. Most nonprofits are so small, they don’t have to worry about it. But as you start to get into bigger nonprofit, and you’ve got 1520 people on you’ve got different teams is as with any technology, make sure you’re not buying duplicative technology. And then as you are buying technology, get the most out of it. We have several clients who we’ve been brought into, we get used in one small area. And then finally, after nine months, we’re like, wait a minute, you can help this team too. And it’s just getting out to other teams. It doesn’t cost them anymore. It’s the platform is the platform, but making sure you get out so as you get new technology or AI makes sure the entire nonprofit team knows about it to see if they can leverage it to 

Mallory Erickson  19:36

okay, I’m wondering if you can just explain to people what a model really is and how boodle uses models for prospecting because this is something that I think is really interesting about your technology in particular. 

Shawn Olds  19:52

Absolutely. So models are built from the data of the nonprofit. Now there’s different data that could go into it. We know that most nonprofits have very limited data. And so we decided to take a focus to what we call affinity based modeling. And we keyed in on that 20 year old statistic that 80% of first time donors never come back, they come in for the Ice Bucket Challenge roommate runs arrays, the 20%, who do they come back because they care about the cause they have an affinity. So what we do is we take a name and an email address of the donor. And that’s it, we perform identity resolution, we map them into our proprietary database. And we then enrich those donors with over 1200 data points that all point towards their affinity. And then we use an algorithm, we actually use a series of algorithms, we test which one works best, when we find the model that works the best. That’s the organization’s model. Again, there’s no PII in the model, we let them apply that model to our database of all 240 million Americans. And now we can immediately give them a total addressable market, we can tell them if this is your best donor, there are 6.2 million Americans who look just like them. And then if they want to start reaching out to them, we can help them reach out to them on Facebook, banner ads, LinkedIn, wherever digitally, they think those people may reside, 

Mallory Erickson  21:05

I would really encourage folks to go check it out. See, I mean, I think the accessibility through which you provide your services is as somebody who works with a lot of small nonprofits is really appreciated. So thank you for sharing all of your wisdom and all the insight into how folks can be best using AI. And I love the stories that you shared about some of those big learnings and takeaways. So thank you so much for your time today. 

Shawn Olds  21:30

Mallory, thanks for having me. I appreciate the time. It’s great to be with you.

Mallory Erickson  21:39

Takeaways inside this episode, but here are a few of the ones that I’m focusing on right now. Number one, identify the preferred communication channels of your donors and meet them where they are using technology to enhance your relationship building. And if you don’t know how to figure out preferred channels, that’s where a tool like Google AI can come in. Number two, you can utilize AI to analyze and understand donor data to make more informed decisions and better target potential major donors or folks who would like to be involved in volunteer opportunities. There is so much potential for segmentation when you use something like AI. Number three, AI has the potential to be incorporated into nonprofit operations to increase efficiency and effectiveness, such as using facial recognition for family reunification, or AI driven text analysis for crisis hotlines. Number four, leverage AI to identify and address potential biases in your fundraising strategies and donor segmentation. Number five, embrace the human machine team approach to maximize the strengths of both AI and human empathy in your fundraising efforts. And lastly, invest in AI tools that can save time and resources, allowing nonprofit professionals to focus on building deeper connections with donors. It’s not about replacing people, it’s about optimizing people. Okay, for additional takeaways and tips inside this episode, head on over to Mallory erickson.com backslash podcast to grab the full show notes and resources now. You’ll also find more information there about Shawn and boodle AI. Thank you for spending this time with us today. If you enjoyed this episode, we would love it if you would give it a rating and review and share it with a friend. I’m so grateful for all of my listeners and the good hard work you’re doing to make our world a better place. And if you miss me between episodes, stop by and say hello on Instagram, under what the fundraising underscore. Have a great day and I’ll see you next week.