תעתיק 🔗
[00:35]
Matt Smith: And welcome to Something to Think About, our podcast about strategy, leadership and change in fundraising. I'm Matt Smith, Managing Director at THINK.
Tim Beynon: And I'm Matt's co-host, Tim Beynon. Today we're talking about something that sits underneath almost everything fundraisers do, but which, I think it's probably fair to say, we don't always make the most of. And that's data. And I suspect this is going to be one of those conversations where we start off talking about databases and dashboards and very quickly end up talking about strategy.
People, culture, decision making and, inevitably, AI. Because let's face it, charities certainly aren't short of data. We collect enormous amounts of information about supporters, donors, campaigns, channels, behaviour and income. The much bigger question is: are we actually using it properly?
Matt: Exactly that, Tim. And our guest today is somebody who spends his working life thinking about precisely that question. The brilliant Sameer Rahman. Let's go to our conversation with Sameer now.
[01:39]
Tim: Sameer Rahman is the founder of DataMonet and one of the UK's leading voices on data, analytics and AI. He's been recognised as a Top 100 Global Data Leader and Wales Data and Tech Leader of the Year, and has held senior data roles and worked with organisations including the Royal Mint, Go Compare, Lloyds Banking Group, Legal & General, Jaguar Land Rover and BT.
Matt: But just as importantly for this conversation, Sameer has also worked with charities including Marie Curie, the Royal British Legion and Anthony Nolan. And, of course, has worked alongside us here at THINK. What I particularly love about Sameer's approach is that he talks about data in business terms rather than making it a purely technical conversation. His business has a very simple ambition: to get data and AI into the boardroom, so that more of the decisions organisations make are actually backed by evidence. So, Sameer, a big welcome to Something to Think About.
Sameer Rahman: Thank you, Matt. Thank you, Tim. Nice to be here. Nice to have a good discussion.
Matt: Yes, really looking forward to this one. So let's kick us off, Sameer. Let's start with your first impressions. When you first start working with an organisation, how quickly can you tell whether they're actually any good at data?
Sameer: So it's a very fundamental question. I ask very fundamental and basic questions, Matt. My first question is: do you argue about the trustworthiness of the data more than the decision? So is it a case that, if you ask for some data or insight from your team, you get answers which you're not sure whether they're right or not? And most of the time in the boardroom is spent arguing about the data rather than working on the insight generated by the data. That's number one.
Number two is speed. For example, if somebody asks, "I want this particular information", how long does it take for your team to get back to you with that information? If it's taking too long, or people are not sure about it, then again, that speaks volumes about your culture, about your architecture, about the quality of your data and things like that.
And number three, I'll look at business cases and I'll ask the board, "Show me the three or four big business cases you have signed up in the last year or last few months." And I'll see what data evidence has been part of that business case. Those three give me a good sense of how the data culture is. If it's taking too long, if the business cases don't have the evidence and are backed more by gut feel, and also if there's more argument about the quality and the trustworthiness of the data, then that tells me that the culture around data and evidence-based decision making is probably not at the level that it should be.
Matt: And to follow up, I think what's coming through in your answer there is a bit of confusion around having lots of data and being data driven. And so is it usually the tech that's the problem? Or is it the people and culture, or a bit of both?
Sameer: People are the problem and the opportunity. I think the biggest underrated issue that people don't talk about that much is business process. Because what technology does is it either automates or optimises, but if the business process is not working, it automates the wrong thing. And that's the biggest bugbear. I think business process, as I said, is the most underrated thing. Now, technology is the most talked about thing. In my view, technology is really the least of the problem, because technology has been commoditised; everybody has got hands-on experience of various technology problems. How people use that, how people use that for good decision making, is more of a problem. So the culture, what do you say to people in terms of how you use that technology, is more of a problem than which technology to source.
Tim: Sameer, I'm really interested to hear what you're saying about getting data into boardrooms. And I'd love to know what that actually looks like in practice, because businesses around the world, especially today, have so much data at their disposal. But what does that look like in terms of getting into the boardroom? What should a CEO or a board of trustees actually expect to know?
Sameer: The biggest thing for me is: do they have the answer to the biggest strategic questions that they can? So in the charity world, can you confidently connect the amount of fundraising that you get to the amount of beneficiary benefit? If I ask fundamental questions to the board, saying, "If I get you 10% investment now, do you know, first, where to invest? Are you confident? Second, do you know how that will impact the beneficiary?" If there's no link between fundraising and beneficiary, then the more strategic question is not there. That is my boardroom conversation. I'd say, okay, why is the link not there? Why are you not confident of where to invest? Why are you not
Load full transcript
confident of whether you're investing in the right thing for the beneficiaries to have an impact? So that, I would call it the unified or 360 view of success of a charity that they should be looking at. That's the biggest boardroom conversation, or one of the first boardroom conversations I would have with a charity's trustees and board.
Tim: In your experience, Sameer, when directors and managers are preparing board reports and preparing documents to put to their board of trustees, do they fall into the trap sometimes of reporting what's easy to measure rather than what actually matters?
Sameer: Yeah. I would even go down the line that, in the charity world, from my experience, they would report on what they can measure, because there are quite a few things that they want to measure but can't. There are quite a few things that they haven't even thought of measuring. So this unified or 360 view of success, not many charities have even thought of measuring. So you won't find that. Again, Matt has worked with the third sector more than I have, but you won't find that in a typical board report: what is the link between fundraising and beneficiary? So there are three layers to that, Tim. One is: are you measuring the most strategic issue? Second is: you want to measure these, but you can't, so why can't you measure this? And the third thing is not just "I'm giving you what's easy to measure", I'm giving you what's possible to measure, because my data only goes so far.
Tim: Yeah. And I think we could all relate to that. Charities I've worked at in the past, I know that sometimes you want to report something, but actually managing to find that data is so difficult because it's stored all over the place in random spreadsheets on someone's computer somewhere and all that kind of thing. It can be really difficult to actually find that information. But tell me, you've worked in big commercial businesses and also, of course, in charities. What's different, do you think, about how we use data in this sector, in the third sector, compared to some of those big organisations? Obviously, there are massive differences in terms of resource, especially what goes into it. But is there anything that we're doing well, or are we just being universally left behind by the corporate world?
Sameer: Obviously, the fundamental difference is in the outcome, really, isn't it? A charity's outcome is "I want to help more people". That's the ultimate ambition, that's the purpose of the charity. And financial services, or any other commercial organisation, retail, will be more around the commercialisation part, and basically the customer and the commercial part. Charity is more about beneficiaries. So the end outcome is definitely different.
What charities are doing differently is that charities rely a lot on CRM. CRM is the backbone of most charities, which is, well, I'd say not so much the case in the commercial sector, which has got a much larger infrastructure, while for
charities the CRM becomes the backbone. The whole ecosystem circles around the CRM, email platforms, communication platforms and things like that, which means that if you don't get it right, it's very difficult for charities to function.
What charities are doing better, I think, is that, again, charities are taking a much more focused approach, saying, "Okay, fundraising is my main approach." So the one thing charities do well is, depending on the resource and so on, they are a lot more focused in their approach. Commercial organisations have quite a wide remit: you have to satisfy the customer, you have to get more out of the customer, you have to capture a market. So you need a much bigger remit, and you become less focused in a way because there are so many objectives to meet. In the charity sector, you get more fundraising, you use your CRM and the infrastructure around it, and then you help with that fundraising, you help the beneficiary. So it's a good single golden thread, and quite a simple thread, rather than being overly complicated as in the commercial world.
Tim: And have you seen, in your experience over the years, Sameer, that charities are investing more in actually coming up with a data strategy now? Because obviously they have strategies and they've got to focus their efforts across both fundraising and service delivery and income generation, and their resources are always spread thin. But are you seeing increasingly now that data is becoming more important and they are actually investing more, strategically, in data?
Sameer: Yeah, I mean, yes and no, to an extent. The one thing linked to this, Tim, is that charities have always lacked, and we know the reasons why, that senior leadership around technology and data, which basically uncovers and educates, and does the data and technology literacy across the board and across the organisation, saying, "This is why you need a data strategy and this is why I'll champion this cause." They've lacked that layer of leadership. So, organically, charities are coming to that. The CRM is not working, or they don't have a single version of the truth, or they don't have a single supporter view. All those different kinds of problems then ultimately lead to "Right, we definitely need a good overview, a data strategy. We need a proper strategy which addresses all these key points in one go", rather than having a tactical, solution-oriented mindset. So charities are coming round more to "we need a data strategy", but because they don't have a data leader, it's slow. And because they are uncovering problems as they go along, it's more of an organic endpoint rather than a natural endpoint.
Matt: Your reflection, Sameer, on the CRM really resonated with me, because the one bit of feedback we always hear is that the CRM is the problem. It's often loaded with everyone's challenges and problems; they just get lumped on the CRM as the central issue. We hear that so, so often. I would love your take on
how often it's really the CRM. How often is it actually that the CRM is the issue versus the other things that we've been speaking about?
Sameer: Look, the main thing I'll say, Matt, is the definition of CRM. CRM stands for customer relationship management. And people think CRM is a technology. So the main problem they cite is "my technology is not working". That's the main gripe you hear. When you look at why the whole customer relationship angle is not working, the main point is: fine, the data feeding into the CRM is not working. The system itself is not being used in the way that it should be used. People are not trained on the system. People don't know what the CRM is capable of and what they should be using the technology for, to further the customer relationship. So the whole ethos of customer relationship, and the strategy around how the customer relationship should be enhanced, or the fundraiser or the beneficiary or the supporter relationship should be enhanced, that's the whole ethos. Technology is just a small part of it.
So where charities are getting it wrong is using the CRM not for the ethos of CRM, but normally as a technology: "I need to use this technology somehow." The strategy, the customer or supporter fundraising strategy, needs to include the CRM as part of that strategy, rather than the technology coming first. It's the other way round. I think the whole problem is looking at CRM as a technology rather than as a customer management enhancement tool. And that's the main core issue, Matt, that I see.
Matt: So it's about getting those fundamentals right before making big investments in CRMs. That's one of the big questions that I think comes our way so often: we're about to embark on a big tender process that often has hundreds of thousands of pounds earmarked against it. It feels the weight of the decision, and the dream that if we just make this one big move we'll end up in a utopian scenario where everything is working exactly how we want it to. So, from your perspective, Sameer, what are the three to four fundamentals that need to be sorted out before those big spends on new systems are made? What are those key fundamentals?
Sameer: One, definitely, number one, is: what problem are you trying to solve, really? What is the business use case, and what is the business strategy that you're trying to solve and align to? So that's number one. Number two will be, because any technology is only as powerful as the data that goes into it, and any technology is only as powerful as the way people use it, so forget about technology. Technology comes to me as a fourth tier. First, the business problem you're looking to solve. Second, what data enables you to solve that business problem? So again, Tim's question around the data strategy: even if you haven't got a data strategy, have you got a high-level roadmap or a guidance document on what your data should be and what data should go into the CRM? That's number two.
Number three,
I would say, again where many charities have got a big gap, is that architecture roadmap. What role does the CRM play in the overall ecosystem? Charities are using the CRM as a database, they're using the CRM as an email marketing platform, they're using the CRM for multiple other purposes. So be very sure: are you using the CRM for all of the above, or are you using the CRM for a particular function or a particular use case?
Then comes number four, which is technology. You've got the business problem, you've got what data feeds into that, you've got the architecture and the role of the CRM in the whole organisation to solve that problem. Then you come to, okay, which technology is the right technology for me? And obviously, running in parallel to all of this is people and business process. That's how I see the ideal scenario. People: you need to train them, they need to have a voice, they need to get engaged, they need to figure out what the business problem is in a collaborative way. And business processes: sometimes you've got the right data, but you don't have the right business process to get that data in, and you don't have the right decision-making process either. So that'd be my top four, and those two parallel strands of process and people.
Matt: It's so interesting thinking about it being sequenced in that way, Sameer, because it feels like the reverse of the conversation that I often have with folk. Eventually you might get to why we needed a new CRM, versus how you've explained it and sequenced it. It just makes so much sense.
Tim: Yeah. It's fascinating. So I'm really interested in what you're talking about in terms of people and processes, because in organisations I've been involved in and worked in in the past, I can see where that's fallen down. And I think every fundraiser listening to this will recognise those common issues that they face and that charities face: duplicates, missing fields, data randomly sitting on spreadsheets that live on someone's laptop, God knows where, all this stuff that sits outside the CRM. Is there ever a case, because many fundraisers will argue this, is there ever a case for keeping it old school? Or in this day and age, is there just no excuse for poor data, or poorly kept and managed data?
Sameer: I think you run the risk of keeping it old school, because you run not just the risk to the benefits, but also the GDPR, security and privacy risk of who has access to what data. You can't govern it, you don't know, and it gets out of control, whether they're downloading it onto a personal machine or whatever. So there's a risk element to that. Whether there's an argument for the old way, it really depends on the people and the process and how fast they come. If people are not trained enough to use newer technologies and newer ways of analysing or collecting data, then what can you do? It's a tail-wagging-the-dog kind of thing, where by the time you get people trained, you have to work in the old way.
Similarly with the business process: if you can't change a process and automate the process, and the spreadsheets don't connect to the CRM, for example, then again, before you get to that stage you have to follow the original process, because you have to keep the lights on, you have to keep the business and the charity running. I don't think there's any... we need to move; charities need to evolve, charities need to move on, charities need to integrate, automate. But before they do that, there's that people and process element which they need to sort out. There is no reasonable argument to work in the old way, both from a risk and a reward perspective, Tim.
Tim: Yeah. And I totally get what you're saying. And quite often in those kinds of situations what you'll find, or what you know, is that charities just become overwhelmed by not knowing where to start when it comes to their data and managing all this data that's sitting all over the place. If you were to walk into a charity that was in exactly that boat and feeling totally overwhelmed, where they are in terms of how they're using the system, or misusing it, or not using it properly, their processes are all over the place, data's all over the place, what would be the first things you'd look to try and help them with or try and fix?
Sameer: I think, again, I'll start with the business problem, saying, "Okay, fine, there's so much to fix here. We can't bite it off in one go. So let's work out what is the main business problem you're looking to solve." And they might say the main business problem is "we can't connect a supporter to a beneficiary", for example, and "we can't say, if you give me ten percent investment, or twenty percent, where should I invest?" And so if that becomes the question, then the second question, where should I invest, becomes the key question. Then I would say, okay, fine. So for this, that second point, this particular data set is more valuable than the others, so I'll concentrate on that. And then the third point: how does it fall within the architecture? Okay, this is just a CRM system, so how do I use my CRM system? Then the people aspect: who uses that? And then how does the data go into the CRM, which is a process aspect? So I'll go with those four, with the business processes, to solve the decision making of "somebody invests twenty percent more". If I suddenly get twenty percent more fundraising, where should I invest, and which campaign should I invest in? So I'll go down those routes, and eventually I'll land on technology as the fourth, in terms of how to use technology. But the business case starts at the very, very top, Tim.
Tim: Can I just ask you, how would you bring along with you someone who's reluctant when it comes to data? There's that fundraiser who has always used the spreadsheet, always wants to use the spreadsheet, can't see the point of a CRM. How do you bring them on that journey with you?
Sameer: Look, people, it is a difficult one. People
are people; we are all used to our behaviours. Essentially, what you're asking is to change behaviours, change mindsets, change yourself in the way you work, which is difficult stuff. But it's more around showing them. First of all, giving them some sort of education, literacy, awareness of why newer ways of working will be better than the old ones. If you can showcase it on a small-scale basis, saying, "Look, I've done this in five minutes and it takes you half an hour to do something", there's often a penny-drop moment, saying, "I didn't know I could do that." It's happened to me quite a bit. And I've been on both sides, really, showing people how quickly you can do that and also learning from people how you can do that. So it's eventually down to them getting onto that kind of level playing field, saying, "Okay, I now understand why you're doing it. You're doing it for the betterment of the day in the life of me. So my day will improve rather than the other way around." Once you get them to that level, saying "this is to help improve your day and your job role so that you can concentrate more on the strategic issues, remove the frustrations from the daily role", then they become more in tune with what you're trying to do. But it is a difficult one, I have to say. Changing people's mindsets and behaviours is difficult, which is why people are both the biggest barrier and the biggest opportunity in a transformation.
Matt: Thanks, Sameer. I wanted to change gears and move us on to talk about predictive modelling for a few minutes, something that I get asked about a lot. You've done a lot of work in predictive modelling for many years, and obviously I am no expert in this area whatsoever. So, taking the layman's position, what can it realistically tell a fundraiser? And how is predictive modelling different from the segmentation most of us already do and have done for all of our careers?
Sameer: Yeah, look, it's a good question. Predictive modelling has been there for ages. In a commercial setting, I'll describe that and then I'll come on to the charity relevance. Let's say a telecoms provider; pick one, Vodafone, for example. The chain of predictive models goes like this. First: who is looking to buy from Vodafone? There's something called an acquisition model, which looks at the market: what kind of people are most likely to buy from me? So even before people have started looking for a new telecoms provider, they will proactively approach them. That's called an acquisition model. All of the predictive models are forecasting behaviour based on either demographics, transactions, your life profile, your family composition, the way you have behaved in the past. So that's acquisition.
And then once you've bought, the other kind of predictive modelling is cancellation. Am I most likely to cancel in my 14 or 30 day window? So there's cancellation modelling, saying: what is the profile of people who buy on a whim and then suddenly cancel because they
bought the wrong thing? What can I do about them, and how can I reduce my cancellations? Then you come on to: they've bought, they've not cancelled, then you come on to the retention model, which is, okay, after my 12 month, 24 month contract, who is most likely to remain? So that's the prediction of who is most likely to churn or not remain, and what can you do about it before they even start deciding. Then you come on to lifetime value modelling, saying, okay, now I've retained them, how do I sell more stuff or get them to cross-sell, get them to buy broadband and things like that? So the whole area of predicting acquisition, cancellation, retention and lifetime value, that's a whole suite of predictive models. What the whole lifecycle does is give you proactive steps in helping people do something that they haven't thought about.
Where segmentation comes into that: segmentation is not predictive, segmentation is descriptive. It tells you what people have done in the past; it creates a profile of what people have done in the past. It helps you make a decision, it doesn't make the decision for you. Predictive modelling makes a decision for you. It scores every individual; consider every individual on the database has got a score, and the highest score is the person who's most likely to buy from you, so target that particular person. Segmentation will not do that. Segmentation will not score everybody. Segmentation will give you the idea of the type of person and the quality of person that will do so.
Segmentation is still very important because it serves quite a few purposes. One is that it can help you decide on your proposition development: this is the type of profile, they care about the cause. And this is my market. It helps you in market sizing, saying what is actually your market. It helps you in campaign automation, saying, okay, I know from a predictive model that this is the person, but actually they look like they're of a certain age, of a certain gender, they have been on the website before. So the predictive model draws on all that information from the segmentation. It helps personalise the email campaign or the direct mail campaign or the other campaigns. So that's the difference between the two.
In the charity sector, predictive modelling is very much useful. Same as in the commercial sector, it tells you which donor is most likely to give. Legacy modelling is a big area of focus for charities. Legacies are obviously a key part of how many charities get their donations. Even before people have thought about legacies, how can you start targeting the right people through a predictive model? So legacy modelling is big. Similarly, converting from a single cash donor to a regular donor is another predictive model. So it's the same concept. It's just future-proofing and being proactive in who you're asking for what. That's the kind of use of predictive modelling within the charity sector.
Matt: I think that's the best explanation of
that subject I've ever heard. In my 17 years in the sector talking about segmentation, I don't think anyone's ever explained it quite so clearly and obviously to me as right there. Thank you.
Sameer: My pleasure.
Matt: So I get asked a lot about whether demographics are still important, because obviously a lot of my work crosses over between fundraising and marketing and brand, and lots of different folk have different perspectives on whether audience demographics are still really, really important. So it's probably too tricky a question to answer in a short space, but I would love your thoughts on whether those audience demographics are still useful, or whether it's back to behaviour being more important, at the heart of what you just said.
Sameer: It's more of a 360-degree view. When I say segmentation, Matt, there are various variables in how we live our lives and consider ourselves as human beings. Yes, we are of a certain age, certain gender, family composition. And yes, some of the decisions that we make are based on our demographics, but not all of the decisions. So demographics is a foundation stone for me for segmentation; it definitely sets the foundation. On top of that, look at past behaviour, because even if we are of the same age and the same gender and the same family composition, we can still make a lot of different decisions. So that's where the behaviour angle comes in. Okay, why is he buying, I don't know, a green trouser while somebody else buys something different? That's where behaviour comes in. And then the third thing is attitude to life, the third layer on that, or the motivation. So I might be giving to a charity because of a personal cause. Somebody else might be giving to a charity because they just like that charity, or it's around the corner from where they are, or they've met somebody who really has recommended that charity. So demographics is a foundation stone, then behaviour is the second pillar, and attitude and motivation to donate, or to buy, whatever the case might be, is the third pillar. And then all those three pillars should eventually give you the why: why they are making the decision they're making.
Tim: Fascinating stuff. Sameer, I want to just come back a step, if that's okay, to predictive modelling again. And with my charity comms hat on, I can't help but think there's a bit of a tension there as well, because isn't it a case that the better we understand people, the more responsibility we have, especially when it comes to how we use their data to interact with them? And we've seen charities get into hot water in the past with things like wealth profiling, potentially. Where's the line between good personalisation, using data really well to communicate effectively and to achieve the results that a charity wants to achieve, and something that just feels a bit creepy, a bit too close to the bone, perhaps, a bit too much information? And what should the test be before we collect another piece of information about somebody?
Sameer: Okay, it's a very good question. The world has moved from mass marketing to segmented communications to personalisation to hyper-personalisation. The creepy bit... as we have moved, people have also got used to some of these things, not saying it's a right or wrong thing, but eventually, I mean, Tesco Clubcard started it: you buy your stuff, I know you, I know what you're buying, every offer that they give is specific to you. Some people might find it creepy, some people might find it very useful, because "I need to buy washing powder and you're giving me a discount on washing powder. Great, I'll go away and buy that."
So there are two things. One is the legalities of it: what information do you collect and how are you using it? There's very much a value exchange in why supporters are giving their information to you, and that value exchange needs to be very clear right at the start, in terms of privacy policies, in terms of other data sharing agreements: why are customers or supporters giving their data? So that's the bare minimum hygiene factor. Then it's a case of testing in terms of where, as to your point, Tim, where do you tilt the balance from useful personalisation to creepy personalisation? So you start testing, and this is different for everyone, because we are people. Again, if somebody gives me a very personalised offer, I would love that, but I'm sure my wife wouldn't. So it's a case, again, of going back to segmentation and attitude to risk taking. You test that with a cohort of your supporter base, saying, okay, currently we are sending them the same email comms. We'll start sending them clustered, segmented email comms. Now we'll say, "You supported us back in 2025 with a one-off donation. We are running this campaign. Do you want to support us again?" It becomes much more personalised comms. And then you become hyper-personalised, saying, "I know you supported me, and you're..." well, you wouldn't say that, but you said in the demographics that you've got children and so on, so we're doing a children's campaign, or whatever the case might be. So it's a case of testing that slowly, to see where you get complaints, or where you start getting into that territory where people really do not like it and your response rate drops. I would say people have got used to that personalisation element. What they haven't got used to is the hyper-personalisation element, which is your creepy bit that you mentioned. So I would say the hygiene factor is definitely that personalisation. Moving beyond that is a case of testing with your supporter base.
Matt: Good stuff. As we start to come to the end of our conversation, Sameer, we can't have you on without talking about AI. I think we're contractually obliged to mention it. It's obviously been in the news a lot of late, and some quite terrifying headlines of late as well. But I'm always keen to try and ground the conversation in the reality of where we are right
now, and I always find that our sector likes that real, practical kind of advice. So, from your perspective, what's the reality when it comes to data and the AI conversation? And how can charities both set about making good use of it whilst also protecting themselves, in what is obviously an interesting moment for us?
Sameer: Yeah. Look, data and AI go quite hand in hand. AI is very data hungry. You can very seldom have an AI application without data. Even if you consider a very simple chatbot, it takes data in, it gives the response back, and that response, it's like the frequently asked questions that we had in the olden days, where people used to look at the frequently asked questions and that's about it. Now a chatbot becomes that at a very elementary level; it becomes a frequently-asked-questions tool. So in terms of whether charities are... first of all, to have any AI use case, they need to have that data foundation ready, in terms of what I originally said. The data feeds the AI application. You should not be afraid of AI, because it's in your benefit to learn how to use AI. You might use it for a productivity increase for a certain function. You might use it, let's say, for answering basic fundraising questions as a chatbot. You might use it to uncover fraud and fraudulent donations. So there are quite a few boxed use cases.
My recommendation to charities will be: fix your data foundations, because every AI use case you try will depend on a certain data foundation. But also, don't be afraid of AI. Again, back to: what business problem are you trying to solve? If you've got a clear view of that, do test an AI feature. It might be customer service, it might be fraud applications; predictive modelling has got a lot of AI applications in there. So you need to test that. And it might be that you give certain boxed access to certain people, for your development team, for example. Claude is good at coding and it has taken software engineering into a new phase, absolutely fine. Give small-scale access to your development team or to your technology team to work on that, see what comes out of it. Test it, trial it with a boxed use case. And then if it works, great. Try to operationalise it slowly and then try another use case. If it doesn't work, try another use case again. But don't be afraid of that, because the fearful thing about AI is investing a lot in it and expecting a lot of magic when the foundations are not fixed: you haven't got the right technology, you haven't got the right people using it, you haven't got the right business process. It's not magic. It's like any other new technology that comes in. It's far more advanced, but you have to fix the foundations to make it work. That'd be my view.
Matt: Okay. And it's really good advice on where organisations should be starting out. Is there anything that you would warn charities against approaching at this point? Again, one of the questions I get asked the most is what you've answered
in terms of "we don't know where to start, how do we start?" But then the second question behind that is that risk factor, and making sure... I think there's a great fear across the sector at the moment that a staff member is unwittingly going to step into something that causes great damage to the organisation. I think that's the great concern: the control and risk mitigation factor to it. So is there anything that you would advise when it comes to that, protecting the organisation? Because these are very powerful things.
Sameer: Yeah, again, my answer, Matt, is to come up with a boxed use case and solve a particular business problem. Because what you don't want to do as a charity is be on the back foot and not have any AI conversation, because then it'll be far too late. Also, what you don't want to do is have a blanket AI philosophy. It's very much focused on: what is your AI application to solve what problem? And then, based on that, who is the most appropriate team or person, and which business process is it fixing? Once you've got that, it becomes quite a focused conversation: okay, so we need a Claude tool for a developer, just two licences. I don't have to waste a lot of money. The end outcome will be this. We will test this in a certain period. We will look at, in three months' time, what we've achieved. That becomes your very boxed use case, which gives everybody in the organisation the message: okay, we are moving ahead with AI in some experimental way. If it works, we'll start to embed that. But be very boxed, be very clear on the use case and the business outcome you're trying to drive.
Tim: Similarly, Sameer, do you think charities need to be transparent with their donors and supporters in terms of their use of AI, particularly if their donor or supporter data is being put into some kind of AI tool? Should charities be upfront and transparent with those donors and supporters and tell them, "We're using your data in this way"?
Sameer: Yeah, I think charities need to be transparent with donors and supporters. It's like any other privacy statement: if you're collecting data, you should say why you're collecting it, where your data is going, and what tools are used to analyse it. Not necessarily naming the tools, but certainly naming, in terms of AI, that your data will go into an AI tool or an engine which will analyse the data. So you have to be as transparent as you can, but you don't have to be very specific in terms of "this is the data that will go and this is the tool that will be used". So I suggest we are transparent to an extent, "to an extent" in a good way, where consumers don't get confused. But you have to be very transparent and careful about it.
Tim: Yeah, good advice. I think having another look at your privacy policy is probably a good idea, I imagine.
Sameer: Absolutely. Absolutely.
Tim: Can you give us, with AI and data science in general in mind, in terms of fundraising specifically, and looking at what's happening right now and in the immediate future, is there anything that genuinely excites you at the moment in terms of where things are heading, rather than hype? And can you give us any specific examples of that?
Sameer: I think, look, it's exciting times in general. It depends; there are two sides of the coin. One is scary, one is exciting. For me, the whole focus... I've been in the data industry for twenty-two years now, Tim. The talk about the importance of data, both from a risk and a reward perspective: initially I used to educate people on why this is important, and now the penny has dropped. So the exciting bit for me is that now everybody understands the importance of something that is such an asset. Now people have started to understand it's an asset. What they don't know is how to use that asset. Somebody said to me, the old saying is that data is the new oil, but refining oil was much easier than refining data. Now the realisation has kicked in that they need a methodology to refine that asset, which is data. So that is the most exciting thing. People have gone over that mountain of "I don't know whether data is important or not". Now the mountain is: data is definitely important, and I need to know how I can refine it to get to that end goal, or to achieve my business outcome.
Tim: Sure. And last question on AI, I promise. Will the future, with AI in mind, actually put data jobs at risk, or are data scientists and people who work in data more important than ever because of AI?
Sameer: I think it'll be a bit of both, because there are some jobs which AI can automate, such as analysis and business intelligence development; AI could make inroads into that. Some jobs, getting data into a shape that AI can do something with, such as data engineer and AI engineer, will be much more in demand, because AI needs data in a certain structured way. AI can't, at the moment, yet think and ingest, although there are technologies coming that way. But from a more analysis perspective, AI will come in more. So you will probably see fewer analyst or data admin kinds of roles, but you'll need more data engineers and AI engineers. It's kind of both. What I am certainly seeing, from where I can see, is that the chief data officer kind of role, or the head of data, whichever title, that leadership role, is becoming much more in demand, only because, again, companies need to know where to invest, how to invest, what foundation to fix. And all that comes with the leadership of, okay, what teams do I need, where do I fix, where do I focus? So again, that data leadership layer, not necessarily chief data officer, will also become much more in demand, certainly in the short to medium term.
Tim: So those roles are evolving, aren't they, I'd say.
Sameer: Yeah.
Matt: One final question for you, Sameer, before we let you go. Jumping forward four or five years into the future, into the early part of the twenty thirties, cast your mind forward. From your perspective, what would a brilliant data-driven fundraising organisation look like, and what can it do that most charities can't today?
Sameer: Quite simply, if you can connect the fundraising to your beneficiary and answer the question of how the amount of fundraising that I have got gets translated into beneficiaries, how many beneficiaries I have helped as a result of my fundraising, if they can answer that question and connect that golden thread, that'll be it for me. Because that encompasses a lot of processes, people, technology and data. If they can answer that question, we'll be in a very good place, because right now I don't know many charities who can answer that question.
Matt: I think that's a very fair reflection. A great place to leave things. Thank you so much for your time. Sameer, such a joy to talk to you as ever.
Sameer: Pleasure. It's my pleasure. Good to talk.
[42:45]
Matt: A massive thank you to Sameer for joining us. You can find out more about Sameer and his work at DataMonet. I thought that was a fascinating conversation, Tim. Any highlights jump out at you?
Tim: It's such a big conversation, data, isn't it? Because it's such a massive subject. I think it's so interesting to hear Sameer talk about the challenges being faced and the role that AI plays, and also some of those classic issues that charities across the country face in terms of dealing with old-school data whilst trying to modernise and keep up with new technology and new ways to use data, and get to grips with their CRM mess.
Tim: Yeah, yeah. I totally agree with your statement there about that always being the thing that people fall back on as the issue. I was really struck by how Sameer laid out the sequencing of attacking the problem, and how cart-before-horse it often seems to be: "We need this big new CRM to fix all of our problems." And when you hear Sameer lay it out, at best that's in fourth place, behind the business problem and working towards that. Is that the right solution? So I thought that bit was particularly fascinating.
Matt: Yeah, absolutely. And people and processes, it all comes down to people and processes. Always culture and people. People can often be the problem.
Tim: Absolutely. And of course, if you enjoyed this episode, please subscribe to Something to Think About wherever you get your podcasts. And when you get a chance, do leave us a rating or a review. And a review, not just a rating, and a review. It really does help more people to find the show.
Matt: Something to Think About is brought to you by THINK, helping charities with strategy, fundraising, insight, people and change. Thank you so much for listening.
Tim: And until next time, take care.