Hi. I’m Malcolm Hawker, and this is the CDO Matters podcast, The show where I dig deep into the strategic insights, best practices, and practical recommendations that modern data leaders need to help their organizations become truly data driven.

Tune in for thought provoking discussions with data, IT, and business leaders to learn about the CDO matters that are top of mind for today’s chief data officers.

Good morning. Good afternoon. Good evening.

Wherever you are, whenever you are. I’m Malcolm Hawker. I’m the head. I’m the head of data strategy at Prophesy Software. I’m also the host of the CDO Matters podcast, and I’m still having my first cup of coffee in the morning here on a lovely Monday morning. I’m thrilled that you are joining us in whatever meeting you’re choosing to consume our content, whether that is podcast providers, whether that is through YouTube.

Thanks for checking us out. I’m thrilled to be joined today by Jason Foster. He’s the CEO of Synergir, which is a data and analytics focused consultancy based in lovely London, but with a growing presence here in the US. Jason also wrote a book called data means business, and I have to be honest. I haven’t read it, but it is absolutely positively on my list to consume. I do have a long list of books, but I I consume them rather voraciously when it comes to data in a little bit space.

Jason, happy Monday. Good to see you.

Happy Monday. Yeah. Good to see you too. I’m I’m on my fourth cup of coffee.

I try to So I’m a little bit further ahead than you.

Yeah. After afternoon, I try I try to cut it off. I found that if if I’m drinking past noon, it starts to have impacts into the evening for me. I I can’t it’s harder to to to gear down.

But four, is that is that normal?

Well, So here’s me with it being a data geek straight away. So that’s a data driven decision, being able to have coffee this late. So I’ve got a fitness tracker that you can you can, you can put in sort of habits, and it will tell you the impact of those habits. And one of the habits that I tracked was coffee and at what time I have coffee. And I can drink coffee up until much later in the afternoon than I thought I could, and it doesn’t impact on my sleep or my recovery. So there we go. I now drink coffee a bit later.

So what are the Okay. Worst this sounds dramatic because it wasn’t certainly the worst thing. I mean, we we’ve we’ve had, you know, more terrible things happen. But one of the more let’s just say interesting experiences I’ve had over the last couple of years as a rabid coffee drinker was my wife making the mistake of buying decaf coffee and not knowing and and us not knowing it.

So so so and and she drinks tea. So she’s not a coffee drinker. She’s a she’s a tea drinker. I’m the coffee drinker in the house, and she drinks for me.

And I remember one day just waking up and had this horrible headache. Look at dull, dull headache, and it and it wouldn’t go away all day. I just I just thought maybe I was a little dehydrated. I didn’t know what was going on. The next day, I’m, like, irritable.

And no like, she didn’t wanna get anywhere near me. My headache is horrible. I’m lethargic, and I’m just horrible to be around. And I don’t I don’t know what led me to do it, but I I ended up looking at a coffee.

It’s decaf. That’s just an honest mistake. Right? But detox was horrible.

I mean Cold turkey on the coffee.

It was horrible. Like, I was just I was an incredible incorrigible human being that nobody would wanna be around. I was like, oh my gosh. Wow.

So now you’re you’re intravenous, but you stop at midday.

No. I still I still I still drink a lot of it.

Funny thing is now, maybe maybe what I’ve noticed is is that the more I drink, the the more sleepy I tend to get, which maybe makes me think I’m, like, ADHD or something. I don’t know. But, because I’ve heard that people that have, there’s there’s some for certain people, coffee actually makes them tired.

More tired. Okay.

Yeah. Yeah. I don’t I don’t I don’t know. Anyway, we we came to talk about data.

Yeah. Yeah. Well, I tried to take us there, and you you took us back to call. Yeah. Yeah.

You you did. I wanted to share a story about going cold turkey at coffee because it was not a pretty time in the in the Hawker household. That’s for sure. So well, yeah.

Well, let’s let’s get into it. I I you’re you’re out there. Right? You’re in the field.

You’re talking with clients all day every day. You’re talking with CDOs.

I I’d love to dive into just kinda let’s let’s talk about AI readiness. Because I think for for anybody in any role within a data function in an organization, if you go to an industry event this year, that will be the thing that you talk that you hear about. You’ll be I I suspect you’ll be walking into the exhibit halls, and it doesn’t matter what those people are selling, whether it’s consultancy, whether it’s software. You name it. You’re gonna be hearing about AI readiness, I I think.

What do you what do you think?

Well, I suppose comp yeah. Conferences will be definitely, they they will lean on and and major on, AI. Whether it’s AI readiness, I think that’s a specific strand of thing that people need to get their heads around.

But, yeah, certainly, AI, what it, you know, software for it, automation, how you speed up, how you how you embed, the thing you need to think about it, the training you might need, the team that you might need, the culture that you’re gonna have, all all of this stuff that you need to think about, the opportunities, the use cases, all the all this kind of stuff. I think so it’s definitely conferences. I think, I think the what’s really interesting is we we talk to more and more organizations in every sector almost is quite a broad range, actually, of understanding of, buy in, of, readiness, to use your word, even even just mental readiness to to take a leap and jump in. So there there is a broad spectrum of organizations who have a broad spectrum of understanding and therefore a broad spectrum of, ability to execute when it comes to to AI. But, yeah, it’s a big it’s a big topic. It’s a big topic. Not hype either, I don’t think.

So it’s interesting. You suggested there that there are some companies that are reluctant.

My company, Prophecy, is just working currently, to publish, some research related to kind of AI readiness. Some of the things that we’re seeing out there are are this idea that that there’s reluctancy or or fear maybe or uncertainty. I I I don’t know what it is, but you suggested that in your in your last response. Are you seeing the same?

Are you seeing some some some data leaders saying, like, I don’t know about this thing? Yeah. Yeah. Just thinking AI.

Yeah. Complete completely. I mean, any to be honest, any any tech innovation, any change, full stop, you get that. Right?

You get reluctance. You get fear. You get uncertainty. You get worry. You get, you know, people thinking it’s a bubble, thinking it’s a hype.

And, obviously, if that if that leans into their fear, then they’ll then they’ll they’ll see it as a bubble more because they want you know, almost don’t wanna lean into the change. So, yeah, toe totally. People are reluctant.

I think the lion’s share is that we conversations we have is that people want to understand it. People want to get their head around it. That might result in them not doing anything. That would be totally fine if that’s right for their organization and their industry.

But I think I think, yeah, that that’s more I think where most people are want to understand it, not really sure where to start, not really sure, you know, where to go, not sure who to listen to.

But it’s it’s getting attention of of of very senior business people in organizations wanting to go, let’s let’s crack on and do this AI thing, whatever that mean ends up meaning to them.

Well, I I think it’s normal for there to be kind of early adopters and late adopters in any technology. Right? I mean, that’s that’s a that’s a very normal thing. There’s an adoption curve. Yeah. But but I get the sneaking suspicion, and maybe this is just me caught up in the hype, but I get the suspicion that this time things are a little bit different.

Not maybe necessarily from a business value perspective, but just from the perspective as of expectations of CDOs.

Right? I I’m I what I see a lot in the market is, you know, some CDOs that were caught a little flat footed with with COVID. I mean, we were all caught flat footed with COVID. Nobody could have ever anticipated that to that degree.

But a lot of companies, a lot of, of boards, and a lot of CEOs said, hey. Okay. We need to transform. Right? Like, we’re we’re we got a little caught caught a little flat footed, but our supply chains were a little more brittle than we thought they were. Our customer relationships weren’t necessarily maybe as solid as we thought they were. And, hey, let’s let’s do this digital transformation.

So two to three years ago during the middle of of of the pandemic, I I think there were expectations that were placed on CDOs. And I think a lot of them, weren’t ready for that. And I’m seeing a lot of the same now when it comes to AI when there’s there’s also kinda maybe this pushback and and a little bit of that uncertainty where there’s a lot of expectations and, some of the people may be expressing some of that uncertainty or wanting to just wait and see how things unfold. Maybe they shouldn’t be. I don’t know. I’m just spitballing it. What do you think?

Yeah. Well, I don’t think this is you mentioned there that this is something that’s new. You know, this AI isn’t new. AI has been around for a very long time, and certainly in the last ten years, lots of AI, you know, technology has been built by vendors and by in house teams to to to learn and and look at patterns and and automate things and and, you know, think about personalization engines and, you know, improving supply chain efficiency.

All this stuff has been based on models. You know? Pharmaceutical industries have been using it for understanding the human body, biology, drug, application, all this kind of thing. So AI AI isn’t new.

I think the thing that is new is generative AI, and the reason it’s got the hype and the the attention is because of how, pervasive it has gone so quickly.

So and I think those two things are quite different actually in terms of in terms of how they manifest themselves in organizations and the particularly when it comes to the AI versus or as well as generative AI.

AI versus or as well as generative AI. So, yes, I think, so so that’s one side of that is is actually this isn’t some of this isn’t new. Some of this isn’t. We need to understand the various merits of all those things and how we get our heads around them and what we do about it. I think this idea of being ready, though, is interesting because, you know, what what does being ready mean?

Be being ready is a being ready is a is in the current world is that you’re you’re ready for what?

You’re ready for change. You’re ready for iteration. You’re ready for, you know, pandemics. You’re ready for economic uncertainty.

You’re ready for new presidents. You’re all these things that that happen around you. And and, actually, I think I think being in an organization that’s adaptable and agile is the most important thing because there’s always something that’s coming coming along, and you always need to be ready for the next you know, nothing’s nothing stays still. Some change is slow and and generational, and some change is quite quick, but you’re you know, you need to be keeping on your toes, and not resting on your laurels.

So I think there’s a sort of general level of readiness required. But when it comes specifically to kind of data and and applications that you need to build and data products and infrastructure, and having the the rules and routines and methods and operating model and the right people, all these things that you need in place to be good at data, there’s never a good time to get ready. You it’s you know, the the opportunity is great whether AI exists or not. The opportunity is great in terms of your business and and your, strategy and your operations.

So it’s always I think it’s always time to make sure you’ve got your foot on the gas with what it is you need to have in place to be good at at using data to guide decisions. AI is then an application on top that helps you go even faster and and do even more things.

It’s such a good question. Like, what does that even mean? Right? Like, us technologists, we like to put everything into these kind of deterministic sets, and it and it’s like, okay. If you do these things, these five things, the project is done. Right?

And then we can launch the project.

Right? But the idea that you could do five things and be I AI ready, but only halfway for the through the fourth, you’re not ready. Right? But then so magically, you hit the the fifth thing and then, okay.

Now I’m all all of a sudden we’re ready. And I I guess that’s your point, which is is that there there is no end Yeah. Here. Right?

There’s there’s there’s no, okay. Launch. We’re done. All of a sudden, now we’re AI ready.

What you’re saying is what you need to do as a layer leader is to build a change resistant.

I don’t know if that’s the right phrase, but change change ready or change resistant or the ability to absorb change. That’s what you should be aiming for, Whether it’s a pandemic or whether it’s a new president or whether it is AI, that’s what you should be aiming for. That’s what I’m hearing you say. Yeah.

Yeah. Yeah. The the the, I did a presentation at a conference a a few years back, and and in in the opening section of that, I talked about the people that were ready for the sort of twenty, so the early two thousands, economic challenges we’re able to better trade through those, trade through that if they already had good insight and good app good handle on data and able to kinda look at the future and look at forecasts and look at what came before and look at their their customer base, they could trade better through that. In the UK and Europe, people that were best, set for when the GDPR guidelines came new guidelines came in were were better able to manage themselves through that change.

In fact, it wasn’t change for some people that got that got ready early on. So they they were best they were best suited to be able to to get through that. When the pandemic hit, the people that had already got good platforms, good procedures, good data leadership, good analytics, good insight in place were able to to to look at what the situation and trade better through through COVID, you know, then then this AI things have come along. And there’s always gonna be these new things.

So as I say, it’s never really a good time to get ready. It’s just I think it’s something that there’s an acceptance, most of the data they have access to, internal and external, to help make better decisions about the operations, about the strategy, about where you’re going, about to help make better decisions about the operations, about the strategy, about where you’re going, about customers, about all these things within your organization. You need to do that anyway. Something new comes along, then you’re able to kind of build on top of it.

This idea of iteratively improving, I think, makes you constantly ready for the next thing, and I think that’s what people should be focused on doing now.

Yeah. Or learning by doing. Right? Right. Like, just just do it. Right? If I don’t think anybody could be fully AI ready, at least, like, right now.

With the pace of change in technology, I think you could make a valid case to say nobody is ever kind of fully ready, but the best way that you can try to keep pace is potentially just learn by doing. Find some use cases that that can be supported with with Gen AI.

Yeah.

Yet I’m still hearing a lot of people say, well, I’m not sure we’re we’re ready. So, you know, I think there’s a lot of opportunities for people like you and companies like yours or even people like me to say, hey. Nobody’s ready, but you gotta go. Right?

If if if there are valuable use cases out there thing to do.

Yeah.

Yeah. Don’t don’t do it just for doing its sake. I mean, we’ve learned our lesson here maybe the the hard way, but if you think there’s value, and I can pretty much guarantee you that there will be, I mean Yeah. The the the number of use cases is is potentially endless.

I think I think it’s I think AI, particularly generative AI, is is big enough, transformationally enough that just answering the question of how impactful could or would this be to our organization, our industry before doing anything is really valuable. Because that helps you determine whether you need to do something now, later, quickly, slowly, you know, not at all, go all in. Do you know what I mean? I you’ve gotta you’ve gotta kinda go.

It’s worth it’s worth, having a proactive understanding of the response your organization should have, then you can decide what direction to take. So it’s not about diving in and going, we have to do something now in terms of, like, building some generative AI solutions or embedding AI into organization because, because, really, their technology is to solve a problem. We first need to understand what the problem is, not not just at a kind of grassroots level, but as a, you know, is there an existential problem in our industry of of that we’re gonna be hit by if we don’t do something? Right.

We better that’s an organization that needs to move move quickly. An organization that’s less, you know, less impacted by that, you know, fine. Then then then do something a bit slower or or don’t do anything at all or, you know, use it as a Copilot tool, but not something that’s embedded in the organization. You know what I mean?

I think that helps shape your thinking about where you go and what you do. But you’ve got to answer that question. It’s big enough to answer the question. It’s like when the Internet came out, you had to answer the question of what does this mean, and and some organizations didn’t do that.

And, you know, there’s lots of lots of organizations that don’t exist anymore because they they didn’t jump on that train quick enough. And this is that this is that again.

Yeah. So what I’m hearing you say, to paraphrase, if I may, yes. There potentially could be value by by jumping in, but you also need to have some idea of where this aligns from an overall strategy perspective. Yeah.

Yeah. Don’t just drive blind blind. Don’t just just go in for the sake of going, but have some understanding, maybe the bigger picture stuff. Where does this does it align to your strength?

Will impact here to you and your industry and your company and strategic alignment to it? So it’s both. Yeah. It’s not just go now, and it’s not just wait and figure it out.

It’s probably figure it out and go.

Yeah. Yeah. Yeah. Yeah. No.

Yeah. Yeah. No. No. Look. I I I always think that we we, the industry, tries to kinda simplify this stuff down so so much that it’s, you know, binary.

Yes or no, do or don’t. But there’s it’s new it’s really nuanced, and and I don’t think there’s a there’s a black and white answer of yes or no in this. I think there’s a there’s a, for me, there’s a very clear answer that it’s worth taking some time to assess the impact now or in the future and then make a call about what you do. That’s just a really pragmatic, sensible way I think of dealing with innovation.

Because you can bury your head in the sand and nothing will happen, or you can bury your head in the sand and be caught, you know, in a year’s time when, you know, a new entrant comes into the market that has, you know, AI ified. Is that a word? I just made that up. The whole industry.

And if that if that happens, you know, that’s happened in industries where whole industries have been have been digitized.

But, you know, actually, I’m gonna contradict myself. If you look at something like Airbnb, you know, they they changed the nature of of staying away from home. And everyone thought that’ll be the end of hotels, but hotel businesses are booming. Absolutely booming.

It’s not it it’s an it’s now an and. It’s not a instead of. And so so I think there’s a real kinda, like, necessity to understand where this could go in your industry and what could happen and then and then have a response. It’s it’s Right.

Right.

Well, it’s interesting.

You had you kind of hinted at this maybe in in something you previously said from a Gen AI perspective, from the perspective of AI being operationalized via chatbots Yeah.

Are you seeing companies embrace that through the application side of the house and less on the data side of the house? Are you are you seeing companies saying, okay. Hey. You know, CTO, you you are responsible because you are gonna bake that into whatever.

You’re gonna bake it into our customer facing application. You’re gonna bake it into our CRM. You’re gonna bake it in somewhere. But this isn’t a CTO thing.

This is is an app thing.

Are you are you seeing that?

Well, if we look at what, I think so generative AI and particularly large language models, I think that’s where we’re we’re talking. And that’s I think that’s more than chatbot, but large language models, they you know, you you would be mad to build one of those.

I know some organizations that have spent a hell of a lot of money now trying to build one and and I’m not going to do that.

You you’d be mad to build your own. Right? Yeah. Yeah. Do we agree with that?

Oh oh oh, yeah. Yeah. No. With without a doubt. I mean, at the beginning of all of this, I used used to hear people say, okay.

I can’t use my data to train an LLM. Yeah. You’re not you’re you’re you’re not going to. The first version of Bard had seven billion parameters.

Yeah. And it and it and it took six thousand GPUs running for twelve weeks straight. So not a lot of people have six thousand GPUs just kicking around.

The moat the moat is too big and deep already for anyone to to build it. So so what does that mean? Looks like really you know, some people have tried to build search engines, but, really, you know, it’s hard to fend off the the main players. Right? It’s the same here now. So if if if we essentially have the really good ones, and I and I’m not and I’m not saying that that there won’t be any others, but it takes us deep, deep pockets and and a lot of time to be able to get good at this stuff.

It’s gonna become table stakes. It’s gonna become something that everyone has access to. Like, you can just goo search for something on Google or whatever you use. So if we’re gonna use OpenAI or or one of the others, then that’s the one we use. So so that’s not the competitive advantage. The tool, the technology is not the combative advantage.

Integrating it into what you do and how you work is the combative advantage.

And it’s a technology, and it’s all really, where I think this sort starts to go is it it’s about the experience, the process, the interaction that you have with it, and how you embed that into what you do. So if you’re embedding LLMs into your, you know, day to day workflow, that’s something that might be embedded into your video conferencing, your emails, your calendar. You know, that’s all about integration. It’s about, technology. It’s about, customer experience and user experience. Those aren’t the things that tend to sit with the chief data officer.

Right.

They might they might do, which is why I’m stopping short of saying it shouldn’t be with them, but they tend not to be. They’d be with digital leaders, transformation leaders, technology leaders, architects, those sorts of people that are that know how to embed technology, but even, you know, experienced designers that know how to build kind of structures that people can interact with digitally. So it’s a I think it’s a different paradigm. Where it comes back into CDOs though is where’s the differentiator?

The differentiator might be in the experience that you have with interacting with the with the large language model, or it might be in the data itself that you add and you add to it. So data all of a sudden and and packaged datasets all of a sudden, not even necessarily all of a sudden because they all they probably should have been before, but all of a sudden become potentially a USP when worked alongside a off the shelf LLM. So I I think there’s a kind of you get two two houses. I think the let’s call it the traditional AI, so the nongenerative AI, that that certainly still probably does still sit with a with a good CDO, and someone who who’s building artificial intelligence solutions inside an organization.

Cool. That is right that that sits with whoever currently is the data leader. That’s correct. But so so I think I think it depends on your organization, but and, but I think this is different.

I think generative AI is different.

Yeah. Totally agree. So, you know, what I’ve been talking about for nearly about a year now is that for now, I I think most companies will operationalize LLMs just as you described. Right?

They’ll be they’ll be Copilots embedded in wherever. Right? Maybe that’s a custom Copilot in something you build or your customer facing application. Maybe it’s an off the shelf Copilot that that is being used in one of the Microsoft products or whatever it is.

Yeah. But the difference there will be the data that you feed into prompt. Right? In the beginning of all of this, we were talking about, you know, complex prompt engineering.

And I was like, okay. Wait wait a minute. Hold on. This this seem this seems like hand rolling code here.

I’m not so sure we all need to be prompt engineers. And it turns out, I don’t think we do.

But the latest version of GPT can take three hundred pages of text in a prompt. Right? You could put a lot of information in a vector database that could actually be consumed in a prompt to say, here is a known fact set. Here are all the the interactions we’ve had with customer ABC over the last year. Now give me a summary about what I should do next with customer ABC or what customer ABC might like to consume next as a product. You could be asking that of an LLM.

The data that you’re feeding into the prompt, if it’s good quality, if it’s accurate, if it’s consistent, if it’s trustworthy, you’re gonna get good responses back. If it’s not, you’re not. So it’s the same paradigm we’ve always faced, I think.

Yeah. Yeah. Good quality datasets that are feeding applications and interactions that people have. You know?

Wow. Wow. That sounds that sounds quite that sounds quite dull, doesn’t it? That’s all we need to do here.

Well, but it’s still it’s interesting, though.

Yeah.

It’s It’s interesting though because I’m still hearing people say, I can’t use my data to train an LLM.

And and, you know, as as as much as I’m out there saying, the p in GPT stands for pretrained.

You’re you’re not training in LLM.

If you’re using an open source LLM, you might be doing something called fine tuning it, which is which is basically a series of questions and answers that you could kind of fine tune the model to lean a certain way. Maybe lean towards a procurement use use case or lean towards a customer service use case or something. But, you know, training an LLM to your point, I mean, that that’s a massive mode. And we’re already seeing, you know, a lot of commoditization there. Most of the the vendors are are just competing on speed and and accuracy.

So I think there’s learning there’s learning to be had here for a lot of data leaders. And and whether it’s it’s FUD, if you’re in doubt or or or whether it’s it’s just a a plain old knowledge gap. I think that that’s a a big part of becoming more ready. Yeah. And I always say just to learn about this stuff.

Yeah. I think I think so. Yeah. And and I get and as always, it gets very complicated very quickly.

We’ve just went into a bit of technical depth there, although not very far, and and even that can be quite complicated for people that aren’t familiar with it. Right? And and I think the the important bit, again, just raising it back up to it is an organization ready. It’s knowing where where to play with this stuff. Where where should we play? And and, you know, you can be purist and go, well, let’s think about the business strategy and and the problems we’ve got and work backwards and only think about problems that, you know, LLMs and generative AI can solve.

But they’re purists, and, actually, you might miss some opportunities, some innovation because, actually, they’re so different, and there’s so many new things that this generative AI thing can do that actually also makes sense to think about in the other direction and meet in the middle where new innovation and problem that we have can be solved in these new ways that we haven’t thought of. That that’s, you know, that’s design thinking. That’s, you know, business analysis. That’s that’s process engineering.

That’s understanding how does the business actually work? And these are all, like, powerful tools that businesses need to have in their minds and and then, you know, actually do. And if data leaders can have though are that way inclined, if data leaders can think that way, I think they can have a really strong, you know, play in this space and and a really take a real strong leadership in in helping to change how an organization thinks about what it it’s able to do and grab this and grab this opportunity, which I I do think it’s really really big and really really strong.

So be the chief digital officer, be the chief transformation officer all at once. Roll all those things altogether. Be be that unicorn.

Maybe. Maybe. Maybe. I think that’s you know, probably with those titles, that makes them it makes it too big and a bit too totally on roles that possibly exist.

But but there but, you know, there’s no I don’t think anymore, and maybe this there never was. I don’t think there’s very neat, you know, squares or circles around roles. There’s always this kind of, like, overlap. And and I think the overlaps the overlap side are really helpful if you’re an organization that is able to collaborate very well.

And the overlaps are a problem when you can’t when you don’t collaborate and you just fight with who’s responsible and try and pull control from each other. But yeah. But I know I know the principle of what you mean is you gotta think a bit digital. You gotta think a bit transformation.

You gotta think innovation. You gotta think data. You gotta think commercial. You gotta think finance, or at least bring the people together that have that experience and and take leadership in in in that respect.

You know, I you said design thinking. I was like, in in in my mind, I’m saying, yes. Yes. Yes.

We need more of that. Mhmm. Because when I think design thinking, when I think, you know, be innovative, take some risks, understand customers with, you know, really, really deep levels of in intimacy, what do they want, what do they not want, to me, that sounds like product. Right?

Like, that seem that that to me, that sounds like an exceptional product leader. Doesn’t mean that CDOs need to become product people and product people need to become CDOs per se. Maybe there could be opportunities there, but Mhmm. Makes me start thinking about products.

Yep.

So what are you hearing in the space? What are you hearing out there related to data products? And does that align to the idea of design thinking? Does that does do do what people what most people are calling data products even really kind of align to that world, or is it something else?

I think, well, conceptually, it’s all all quite similar. I mean, we we’ve been talking to customers about, sort of data products and data product management more specifically, for a number of years. And getting getting there was almost like is is almost like trying to create, you know, an organization that it that uses agile in its, you know, deepest form, the methodology agile, not being agile. So they’re quite it’s quite a it’s quite a transformational thing that needs to happen to become fully agile.

In the same way, it’s quite a transformational thing to become fully kind of product led with how you think about building solutions to solve problems. And I and I, yeah, I I think, you know, when my example before about where we were talking about generative AI and and, you know, if it if that’s the the tool that you bring in and now it’s about embedding it, We need to think about it from a product perspective. Like, what what what’s problem does that solve? What metrics can we use to test whether it’s been work whether it’s working?

You How do we keep improving it? Who’s gonna look after it? Who’s gonna make sure it’s doing what it’s supposed to do? Who’s gonna make sure it’s iterating and improving and we’re doing new releases and new versions of it so so that it’s getting better and solving bigger problems than it did before.

I think if we can package up, solutions that we build with that mentality of of product in the same way we think about physical products like a phone or, you know, an ironing ironing board or a cup or a whatever you buy, then then we can then we can think about AI solutions and actually even any data solution in the in the same way. It’s gotta serve solve a problem. And if if we build around solving that problem, then then we know it that we’re that it’s supposed to be doing something. It’s a me it’s got a means to an end, not just something that exists for the sake of existing.

What you just described, data product management. Exactly. Love it. Like, love it. Applying product management as a discipline to data management. All those things you just said, like life cycle management. Right?

Understanding customer needs. What problem does this solve? What value does it bring? Maybe even quantifying that.

So Yeah. A lot of I’ve been in a number of interesting dialogues on LinkedIn recently, related to data valuation. Where do you sit on this? Is is is the idea of valuing data is is is this just a fool’s errand?

Is it like, you know, the Sasquatch and the Yeti and the Loch Ness all locked, you know, all as one, or is it something that you can actually do?

I think there’s kinda two two angles to that actually. One is the value that applying data to something has. So if I apply data to better understanding customers, then I can, you know, target them more effectively or make more of the right products that they need. And that adds value.

That makes them spend some more money. So there’s an incremental gain by applying data to something. That you can put a value on, and I think we should. And and I don’t think we should do much with with data generally where we’re not thinking about that that kind of, like, upside, that impact.

It might be it might be cost saving for automating something. Of course, it might be reducing cost, but I think that for me is monetizing data. That is putting a value on the application of data rather than the data itself, but the application of data. And and there’s loads that organizations could do and to do with that and definitely should.

Then I think there’s the value of the data itself.

And, and for some for some companies, I don’t think that is relevant, actually. I think at some companies, the application of data is what makes money or or saves money. I think if you’re in an industry where you are sitting on a a a pot of data that doesn’t exist elsewhere and other people could get value from it, great. Let’s put a value on it, and and let’s try and sell it in the in the broader set in the broader sense. And that might not be actually charged for it. It might be about creating more sticky relationships with your suppliers so they they renew their contract.

So I think you can put a value on data if it’s a unique data step. So just a list of customers that we have is not unique because I because customers shop in lots of places, so I don’t think that’s necessarily necessarily unique. But if you’re in an industry where there’s only two or three players and you’ve got, you know, thirty, forty, fifty percent of the market share, that’s a really interesting dataset that that people will be that people will be interested in. Or you’ve or if you’ve added an added value to that data in some other ways by applying other things about about customers, for example, like, you know, like Facebook does and turn that into an advertising engine. So I think you can monetize data by how it can be applied. You can monetize data by if it’s unique.

They also think you can monetize data through sticking out as talk about regenerative AI, where adding data to a certain situation multiplies the effect. And and I think that’s where people should start looking at their own data saying, how can we use this data to to add value to a to a business challenge somewhere?

Yeah. The idea that you you can’t valuetate data, to me, just doesn’t pass the smell test because we all know like, intuitively all know that if the data was better, the business processes would be smoother and we would make more money. Right? Like, that’s just it’s just it’s just intuitive.

It it it it’s logical to to think that better data will optimize processes, which will yield business results.

We like to get kind of wrapped around the axle around thinking about, like, the intrinsic value of data. If a if a if a field of data is sitting, you know, late into, you know, in a database somewhere just collecting digital dust, does it have intrinsic value? That that’s where I think we we as data people love to fall down rabbit holes and and, you know, to start talking about, you know, generally accepted accounting principles and those types of things. But to to me, day data has to have value and it has to be able to drive value. Because if it doesn’t, I think I’d have a hard time justifying my job.

Yeah. As I said, but I think it’s the you’re you’re right. Is it maybe it’s a splitting hair or nuance point here, but it’s about the application, the use of the data, or or the assurance that it’s that it’s good. So, particularly in the data management field of data, which is obviously one strand, poor data management causes problems somewhere.

You know, processes break or, you know, customer interactions fall over, caused by data problems. So fixing that it doesn’t mean that data has value per se, but it does mean that that block that that resolving a problem stops you from losing money or stops you from having brand reputation damage or stops you from some risk that you have. So I think the application or the misapplication of data is what has value. And then for some companies, packaging up data into little sets that you can do something with then also has some additional value.

I’m attracted to the idea of kind of running a data function as if it were a p and l. Like, not charging for everything. Right? Not putting a price tag on everything per se, but but running the organization as if you were making. I mean, you’d you’d made reference before about, like, products. Maybe you’re making an ironing board.

If you had to make a good iron if you if you were making them, you had to make them good. You’d have to meet a customer need. You have to do it in a way that was cost effective and cost efficient.

So I kind of conceptually like the idea of running a data function that way even if your CFO doesn’t think the data has any value. What what do you think when you hear that?

I think most, organizations don’t have the operating model that supports actually actually running in that way, as in physically having a team that has a p and l in in that way. And often, you know, it’s phrased wooden dollars. Often often because it’s internal movement of money. It’s not really any no one’s really paying or or losing money in that in that respect. But, yeah, of course, there’s costs. But but so conceptually, though, there are costs associated with running any any function, even if it’s just one person.

And in large organizations, the data teams can be hundreds, of course, if not more. And there are and and those costs have to they have to go somewhere. They have to deliver some value because otherwise, what’s the point of having those costs? And and some people are better at having rigor around that.

And the cost if that cost, produces results, it’s making something. It’s either it’s either washing its own face, it’s it’s getting enough back that that covers the cost, or it’s making more. So, yeah, in that respect, that is profit and loss. That is making money or not making money.

And and and a I think a good, data function with a with a strong commercially minded data leader who wants to have a long fruitful career will be thinking about both the sides of that equation, the the value that it can deliver and the cost. And and I think you have to I personally think as a as a CDO, you have to earn the right to spend the company’s money in some way. You have to earn the right to get investment to do more things. And and sometimes you get you can get lucky and and someone can go, yeah.

I get it. A leap of faith. Here’s some cash to go and do something. But that that does run out.

So I think I think the mentality of p and l, the mentality of let’s spend some money to make some money, let’s make let’s make let’s spend some more money once we’ve made some money, which is how you run a business. You can only spend it if you’ve got it. And you either got it get it by borrowing it, getting investment, or making money and and reinvesting. I love I love that concept, and and it’s not and I think I think mentality wise, that’s something all data leaders should be thinking about, and it’s only what we try and instill in in customers.

Yeah. I think it is really about a mentality at the end of the day. Right? It it it’s it’s about that p and l mentality.

Right? It’s it’s about driving value and thinking, hey. My job entirely depends on this. Yeah.

And and in in my experience in having worked with hundreds of CDOs, largely through my time at Gartner, I could kind of sense the mentality. Right? And you you’ve used the word commercially minded. Right?

You could use you could use a phrase to say kind of P and L centric. You could say maybe business centric.

But that mindset that’s associated with with that type of leader versus the leader that sees himself more as a cost center and and as a pure cost center, as more of just an operating function.

I could find I I could remember kind of palpable differences between those two mindsets, and I think I think it’s it’s interesting. Because if I were to paraphrase everything you’ve been saying for the last forty minutes or so, what I’m hearing you say over and over and over again in different forms is get closer to the business.

Right? I’m I’m hearing you say get closer to the business. Understand the business. Understand the cost drivers.

Understand business processes.

How does AI get operationalized? What does your operating model look like to best drive outcomes for the business? Yeah. Which is often I hear often I hear the exact opposite, believe it or not.

I I hear what you’re saying is kinda shift right. Mhmm. Right? And what I hear often is more of a shift left.

Like, get closer to the data. Get closer to the business applications. Get closer to the source.

And I don’t think it’s an either or, but if I had to pick, I know which one I’d be leaning towards.

I suppose, again, some of there’s there’s how does that physically manifest itself? I think either of those. You you you do need to do to do both. I I think even if even if, you’re a support function to other organ other parts of the organization, which I think is sort of more the shifting towards data, then the you know, then you you’re you’re raising up a level, and the organization needs to understand the return and cost associated with with doing data.

And but, yeah, I I just yeah. I I kind of I kind of feel that, a chief data officer that doesn’t care about, and I’m and I don’t think I’ve seen any that doesn’t care about upside and value and but but there’s certainly some that act and behave in ways that that don’t lean them closer towards delivering delivering value. But it but it does come back it does come back to the sort of operating model of of how an organization sees data. And this isn’t just down to the CDO, frankly. This isn’t just about the CDO going, I’m either a cost center or I’m a profit center or I sit near a data or I sit near commercial. This is about a conscious, choice by an organization about how best to set up and how best to get value, and how do we work and collaborate as an organization to ensure that we’re that we’re solving for the right things.

And and I but I do think there’s an obligation on the CDO to to help the company that they work for to get to that mind that sort of mind space. Yeah. Because we gotta we gotta think about it in that way.

What you’re describing is is a change change agent.

Right? What you’re what you’re dis what you’re describing is somebody who can lead an organization through change, whether that’s changing a mindset, whether it’s changing a process or an operating model. Yeah. That’s another kind of recurring thing that I think I’m hearing here, which is digitally transforming okay. That sounds great. But, really, what we’re talking about is building a change resistant organization.

Really, what we’re talking about with becoming AI ready is building a change resistant organization. To summarize everything I’ve been hearing you say, that that really seems to be job number one.

Now how you do that? Yeah.

We we didn’t book a second hour. I mean, but but but but how you actually do that, you know, whether, you know, you mentioned the adoption, of agile methodologies. We’ve talked earlier about experimentation and and trying and earning by doing. I think there’s a lot of interesting conversations we could have about how. But maybe that’s a great last question to ask here as as we as we wind down our time.

I’m a I’m a I’m I’m your customer, and and I’ve engaged you. Okay. I I I I need to do some of these things. I need to become more change resistant. What what are the the one or two or three things that you would counsel your customers to focus on to to do that?

Yeah. I would, says a few.

So I would And and I I understand.

To to answer this in, like, in three or four minutes is probably you know, it’s it’s this is this is a little unfair of me to but but but still give it a whirl.

I think it’s a good question. I I think there’s a few things. So there’s, start small and iterate.

So there’s you don’t need to go all the way and solve all the problems straight away. We need to find we need to find something to start with and test, learn, prove value, establish the right agenda, and then iterate from there. So start small and iterate.

I would focus laser focus on business problems, business outcomes that the organization needs to solve. I wouldn’t start with trying to sort out data. I think through the problem that you’re trying to solve, we get to what data we need to fix and clean and put in the right place in order to solve that problem.

And I would focus on, collaboration and communication. So very high communication, very high collaboration across the organization.

I think the combination of those things builds you credibility. It builds trust.

It gets you known. It gets you, it gets the awareness your awareness up about where the problems are in the business, and it focuses you not just on strategy and and long term, but also execution and solving problems today.

Brilliant. That’s a great way to end the conversation. If you were interested in hearing more, please check out Jason. Your LinkedIn you’re you’re prolific on LinkedIn. Your company, Synergir dot com. Correct?

That’s right. Yep.

Alright. Please check them out. Jason, thank you so much for carving an hour out of your busy Monday to chat with us. I really appreciate it.

I’ve loved it. Thank you. And, yeah, appreciate you having me on the show. Thanks a couple of weeks.

Great. Well, to all of our listeners, thank you. Please consider subscribing to the podcast if you haven’t already. I really, really appreciate your engagement.

We are building an act active community here through CDO Matters, whether that is engagement with me on LinkedIn or whether that’s engagement through our podcast or through our blogs or through our live events, the third Friday of every month. If you’ve got questions about data, if you need to build a data strategy, if you’re struggling with with getting a governance program off the ground, come check us out. We’re sharing insights all day every day through multiple channels. My goal here is to help extend the tenures of CDOs and everybody that wants to be a CEO.

Thank Thank you for checking us out. We’ll see you on another episode of CDO Matters sometimes very soon. Thanks all. Thanks, Jason.

Thanks.