Aaron Phethean (00:26)
Welcome to today's show. Today we're joined by Dylan Anderson, technology strategist and data strategist, helping companies figure out how to make data fit and achieve their objectives. Today's show is packed full of strategic advice and we dive into technology and we look towards the future. So without further ado, let's dive in.

Aaron Phethean (00:48)
Hello Dylan and welcome to the show. I'm absolutely delighted to have you on and you're an author of a lot of things, you've been on podcasts before. So I feel like you're really able to relax and dive in. I'd love to hear a little bit about what kind of excites you about data to start with.

Dylan Anderson (01:06)
Yeah, sure. So yeah, by way of introduction, Dylan Anderson. So I'm a head of data strategy at Perfusion, a small consultancy in London, UK. And I think that does lend myself to why I'm excited about data. It really is such a growing and evolving space. And I actually started my career in pure business management consulting and transitioned into data and data strategy because I saw, ⁓ data and AI is the future.

And then once I kind of joined into there, I saw, oh, wow, there's so many different facets of it. It's not just machine learning models or AI models or dashboards or whatever. There's so much underneath it all, like data engineering, governance, or security, or privacy, or how you connect the business and the data world. And that's just honestly an endless task to better understand it and explain it. So it keeps me going, especially as a consultant working with a lot of clients in this space.

Aaron Phethean (01:41)
Mm-mm.

I think you've nailed it. It is kind of endless. one of the things that I find kind of most intriguing is there's not like data itself still hasn't really fit and found its way into a sort of standard part of the organization in the same way that like finance and marketing and sales sort of have. Data is in different parts in different companies, which I just find like fascinating, which means it must be just such a long way to go before we really standardize our ideas.

Dylan Anderson (02:32)
I mean, you inadvertently touched on one of the biggest blockers of data success in organizations is how it's organized and the org structure around it or the operating model around it. Because it doesn't have that home, because sometimes it's just been locked in under IT, it doesn't produce the value that it really could, even though everyone hypes it up and everything. it's, yeah.

Aaron Phethean (02:57)
Yeah, okay. Well, I think that is probably the main thing we were chatting about when we're thinking about what what can we talk about that's kind of new and interesting about, you know, all the stuff that you write about the customers you speak to. I think the the catchphrase that I recall was not only technology, we had strategy is not only about technology, and it seems kind of obvious to me because speak to a lot of people, but maybe frame it for everyone, you know, everyone sitting in a company thinking they

might be a bit too aligned to IT or they might be really succeeding but maybe just frame that. What is what is kind of a good data strategy look like?

Dylan Anderson (03:35)
Yeah, so I mean touching on what you just said, there's are lot of vendors out there who are great sales people and they're great at telling a company, hey, if you buy X brand, we'll solve all your data needs for you and all your data problems and don't worry about it. And I think as a, it comes with your other thing that you said, data doesn't live in one specific place. So you get a lot of senior stakeholders, C-suite individuals who don't.

quite understand data. So they hear that and they're like, ⁓ that sounds great. Let's do that. Let's solve it. Data is technology. It's the same thing. Whereas really what data strategy is and what it uncovers is it's basically business strategy, but with the lens of data and AI. How can you accomplish your business goals and your business needs using the data that you have in your organization?

using data tools and products that you can build, using AI and new kind of solutions that didn't exist yesterday. And how can you grow revenue? How can you increase your ⁓ customer base or improve their experience or reduce costs? All those things data and AI could enable. And I think the main purpose of a data strategy, in a small sentence or two, is to enable the organization.

Aaron Phethean (05:00)
Mm.

Dylan Anderson (05:00)
to

do what they do in a better way. And I think at the heart of that is also uncovering all the little bits that make up this data ecosystem in every company and figuring out how it all works together because data is just not one thing. It's tens or dozens of different little pieces put together in the right way. And if you don't put them together in right way, you're going to have problems.

Aaron Phethean (05:03)
Yeah.

Yeah, ⁓ there's quite a few things that I see day in, day out. If I revisit the vendor point, I don't feel uncomfortable at all. Obviously a software vendor. I feel like the very abnormal approach that I'm not there to sell my product. It's really there to figure out whether it's a fit. one of the reasons, that's how bad a teaker actually means, ethical and boring.

that we started was because I looked at the enterprise sales process and just thought that that didn't often match what people wanted as an outcome. was sort of unethical in some ways. It's just like, that's not how technology should be bought and sold. Coming back to the data though and what companies make of it. Like if you look across a company, I definitely get why senior leadership doesn't necessarily know what...

data is or does for them. I kind of like the analogy of ⁓ the gauge, you're driving your car, but you're looking at the data gauges to see what's going on. It's a representation of the world. When you start that kind of conversation with senior leadership, how do you go about explaining what it is or the way they should be thinking about it if the mindset is not quite right today?

Dylan Anderson (06:49)
Yeah, I mean, it depends on two angles. Sometimes if you're trying to kind of sell in the idea of data and the idea of AI, you kind of you do talk about kind of that gauge. You talk with the value that can be delivered. I think business value is the most overused term.

data, let's create some value out of this, like what does that mean? And it's really about getting and understanding what they want from this and be like, okay, this is how it can drive your revenues better, or this is how you can get a build your customer intimacy and experience. And then it also flows to the next side of where on each of my projects, we talk to executives because they want to see data, they want to use data better. And so part of the foundational piece of really understanding those use cases and

how data can enable the organization, you have to speak to those executives and senior leaders to be like, what do you need? What do you do on a daily basis? then working with them to figure out where data can help bridge the gap. I think the biggest problem with a lot of when people do data strategies is they ask, what do you want?

from data, like what tools and solutions do you want from data? And it's like, no, no, you can't lead the witness because that's where your expertise has to come in to be like, okay, let's figure out what you actually need and what we can provide and then find that kind of middle ground. And that is when you get people bought in. And that's when people really start to see, oh, this is why we are going to spend this much money on building this out. Because otherwise they're like, actually, you know what? I'd love that.

half a million quid to spend on a new marketing campaign, not this data platform. Like, I don't know what that.

Aaron Phethean (08:31)
Yeah, they can understand

how that might bring some value. They can't necessarily say, how is this going to bring some value?

Dylan Anderson (08:38)
Yeah,

so you really have to shift their mindset and their way of thinking to this can bring you value. You just have to see how it gives you value.

Aaron Phethean (08:47)
Hmm.

And I think we touched on earlier how you got into data. And I think it makes total sense that you come out of that point of view. I wonder, thinking about someone who's come from other directions, they might have come through finance quite close to data, or they might have come in through business strategy. If they're a technologist in a data leadership role, then looking at how to change the organization, I feel like there's some really quality advice to give to them about

had changed the way they go about things. I wonder what you would say to somebody who's perhaps finding themselves in that position for the first time, like, why can't I make a difference here? Or, you know, why are they not understanding what data could do? know, what would your advice to them be?

Dylan Anderson (09:35)
I think my biggest advice would be listen. And that listening to and making friends with business stakeholders is probably the best thing you can do because then you actually hear their issues, you hear what they're going through, and you begin to think with that mindset. I mean, how people think and how people do things in their job is a product of their environment.

A person who's stuck in the technical realm who's only talking to tech people will think from that mindset and they will explain things from that mindset and therefore you're left with them getting into the weeds about these different data points when someone who's like a senior marketing lead has no idea about that. But what you want is this and there's no and there's a constant barrier between that. So I think it is funny because a lot of organizations look for tech leaders who are like 15 plus years in Python.

Aaron Phethean (10:15)
Yeah.

Dylan Anderson (10:27)
or

whatever, but your best tech leaders probably haven't coded for another, for three to five years because they've been helping solve those problems for the organization they work at ⁓ while understanding the underlying logic and rationale behind the technologies that they use. So there's a difference between coding and actually being a technologist from a strategic perspective.

Aaron Phethean (10:28)
Yeah.

Yeah.

Yeah, I totally agree with that. And I absolutely love the advice to listen. It's really interesting how many different areas of life that could be applied to. If I think of kind of my day to day, you you're to build a successful software company, you know, it's so important to listen to your customers and what they want.

not think about how this technology works and what kind of find a problem for the technology. Like that way around just doesn't really work. So yeah, if we can all listen a bit better and really understand the challenges. And that leader who's developed the awareness of technologies then has a much better position to make the fit. I definitely get that as well. That's cool. ⁓ Coming to you and your kind of career and how you've done things.

I'm really interested to dig into what stands out as some good moments where you've done this kind of thing, where you've really made a difference or perhaps a project that feel like was technologies, it was significant in some way. I wonder if there's something that comes to mind. You're like, I absolutely nailed that connection between all the business wants and the technology delivery.

Dylan Anderson (12:04)
Yeah, think, I mean there's a few examples I have on that realm. think one where I worked with a huge global events company, like FTSE 100 kind of thing. And they didn't use data properly. It was like classic marketing thing. But then when we worked with their team,

⁓ through some workshops and through some ideas, we got them to really reimagine their business model. So wasn't even that we were coming in and recommending what to do with data. It's like, you know what? You're actually not a marketing events company anymore. You're a data company.

Because everything that you do from a marketing and events perspective is actually connecting the right people. And to do that, you need data to do that better. And if you want to succeed in this world, this is how you do it. And then the data strategy and the use cases kind of followed on from that. And it really changed their mindset. And they started investing a lot more into like their data quality, their platform, like how that came about ⁓ and training up people to think that way. And I think that one really hit home because it's, you do have

have to go in and really not take things at face value. You have to look, be like, are we really doing the right thing here? Yes, we're making a profit. Yes, we're making money, whatever. But if you really want to achieve your goals, how do we think about this and how do we think about more from a future-oriented lens? And so that was one big project that I loved from that perspective.

Aaron Phethean (13:30)
Yeah.

Yeah, to me the most exciting sounding part about that is that...

I hear quite a lot companies talking about we want to be data-led, we want to be data-first, we to be data-centric company. But what you've nailed there is, it's a business model change. It's not do the same thing and then just think about data all the time. You've changed the center of the universe and decision-making from a, just shift it to somewhere else. That could mean a whole new revenue stream. A lot of companies can monetize data directly, but don't.

example you get.

Dylan Anderson (14:12)
Exactly. I think, I think a lot of strategy companies do that part well with, ⁓ with our strategy consultancies do that part of the data strategy well, and just kind of propose these things that could potentially shift the model. But then there's the second part of it. So this was my second favorite project that was it gonna, that it was a nice segue into, ⁓ where we were working with a big global logistics company, 200 people on their data team. And they were, they were operating pretty well, but they were kind of each

different function was kind of operating in their own silos. like the technology team, the data governance team, the analytics team, all these things. And we went in and kind of brought it all together and like combined it all. Cause it's like, yes, you've got this, these transformative goals and you're changing the business model and you're becoming more, and they were also becoming more of a data company. ⁓

Aaron Phethean (15:02)
Yeah.

Dylan Anderson (15:04)
But it's like how do you measure everything up and how do you stack it all together to do that and enable that from a data and technology perspective? And that's kind of where you start to lose some of these like big, these more.

idealistic strategy first minded people because you need to have that holistic perspective of both what's the strategy mean, what does this mean for the business model, but then how, what's actually realistic and tangible from a data and technology perspective.

Aaron Phethean (15:34)
Yeah, yeah. Well, I can see how that can happen.

I can see how you get sort of focused on ⁓ the data governance outcome or the analytics outcome, and they perhaps don't join up together. I wonder on that, a technology podcast, like to talk about technology, I feel like for a minute I'd like to dwell on the technology, self-indulgent ourselves. What's kind of cool out there that's, we're in this really exciting space where companies do want data to do more.

And we're also at the forefront of a lot of the technologies. You AI is probably an obvious one people think of. There's probably 10 others I could name. I wonder if like from your perspective, what are you seeing that companies are like really able to benefit from that's like forefront of technology, know, the stuff that they might not be aware of that's there ready to be good for their business in some way.

Dylan Anderson (16:34)
Yeah, I mean, I think there's a lot out there, obviously, that is really pushing the boundaries. ⁓ But I think, and I'm not going to name specific vendors or technologies or anything.

Aaron Phethean (16:45)
Fair, yeah.

Dylan Anderson (16:47)
But what I am seeing is where companies really reap the rewards is when they approach it strategically and they know exactly why they're buying that technology and what they're going to get from it. Like I've seen, for example, I've seen a lot of failures in data catalogs because they buy them and they don't implement them properly. They don't have a strategy around them. They just expect that it will improve literacy, which it doesn't. But, ⁓ and I've seen a lot of successes around data quality platforms, right? ⁓ I think that's one thing that's exciting is

is where AI is coming around is like taking on the whole data quality journey because it's not just observability or contracts, it's the whole thing and then how do you apply AI within that journey to improve data quality. It's not the sexy AI that is driving new models or creating new chatbots, it's the unsexy but necessary AI that's...

re-engineering and helping improve your data quality so you can use it later on. And that's kind of where excites me. I mean, for me, as a more strategy-focused person, it really just excites me when I see technologies being used properly because that ends up being the biggest barrier to success with technologies is the company just buying it and then not using it well.

Aaron Phethean (18:07)
Yeah, I definitely agree with that. You know, the number of times I've seen a perfectly good technology, you know, great vendor, you know, exactly the right fit for exactly the wrong problem. Because they've bought it to try to solve one problem and it actually wasn't their main problem. And, you know, sometimes that leads to the kind of technology sprawl where it's the tool you have, so you have to use it for things that it's not perfect for and...

It's such a challenge to keep your technology aligned with the business. I guess that's part of what we see and do is that you're making a migration to our technology as easy as possible because they might have a vendor that no longer fits or a technology that no longer fits. I feel like that's one of the ways that we can make the biggest difference. There are good technologies available, but if no one can access them, then...

Dylan Anderson (19:04)
Yeah.

Aaron Phethean (19:04)
What's the point? The project is not to migrate, it's to have the outcome.

Dylan Anderson (19:10)
Exactly. I think the underlying...

solutions architecture and the technical and data architecture is really important to set that up well. I'm going back to what you said at the top of the podcast and saying like in terms of vendors being more sales focused. I think the original vendor CEOs or founders were not sales focused so they could actually explain it and see it through. But then once you grow to a certain scale, you're hiring people who have no idea how that technology actually works. They just know how to sell it.

So really like looking, going past the person who's trying to sell you and being like, okay, you have a sales, you're very good at sales, but do you know what data actually is and how it's used? Um, so I think having that extra view and making sure that you do your due diligence to, figure out why am I going to spend hundreds of thousands of pounds on this a year? Okay. For this reason, and, and make sure that it lines up with what your team.

Aaron Phethean (19:53)
Yeah.

Mm.

Dylan Anderson (20:11)
wants and can

Aaron Phethean (20:13)
Yeah,

you know that, I mean obviously as a vendor again, kind of, it always touches the nerve because when I speak to people and as the technical founder, know, technology guy by background, I can get really in touch with what they need or what the challenges are. And I definitely can see the next stage not that far away where we've got less experienced people in front of clients also needing to talk about the technology. So you've got to bring them along. ⁓

Like you actually be that level or near that level to have good conversations with prospects and actually nail it. I wonder like a risk of like challenging the kind of being a vendor. ⁓ What do you really think vendors should be doing differently? Like, know, tell me what is it that we do that could be better for the client and really help?

Dylan Anderson (21:11)
I think the biggest thing that I would love to see, and again, I'm a consultant working with a lot of clients, is being kind of realistic about what your tool should do versus what it shouldn't do. And then providing options for, ⁓ what would fit in with that? Because I think that level of distrust and why there's so many migrations happening all the time comes down to the fact of like,

the vendor being like, more and more and more, which makes sense. Like the salesperson's bonus is tied to that, right? ⁓ But what would be good is like, okay, yeah, you've gone, you've come to the limits of where we can actually deliver value. I would also recommend this tool to bring them in and do that, connect that, or you can build that yourself with open source or whatever that might be. And I think where I've seen it work well before is when, a solutions architect is brought in from a client, from a vendor side.

and they actually offer that kind of more technical know-how of how to set it up and customize it. Because every tech stack is always going to be customized right? There's always going to be legacy tools that you don't want to deal with, but will have to deal with. There's always going to be kind of unique team situations. So how do you think about that? It's interesting, I I mean, I think my role as a data strategy lead or a thinker,

Aaron Phethean (22:05)
Yeah.

Yep.

Dylan Anderson (22:33)
should exist in each technology company if they're big enough because it would be great if as a client to go to your vendor and then give you kind of that advice of like, you should structure it in this way because this will deliver this value rather than going to the vendor and being like, what should we do? And they're like, ⁓ just add some more on some more data and we'll go from there.

Aaron Phethean (22:46)
Yeah.

You know, that is, I can see that working really well. You know, the solution architect, and I guess technology people, some of the very nature of how we work and think, it has that kind of negative, perhaps, that they're looking for the technology to suit problems. But in the positive, when you talk about it from that sense, if you've got someone who, even though they work for a vendor, they have a more unbiased point of view of, in your situation,

the technology should be used this way. And you're actually having that as part of the process and getting the value out of it tons of sense. They want to see the best technology used, not necessarily just the other module or whatever's available to buy. You really like that. I wonder, moving on to the last couple of topics, I'd like to dig into the future. So we talked a little bit about your

Dylan Anderson (23:43)
No, exactly.

Aaron Phethean (23:54)
you know, experience and how you work with clients and how companies can benefit from data. And then a little bit about technology. It seems to me the kind of elephant in the room is what's next? Like what do we need or what's coming that probably raise awareness of if nothing else? What would you think people should be looking at?

Dylan Anderson (24:16)
I think, I mean obviously everyone will probably say AI and I think it's probably in that realm but I think it's in the realm in different ways. think what I mentioned before of shoring up your foundations, improving your data quality, your data governance via AI is a really good step and probably will deliver so much more value than this new chatbot that you want to build or this new document conglomerate thing that you want to build.

⁓ To go beyond that is the culture and the org structure and the operating model that you, that AI, how AI will change that, right? People are not going to be like, I know so many managers who don't manage people anymore. They might manage AI agents in the future. ⁓ Or what does that org structure look like with AI? Like how are people empowered by different AI tools and why are you building it for them? And then how do you make sure that they are able to use them as well? Right?

Aaron Phethean (25:03)
Hmm.

Yeah.

Dylan Anderson (25:14)
And just kind of, so I think the future is really about thinking about AI, but in a more foundational, how does it help enable the business perspective rather than just cool new tools and solutions.

Aaron Phethean (25:29)
Do

you know where my mind goes to with that kind of description? ⁓

A few people are talking about the challenges to CRM tools because of course, if AI is able to absorb the exhaustive operations better, then why do you need a tool to capture what happened? That's the AI disruption where I feel like that's almost what you're saying. It's data governance, the next step further is just it observes what's going on and that's already your capture determined and nailed to some degree.

does sound pretty cool. Sounds like a big part of the challenge of companies actually solved. I'm not sure the management of agents is any easier than the management of people though. That feels quite challenging.

Dylan Anderson (26:11)
Yeah.

I

mean, I just think of one major CRM company that is absolutely massive and is really leaning into their agent force, how they speak. So that's my reference to who they are. And I know so many clients who have that tool and they aren't even getting any value from the data in the first place. So like, what's the point of putting agents on top of that data at this point, especially when you actually look at the data quality in it.

Aaron Phethean (26:27)
Mm-hmm.

Dylan Anderson (26:46)
and the data quality standards, the rules, all that kind of stuff isn't set up in that tool. So you have like seven different names of a customer because they're spelled slightly differently. And so like, why don't you focus those agents on fixing what's wrong with the underlying tool and how it manages data before going off and building some value add that won't add value because it's not doing it off the right data.

Aaron Phethean (27:13)
And you must see

tons of that actually in your role in the like that like.

quality data capture, however you define it, it's enormously difficult and often like the least governed process and you know, as data professionals downstream of that, there's, know, I'm seeing people say how awful it is to be receiving that low quality data. That must be a really interesting position to be in to just observe that and try and fix it.

Dylan Anderson (27:46)
Literally every client and the biggest problem is it's not necessarily a data team's fault, but they get the fallback on it. It's the fault of not having a data governance program, of not having the standards for when people input data or when they sock data. They just let it come whichever format it's in and therefore it's poor quality. And who fixes it?

Aaron Phethean (27:48)
Yeah.

Mm-mm.

Yeah.

Dylan Anderson (28:14)
who knows and it's like that Spider-Man meme where everyone, every Spider-Man is just pointing at each other, right? No one really knows who's fixing the data and it doesn't get solved because it just gets pushed on.

Aaron Phethean (28:19)
Yeah.

And do you feel like that's a sort of positive of data products, data ownership? That feels like very much that's trying to solve that problem. And ⁓ maybe there are subtle nuances of that one bigger problem. But that feels like an important area.

Dylan Anderson (28:42)
It is, and I've seen some companies do it successfully. ⁓

It needs to be a holistic approach, right? Like I think from a technology and architecture perspective, like you need a data model that actually works for the business and actually structures it in that kind of way. I've gone into companies and they're like, don't even know. Their engineers are like, I don't know what Kimball means. And I'm like, okay, how do you not know what Kimball or Inmin are? Like, this is literally your job ⁓ because they've just done point to point for every data pipeline ever. And it's a mess. then that needs to be, those models need to be

Aaron Phethean (29:11)
Mm-mm.

Dylan Anderson (29:16)
supported by like some sort of governance strategy. Setting data owners, and I've seen it in many clients, they set data owners, but then there's just no reason for the owner to update the data. There's no vision. There's no association with value. So how do you kind of connect that governance with the technical underpinning engineering? And then hopefully in the future, as these data quality tools...

Aaron Phethean (29:31)
Yeah.

Dylan Anderson (29:41)
proliferate and get better and like there's a lot of them now and they are getting better quite quickly. How do you link that into the processes and the thinking on those two different areas to make sure to automate some of this stuff?

Aaron Phethean (29:53)
Yeah, I mean, I feel like there's tons of great advice in there. you know, the idea of using a tool that's called the data warehouse without actually like centralizing and warehousing the organization, it feels crazy, but you do like I see it all the time. I don't think of that as the mission. That's probably bringing us to the, you know, perhaps one of the last things that I'd like to hear from you is.

Dylan Anderson (30:11)
Yeah, you do.

Aaron Phethean (30:23)
If you had to give perhaps the company and the team some practical takeaways, what they should be focusing the attention on now, and I suppose I'm thinking of it in the terms of the current technologies available, the current problems that you see a lot of, and that feels like there's both sides there that you could probably have a really good view on.

Dylan Anderson (30:49)
Yeah, I think the biggest thing I'd say is approach it small and methodologically, right? And make sure that you're able to scale what you're doing in a way that makes sense. And how I usually like to approach that is, and I obviously am a big advocate for strategy, but strategically,

Look at where you need to prioritize what those biggest priority use cases are from a technology and data perspective and go there. Figure out if that works, tweak it a bit because it won't.

the first time you'll have to tweak it and then expand it as you go to new technologies or new parts of the business or what that might be. And I think companies like to take on everything. we have one, I've one client who's like, Oh, we need to create a customer data model. Can we evaluate all, all the technologies that touch customer data? I'm like, how many technologies that they're like 70, 80. I'm like, no, we'll be here forever. You need to focus on the top 10 to 15 because those are probably 95 % of the data that

Aaron Phethean (31:49)
you

Dylan Anderson (31:54)
you need and then expand from there because if you're building it for these esoteric legacy technologies it's not going to work you're going to build something that's useless or it's not going to be finished

Aaron Phethean (32:04)
Yeah, is another big one for me personally is the kind of finishing is so, so hard but enormously important.

Dylan Anderson (32:11)
Well, especially

like, I mean, I code in R. Like I know how coders think and technology deep tech people think, right? You want to make it really good. You want it to include everything. You want to encompass it at all. But like when you're thinking from an enterprise perspective and from a strategic perspective, that's impossible. So how do you think of it more modular and prioritize the right things?

Aaron Phethean (32:16)
you

Yeah.

Mm-hmm. Do you know I feel like if I think back across the things we've discussed Listen, but that feels like absolutely amazing advice I'll summarize what you've just said. It's probably walk before you try and run that feels like again amazing advice for a data team and then this idea of like finishing and Scope like modular pieces of work again like this brilliant advice

Dylan Anderson (33:05)
Yeah.

Aaron Phethean (33:06)
Dylan, really great to have you on the show. I feel like you've absolutely nailed how people should think strategically in just 30 minutes. I really appreciate it. Thanks for coming on.

Dylan Anderson (33:16)
No, thanks for having me, Aaron. It was a lot of fun.