Transcript
Speaker: If you're framing things as like customer buys products, products require services, we're then reframing how we're thinking about the actual delivery of data to our customers. If we're going out going like, hey, what do you want? People are gonna go like, I want the moon on a stick, please give it to me. And you're like, what problem does that solve?
Speaker: and it doesn't solve any problems. And that's not like, again, like the ticket trap of like asking people what they want rather than actually going out and finding what problems they have. The best way to list the requirements is just to ask people what problems they have. Hey, what keeps you up at night? What stops you from sleeping?
Speaker: Today's episode is brought to you by Omni. Most companies I speak to want AI analytics but are failing to put projects into production. That's where Omni is different. It's the semantic brain that grounds AI in the heart of your business logic, giving you governed answers, whilst also the depth to identify root causes. It's intelligence everywhere, from your spreadsheets to their chat feature, even within your product. Omni moves you beyond the dashboard. Don't just take my word for it. Trust teams like Perplexity and Synthesia that are already using Omni to deliver intelligence that people trust.
Speaker: Check out them in the show notes or visit omni.co. That's O-M-N-I.co. Now, back to the show. Hello everyone, welcome to another episode of the Stats Data Podcast. I'm your host, Harry Gollop, and today we're getting into data products, what they actually are, why but also why so many of them fail,
Speaker: I think data teams spend a lot of time building and not necessarily a lot of time focused in on adoption and making sure what they've built is really having impact. um Today, I'm joined by East Oldfield. Hi, everyone.
Speaker: ha The CDO from Advancing Analytics and yeah, we're going to get into, suppose, some of East's strategies and playbooks to how he makes sure he drives real impact with the data products that his teams build. So yeah, East, welcome to the show.
Speaker: Cool. Thanks, Harry. Really excited to be here. it's I love talking about like the stuff we're doing and helping like people change how they're thinking with things as well. and So thank you for inviting me on.
Speaker: No, no worries at all. I'm looking forward to And yeah, I think it's one of the biggest challenges that we see um is the ability to talk about impact. um Whether you're a data leader, an analyst, scientist, engineer, you know, the the importance of being able to talk about impact has never been so prolific in the industry and or for businesses. So yeah, excited to hear some of your your strategies around this. But um I've given you a little bit of an intro. and is yeah who Who are you? It would be great to to give a bit of context to the audience about yourself, your sort of career and your role today.
Speaker: Cool. um so yeah so i started my data journey a long time ago, about 15 years ago where I did I guess, a master's in economic history, which I did after graduating into the financial crisis.
Speaker: So not a great time to find the job, but a great time to carry on with their education. um And that taught me a lot of nice skills like how to use up the Python and SPSS and how to do data analytics.
Speaker: And I found myself like a job in a small marketing company doing things like propensity modeling and and um and other things.
Speaker: Long so long ago, i can't even remember. um But from there, I then went into like the music industry and spent a lot of ah for about three years and looking at like recommendation engines, um both for like, uh, uh, as well as like the A&R teams to go and find like new artists, uh, to go and sign. Um, using lots of like disparate information, like whether that's from my consumer surveys all the way through to like Spotify. And this is again, like early days of Spotify. So, and also the early days of big data.
Speaker: And it's just a lot of fun, um, to have an impact on what the business was doing and decisions it was making. um But more importantly, um it was just a fun industry to be in. There's a lot of like you got um face time with artists and taking them through like what audiences think of them and how they can then change their like profile if they want to.
Speaker: um But most of the they didn't. And and Then I moved into consulting, learned a lot of technical abilities, like how to do um like ETL and BI and all that, like what you now call like data engineering.
Speaker: But back in the day just BI. i and then had a little break from consulting where I headed up the the data function for Chanel, luxury retailer, um and then moved back into consulting five and a half years ago at Advanced Analytics.
Speaker: Started off like principal consultant and then head of analytics and about three months ago or so moved into MunYu new role of CDO, which is not, I guess it's not a traditional CDO role um because I have very little responsibility for data at Advanced Analytics. I'm more like the field CDO where I have a lot of responsibility for our customers and their data strategies and how they're best leveraging data.
Speaker: um and helping them drive um a lot of their data propositions forward. Excellent. Fascinating career. I think it's always great to hear so many different journeys to how people get to where they they've got to.
Speaker: Getting into the way to the meat of the episode then, East, it's all going to be around how do we sort of bridge that gap to really driving being impact with with data products. and But I think first off, teams are always told and and really focusing on building the right platform. You need good good foundations um in in place, but there still seems to be this gap from building a platform to building products that actually drive impact. Is that something that you see? why did you think that Where do you think that gap comes from?
Speaker: Yeah, I mean, like most ah most data teams are deeply technical people and they want to build like good things for themselves to use.
Speaker: And there's nothing wrong with that. like If you have a good platform, it helps make your life easier. But a good platform is only a very small part of the journey, but it's also the one that takes up most of people's time and attention.
Speaker: um So there'll be things like, hey, let's just focus on how this Terraform works to make sure that we're able to keep redeploying again and again and again. it's like, well, do you really need to do that when you've just got like a storage account and a delivery cluster? And that is that's your platform. like Why do you need to keep going around and around and around to make sure that's a rep repeatable process if it's just like a one-off thing or it's a very easy thing to do like DR? there's
Speaker: I mean, with most like architectural decisions, it becomes a trade off. What are you trading off um for like not having like automation as an example for a platform? um But there's also like confusion within the wider industry. like and A few years ago, like data mesh was like a big thing.
Speaker: It's no longer a big thing, mainly because everyone jumped on the hype wagon and realized that there was a load of like organizational change that needed to happen in order to make it a success.
Speaker: um But where people kind of got like focused on was the the technology to make a federated data platform, um but also things like data products and like data products from a um from a data mesh definition just has way too many attributes, way too many things that define it.
Speaker: um Whereas if you look at like the wider technology and software industry, a product is four things. It needs to be valuable.
Speaker: So it needs to be able to solve like real problems. It needs to be usable so that customers want to keep coming back. It needs to be feasible with the technology and skills that you have.
Speaker: It also needs to be viable. So it needs to be like worth the effort, worth the cost in making it. It's also not going to open you up to any like regulatory mit issues or legal issues.
Speaker: So have these four things and those those four things which like we as a company have focused on as like this is the definition of a data product. why the product is A data product is a product that just uses data rather than it's a software product, but it's a thing that hack is yeah has data as the focus point.
Speaker: And is on that, i think that's really good to unpick a bit more. and you knew That definition of them four sort of variables that should be make something a data product. How, like if you're a you know leader, a practitioner, how do you actually define them? What's your process for understanding where a product will sit on them them them scales? And how do you actually build the definition of what a data product looks like when you're talking about an actual product?
Speaker: Yeah, i mean, so there is and we have like we have a framework which we call Sunbeam. There is if you go and have a look at the American League website, there are links to it. But but like It's a framework in the sense that it helps guide us, but it's not necessarily like a one size fits all methodology.
Speaker: um Because everything is going to change depending on i what what your company itself is doing, um which industry it's in and what are the most pressing needs for it.
Speaker: um So if i look at, I guess, any company strategy, If I'm going to be super crude about it, it's going to boil down to one of three things or maybe like a mixture of the three things.
Speaker: I need to increase revenue. I need to cut costs. I need to manage risk.
Speaker: And a company strategy will be a mixture of those things. That's how companies operate. this That's your mini MBA done in like one minute.
Speaker: um But for the most pressing things, it's there like, hey, I'm in a regulated industry industry. I need to be able to do like ri regulatory reporting because if I don't, I'm going to get fined.
Speaker: Like that in itself means that you have like the basis of product because it's valuable. um
Speaker: Then you then scope that down into, okay, so I know which area I've got like value for. um who are my customers for this?
Speaker: Are they internal or are they external? and If they're internal, it's going to like, if we're going with the regulatory reporting element, it could be the risk function or it could be the regulator themselves.
Speaker: If it's the regulator themselves, then it's like, well, okay, what are your specs? Like you're going to have to have, like for a regulator, there are going to be things which are going to be common in standards because that is how they're going to apply the same regulation across thousands of organizations. um sorry once you've got you your customer, then you know what their needs are.
Speaker: Therefore you've got usable. then you need to then understand like what data needs come for it, come from different sources that you might have access to in order to build this product.
Speaker: And that comes onto your viable like your vi like what is the effort for this does it outweigh the risks like does it outweigh the the potential like profit or like cost factors so if um i have like need to do like regulatory reporting and if i don't do it it might fight i might get fines like a million a million pounds every year for not doing it but it might cost me 10 million pounds to build this thing
Speaker: Not going to work. it might cost me like a hundred thousand. It's like cool. Easy peasy. It's paid for itself. um um And then from a feasibility perspective, it's like, okay, do we have access to this data that's required? Are there things which we can like, like filling gaps with, with things? If we're missing data, are we going to get in trouble, like other trouble if we do, if we, um if we bring through this data? So for example, again, like,
Speaker: If you're in a regulated industry, you might not want to bring through things like customer information or PANS, but you might be obliged to based upon regulation.
Speaker: Like, I know your customer right as a regulation means that you have to know exactly who your customer is and what they're, what they are doing and how they are linked to potentially other customers inside your organisation as well.
Speaker: Really interesting. And these frameworks, East, do you find that they they can add noise to the environment because I think you know you're framing there what a data product should be and many folks are trying to move over to a data as a product model but you is this just the flavor of the month as you said or year you know we have data mesh where data contracts and there's always something new Why did you draw the line of understanding what you should be working, what frameworks you should maybe try to apply and and which ones maybe are just noise?
Speaker: Yeah, it's a good question. um like there There are aspects of any framework which provides noise. um and You as a practitioner just have to ask yourself, like what is the immediate problem in front of me and how do I go about solving that?
Speaker: um And then you can borrow from any framework that you have. um But I guess typically most data teams don't operate in going going out and looking for problems to solve. They just take orders from ticker queues.
Speaker: And that is what ends up killing data teams. it's like It's the IT model that doesn't work. and It doesn't really work for IT model, like for IT service either.
Speaker: um But it's how how we got into into being and how we've just adapted to the world. Mainly because data typically falls under IT t and IT t just goes, I can't deal with these things. I'm just going to pop in the ticket queue and we'll just go from there.
Speaker: What's your advice? Because I hear that a lot. You know, one of the things when people are looking to move is that I don't want to be, don't just be sitting churning out dashboards or just offering these tickets. um What's your advice? Again, maybe it's slightly different for leaders and for practitioners, but if they're in that type of environment, they want to start making change. How how would you go about doing that?
Speaker: So the first thing is, it's more like a mentality shift. So whenever I hear, here um they need to change the mentality because especially you are data team who only has internal customers, your users are basically a captive audience, but they're also not a captive audience because they have agency. They can like go and spin up Excel and do their analytics outside of your platform, outside of your service.
Speaker: It's more that the yes the weight of the organization um sort of either there like with a carrot encourages them to use the platform or from a stick forces them to use the platform, um whether they whether they want to or not. um So the first thing is just reframing them as customers.
Speaker: customers will have problems that need to be solved and for that problem to be solved you need a product and that product um needs I guess a needs to be valuable to them it has might if you think about like maturity models it might have an economic cost to um to the customer.
Speaker: Although like when it comes to internal customers and internal products, like these costs are pretty much um it's like paper money, like monopoly money. It doesn't actually have any impact on anyone's my proper balance sheet. and But it's again, if you're framing things as like customer buys products, products require services,
Speaker: We're then reframing how we're thinking about the actual delivery of data to our customers. We're delivering them by products and we're providing in service.
Speaker: So we're kind of... I guess it's like, there let's empathize with our customers. understand where they are, what they require. It's not even like what they require, what problems they have.
Speaker: Like if we're going out going like, hey, what do you want? People are going like, I want the moon on a stick, please give it to me. And you're like, that's what problems that solve. And it doesn't solve any problems. And that's go to like, again, like the ticket trap of like asking people what they want rather than actually going out and finding what problems they have.
Speaker: um like If I kind of take like another like industry term, might we've got the forward deployed internet engineer, something that Palantir came up with and it pretty much describes like most most advanced consulting models of going out and finding problems and solving those problems.
Speaker: rather than just churning through tickets. Because the tickets, like it's useful if people got bugs for the products that you've created, but they're also not a great way to elicit requirements.
Speaker: The best way to elicit requirements is just ask people what problems they have. Hey, what keeps you up at night? Yeah, i feel subsidy from savingping yeah i think- It's a common theme, right? It's about really really understanding your your stakeholders and not being scared just to to sit down with them, right? What are your problems? And and really you know not taking stuff on on face value. One of the approaches that when we've spoken previously, east that I think is a really really poignant, particularly with the industry where we are at with, with I'm going to say the word ai Yes.
Speaker: um but it's around rapid sort of prototyping before you're you locking down requirements and i'm sure we can get on to how how important that is with or how intertwined that is with partnering with your your stakeholder but yeah what what's your view on on sort of this um prototyping type uh approach before for building why is it important for from your standpoint Yeah, so I guess like from a traditional data perspective, you might kind of treat prototyping as like report wireframing.
Speaker: It's just that with AI, you're able to go one set further and and deliver a like a working prototype in the sense that it's using real data.
Speaker: It's fully functional, it's just not got the hardened edges around it. So it's got the functionality, but it needs to still be hardened up. that It needs to have the engineering applied to it so it can just be deployed and be scalable and supportable rather than just something that just have.
Speaker: um as your vibe code. I hate the term vibe coding because it's not. like You're still applying engineering principles to so what you are creating.
Speaker: It's just that you're using Genitob AI to produce it. So by taking on board that accountability for what is being output, you are being the engineer still.
Speaker: um But my AI is so capable. might you can Within a couple of prompts, you can have pretty much any like have a working data bricks pipeline, can have like working dashboards in a single prompt, you can have singles with mine complex so so solutions and systems in maybe like a dozen prompts.
Speaker: But that means that we are able to get to something tangible in front of the customer super quickly and super easily. um So that when you are like talking to them about their problems, you can then put very quickly go like, hey, does this help solve your problems?
Speaker: And they can go, yes, no, well, how about if we tweak this? And then you're starting to go into the like that instant feedback cycle where um you're able to iterate quickly or you're send you also you also know very quickly and very early on whether something is going to be valuable or not.
Speaker: So we spend probably too much time just creating things that people don't use but if you start with hey customer what problem do you have um Go away, do a quick prototype, come back with them. Hey, does this solve the problem that you have?
Speaker: Oh, it does. Great. I can now move on and make that into a proper product or no, it doesn't. Fine. I've wasted maybe a day or two or even less sometimes. Maybe I've wasted like an hour and you've got that feedback and then go, okay, that didn't solve your problem.
Speaker: Like, where did it go wrong? my Did i misunderstand the problem? like Is the problem no longer actually a problem that needs to be solved? like Where do we go from here?
Speaker: I think it's CI, CD, not of your pipelines, but of your requirement gathering, um right? Which allows you to quickly iterate, pivot, make changes before you've invested that time. Exactly. What are the risks, East? Have you seen any any sort of, I suppose, challenges yeah or risks of where you people can build things really quickly where that's maybe something they should be aware of?
Speaker: yeah Yes. um So the first one is that they start doing development in or like the prototyping in an environment in which it's just not like it's not secure, it's not sanctioned, especially if you're using like live and real data.
Speaker: Like, do you really want your your organizational data going out to Replit or OpenAI or wherever? um So there are like decisions based, and and again, these are more organizational ones, like how do people use AI safely?
Speaker: How do they use it within am a recognized harness where the data is still retained by the organization, it's like still within the organizational boundaries without them being leaking here than everywhere.
Speaker: um There are also risks where and people get way too attached to what they've created because it gives them power which they might not have had before.
Speaker: by Going from like report dev to creating like very shiny UIs and systems is is very different.
Speaker: but it's then like, hey, I've got this power now. I'm gonna go crazy with the power and treat everything as like, hey, it needs a UI, it needs a UI, it needs a UI, and it's like, I'm just gonna have similar problems of like report and dashboard live proliferation with without but then it's now an app it's an app it's an app it's an app um so again lots of these things just need to be thought about um more appropriately before then making the decision about what actually is going to be solving this problem and where it's going to be solving the problem uh
Speaker: I saw a few weeks ago, like an approach to AI, which was like slow thinking. And it's doing that all that thinking process about what we're trying to achieve before then doing it. It's like, think slow, move fast.
Speaker: think that's a good motto to have. um Bruno, who joined us on the podcast recently from from Monzo, said that yeah that's one of the powers of AI, you know that execution um layer is is being handled for you, which allows you to step back and to think more deeply about how how something is being architected or yeah is this driving the right things and asking them questions because before you needed ship stuff, you you people needed to get their heads down and and start building. So whilst we can build quickly, I think yeah it it gives you now the ability, you need to sort of restrain yourself um to to really ask the the right questions first. um
Speaker: East, wall I think one of the other things people might struggle with is when when you are prototyping, when is what he's good enough for for for what you've built and when do you need to kill it and throw it in the bin and and start again? And yeah when when is it, okay, this is more than a prototype, we need to productionize this and this needs to scale. So how how would you, what's your sort of frameworks to sort of, or or approach to breaking down that?
Speaker: Yeah, I mean the the easy one is when to kill it and that's basically when a customer says I don't like it or it doesn't work for me. um when it The harder ones are knowing when to stop prototyping and start productionising and it's usually when you're getting um
Speaker: signs like, oh, could we tweak that? Could we tweak that? Or could that be a green? um That's when you need to stop and start productionizing or putting in like the CICD, the source control, the robust pipelines that support the delivery of data, um any infrastructure that's required to host whatever you've prototyped.
Speaker: um Maybe even also so ripping out some of there the the code that AI has written. likes to be verbose. It's not always and the right thing to do.
Speaker: and But it's a like from a prototyping perspective, it's a means to an end. doesn't have to be perfect and you will have to do lots of refactoring. um but that's when you then have, that's when you then like stop, but you still maintain the prototype in the sense that it's it's available for the users to use and make decisions from, but it's not necessarily in a fully productionized state, but you're also not making any more changes to that prototype.
Speaker: Excellent. And on that, we've now sort of covered yeah partnering closely with your stakeholder, really understanding and challenging what it is they're asking for. You've worked on this sort of CICD process with requirement gathering and prototyping.
Speaker: um You've got to a stage now where you've got something which the the your your customer is nitpicking at and is is there and you've no you've confirmed it's or you're you're probably starting to see that it's driving value for them.
Speaker: How do you continue to drive that adoption? Because I think that's where most data professionals tend to draw the line and write, it's built, um chuck it over the fence and it's done. and I'm to go and move on to the the next stakeholder, the next team, the next project. But I think, you the work should only just be be starting. So yeah, what's your views on on how you look at um really sort driving adoption, measuring success and having, adding sustained value?
Speaker: Yeah, so I guess with anything product related, you then need to market it. You basically need to sell it to to other people. um so Selling it to one person is great, but the only way you're going to make is a success of it is selling it to lots of people.
Speaker: um So it's things like doing like little road shows around your organization to demonstrate the value in which you've brought to a customer. Hey, this thing also solves these problems. And then people will be like, oh, that will solve my problem. Or actually, i think it might solve my problem if I use it this way.
Speaker: People are actually very good at thinking about i guess the lateral impacts of a product and how they can utilize it even if that isn't always the primary intention of a product.
Speaker: um So it is... like market it, make sure that um people all right are basically aware of it um and you're driving hype around it.
Speaker: Then there's also like the long tail of enablement and onboarding. and So like having things like dropping sessions and embedded training and embedded documentation and things like that where people know where to go.
Speaker: if they need help, whether that's like, hey, watch your video, hey, look at these documents, hey, go and talk to talk to the team. um There is um a good set of I guess what because I guess at this point we're now kind of shifted from like product delivery to service delivery and there is a very good set of definitions of what a service is from the government digital service so GDS and there's about like 10 of them and
Speaker: They all describe what makes something really good to make sure that people want to keep keep coming back to use it. um Things like setting expectations. So from a data perspective, it's that, hey, this product gets refreshed daily um and it contains these things and it will always contain these things.
Speaker: But it's also signposting people to um so that they can get human help when needed. um and a load of other things but it's at that point when you are kind of gone in from like dev mode and production mode to like service mode where you are then delivering a service to your customers and the mentality then is like oh cool I'm just gonna go back into my ticket queue like no stop no ticket queues like you need to help people serve themselves and at the very last resort ticket queue it's the how do you make sure that
Speaker: um people are continually continuously able to make and decisions from the product that you provided. How um how are they continuously able to know what new features might be delivered to your product? um You might call it like, hey, let's deliver a roadmap.
Speaker: um Again, it's making that publicly available to your customers. rather than again, just sitting on a DevOps board somewhere or a GitHub project or whatever you're using, just making it part of the product itself.
Speaker: Like lots of, um I guess, digital tool native companies do this really well because they have customers already. So it's shifting from like, hey, I've got my external customers for my software product, I'm now going to shift that same mentality to serving my internal customers for data products. It's the same approach.
Speaker: It's just different audience. Really interesting. So i would you say that's your, I mean, it's obviously borrowing from other industries, right? Products that are and maybe even going and sitting down with your products team and and asking for some mentorship guidance. i Look, this is the these are some of the data products we're looking to roll out. of What is your what are your approaches? How could we do this internally could be a a valuable thing, especially if you're in a product driven organization. Okay.
Speaker: When, um what do you do when you, how do you keep this nice and clean and tidy then? Tech debt is hard. There will inevitably be things that, products that that that do decay over time and their usage fails. How do you keep that, um how do you understand when that's happening ah and how do you close that gap with stakeholders and products and they say when to to build something new as I suppose a a rounding off?
Speaker: Yeah, I mean that's that is like really difficult. like again Again, if you do work inside a products led organisation, go ask your product team, they already have like plans for that. um But there are things like having observability inside your platform and your products to know who's using it, how they're using it,
Speaker: So that when you can see things like usage drop off, then you know that you need to start having a conversation around deprecation. um And what does paths off the product look like for those customers who are still using it?
Speaker: Is it still valuable to them? Is it still viable to you? um Is it still feasible to run? like Having like one customer who it's super valuable to and doesn't cost a lot to run is probably still worth keeping.
Speaker: um But if it's not that valuable or it's super expensive to maintain, then you need to start having difficult conversations. And no one wants to hear the words, we're going to kill this thing.
Speaker: but No one wants to hear that. like the Not the people who built it and not the people who use it. um Like you know, probably first hand when you've been using something and cu and the companies come out and go like, win we know we no longer support this.
Speaker: And you're like, oh, I still use it. It's like if you've got a first generation Kindle as an example, like no longer supported, still use it, but it's note you can no longer use, you can no longer get books from it. And like, well, great, thanks Jeff.
Speaker: I think um although we've had a lady called McKinley back on season one from the Financial Times who you spoke about their BI sort of migration and adoption. One of the challenges they had was typical, you know dashboard bloat.
Speaker: think they had about 1,200 dashboards across the organization. um They turned off about 800 of them. um and They got two or three people that said, where's my dashboard gone?
Speaker: Yeah, screen test. Which um I think reiterates your point of who's using it um and setting up you know from the technical side, have you got observability, have you got usage metrics? um Because you can quite quickly, so you might not even have to have any conversations before before you turn it turn it off.
Speaker: um Is, we're almost at the end of the episode. I suppose rounding off, what was your um what's your advice for um folks listening that I suppose want to want to be closer to their stakeholders and want to drive impact? Like we've spoke about a lot of points here. what are your sort closing sort of key takeaways for for a practitioner to really improve in that space?
Speaker: Yeah, so I guess the first thing is to like read a book. and So we talked a lot about like product and what makes a product and prototyping. These things all come from a book called Inspired by Marty Kagan.
Speaker: So again, it's not new. There's nothing new underneath the sun anymore, especially with like the rise of AI. You can regurgitate things pretty easily. um And then when we talk about like service design, there is a book called Good Services by Lou Down.
Speaker: And that's then how like so we've got our product, we've got our product, we're now in service mode. So the two books go hand in hand, although reading them and looking at them you're like, how?
Speaker: And it's basically what we've just been listening to for the last 35 minutes or so. And in the great theme of like being a great product person is experiment, try what works. What I am talking about, mike now might not resonate with with how your organization works. You just need to try what works for you, try what works for your customers um and see what sticks and keep going from there. is
Speaker: There is no one single approach to things um other than like make sure you've that you've got customers, make sure you're doing products, make sure you're delivering services. But other than those three things, like how you do about that is entirely up to you.
Speaker: why do What do you think the biggest challenge facing data professionals over the next year is is going to be? And maybe, you know, it's big question, but like on ah on a macro level.
Speaker: On the macro level, um i mean, the fact that AI has come ah come around and will continually disrupt how we do things. um If you look at like the Data and AI Summit, so Databricks flagship event that was only a month ago, AI features heavily and everything which they've announced as features will change how we do things.
Speaker: um Not always for the better. and but generally to make lives easier. um There is also so the economic impacts of AI as well, in the sense that you can do more with fewer people.
Speaker: um So there's going to be changes to how teams want to and need to organize.
Speaker: Again, another book to read is Team Topologies, if people want to go and go down there get down to the library. But that, again, is like, how do you like structure to deliver value for customers in a product organization?
Speaker: um There is also the fact that AI, because you can do more with AI, do you need to do more with less or do you just do more with more people? um Like if you look at IKEA, like they introduced AI to things and they like hired more people because they've spotted more work. But I think like most of these things are currently like looking like outliers.
Speaker: um Just need to keep an open mind um with how the industry is going. like AI does change a lot, but does it change everything?
Speaker: We'll see. lovely place to leave it east um it it's been a pleasure chatting through approaches to um i suppose defining data products rapid prototyping how we partner closely with stakeholders and and i think the big takeaway is um don't just throw your product over the fence and you have to sit closely with your your stakeholder from inception through to adoption and beyond and and yeah, being being close and understanding the the the pain points and the um and what's going to really drive impact is is really going to help differentiate you as a as as a data person and a data team. So thank you so much for thank for coming on.
Speaker: Excellent. um Well, that's it for this week, folks. We'll see you in a couple of weeks. Thank you and goodbye. Thanks all.
Speaker: if you've enjoyed today's episode ah be so grateful if you could hit that follow button leave us a rating even better after to show on to a friend who might also get some insight from it really helps us grow the community and continue to share amazing conversations I also wanted to take a minute to talk to you about Cognify.
Speaker: Those of you that don't know, Cognify is the leading recruitment partner for modern data teams. We help some of the world's best organizations scale data and drive real value from the buyers that they make.
Speaker: you're thinking about building a team or making a hire and you're struggling with talent or just want some insights on the market, then I'd love to jump on a call with you and tell you a bit more. Equally, if you're looking for a job and want to find your next dream role, then reach out to myself or any other Cognify team.
Speaker: We'd be happy to see if there's anything on our books that we can help you with and give you general advice on the industry. Finally, big thank you to Omni, this season's sponsor. If you'd like to learn more about the AI analytics that Omni can deliver you, then check out the link in the show notes or come speak to me. i can happily point you in the right direction.
Speaker: Again, thanks for listening and look forward to seeing you a few weeks time.

