Subscribe
Share
Share
Embed
A show to explore all data matters.
From small to big, every company on the market, irrespective of their industry, is a data merchant. How they choose to keep, interrogate and understand their data is now mission critical. 30 years of spaghetti-tech, data tech debt, or rapid growth challenges are the reality in most companies.
Join Aaron Phethean, veteran intrapreneur-come-entrepreneur with hundreds of lived examples of wins and losses in the data space, as he embarques on a journey of discovering what matters most in data nowadays by speaking to other technologists and business leaders who tackle their own data challenges every day.
Learn from their mistakes and be inspired by their stories of how they've made their data make sense and work for them.
This podcast brought to you by Matatika - "Unlock the Insights in your Data"
Welcome to today's show. Joe Wright from CitySprint is gonna tell us all about the realities of a legacy migration project. They're big. They're messy. And I think the key takeaway from our conversation and the thing that I got such a buzz out of is that data is people and data.
Aaron:You know, you can't just do the technology. So let's get into it. Hi, Joe.
Aaron:Hi Aaron. Thank you. So we've, worked together for a while now. This is a great opportunity to talk about what we've been doing and your vision at CitySprint. But maybe a good idea if you just start with who on earth are CitySprint. What do they do?
Joe:Sure. Yeah. CitySprint are, on the largest same day courier providers in the UK. Very strong footprint, in and around London, which is where they started. I don't operate across the whole of the UK.
Joe:I have a a UK wide network. Same day logistics is is literally predominantly picking up parcels from one location and taking it straight to another location. So we've got a fleet of couriers who use all manner of transport around London. They they they use a lot of bikes, for example, because you can move around London quite quickly on a bicycle. And they differ to, sort of a hub and spoke operation in some words, but they will literally take the parcel from one place to another, from a to b rather than taking it back to a hub.
Aaron:Yeah.
Joe:Handling it, trunking it somewhere else, and then delivering it then the final mile. That's what we do the we do
Aaron:Yeah. And it strikes you, they're quite a cool company, actually. You did a, bike ride all over the UK last year, from depot to depot kind of thing, didn't you?
Joe:We did. Yeah. Yeah. It was, quite a team of us, involved in the end. So we took one of our cargo bikes, which is a a push bike, an ebike, with an extended front fork, if you like, for for carrying parcels.
Joe:They're heavy, I found out. And, we basically took that bike right around the entire network, as you say, from site to site, right up from Perth, I think. There's those Aberdeen at least in Scotland right down to, Southwest.
Aaron:Yeah. So Basically.
Joe:There's a team of people that that that they're second chance to take the bike a little a little way of the gym now, and we had support riders all the way along as well. So some leader would be on the cargo bike and some would be on another another bike. But Yeah. It was a brilliant experience.
Aaron:So, as much as I love bikes, this show is not meant to be about bikes. It's probably meant to be, all good time to get to the data, and the and the people involved. What what does it look like inside CitySprint? You know, how does the how does the company interact with data? What do they expect of you?
Aaron:What what what does what does the kind of operation look like? I suppose it would would help people understand and give them a little bit of context.
Joe:So the operation is it the the business is structured to me, almost run locally, operationally. So each site is responsible for organizing its own workload. They're all cooked up centrally to, a management system, as you might expect. We then take the data from the management system and make it available for analytics in in in my function, in the BI function. So we're servicing the sales team.
Joe:We're servicing the operations team, so we are able to play back to them their their performance stats. Yeah. And we service, the, the management team as well by providing analysis. So because we because we're custodians who've got a lot of data coming out of that system and other supporting system systems like the CRM, we have access to a lot of data in one place, and we can therefore hang it together and and and learn a doc Analysis as as in Yeah.
Aaron:And, of course, being a modern, fun growing company, that management system is, like, super modern and, like, just brand new, isn't it? It's like, you know, it's the it's the latest, piece of software.
Joe:Sure. Yeah. If if you mean newest by 12, 13, 14 years old, it'll still range from internal. But but the company is doing a lot to, modernize that. So it's the same story, isn't it?
Joe:What why why fix it if they ain't broken? So the the the main management system has been in place for a long time, but it works, and it works really, really well. It's it's quite heavily customized for the operational needs, and we have the opportunity to continue customizing it as the business evolves. So there is a school of thought that says, as I say, if any broke, I'd fix it. However, we've we've sort of recognized that to to remain competitive, as everybody else moves to the cloud and brings the latest architecture and new innovations in IT in, we we definitely don't wanna fall behind that curve.
Joe:So there's a big IT project to to modernize the whole system at the moment.
Aaron:And I I suppose that the question was tongue and cheek, of course. And I think this is really, you know, important, you know, reality for everyone listening. It's an old system. It's been there a long time. It's got a lot of history embedded in it.
Aaron:And the data that, you know, we deal with and, you know, we're we're working on a project together as part of that modernization. It's, you know, it's, it's difficult. It's got a lot of things that has just built up over time. And, you know, the amount of things that we're bringing together, you know, that that's the reality for everyone else. There's just a lot of, you know, legacy and and and technical debt that builds up over time.
Aaron:Maybe interesting to tell everyone about the project from our perspective. So, obviously, we're working together replacing some of the data legacy with with Matateka. Why did you spot that was necessary? What what was life like kind of sort of before?
Joe:Sure. Okay. So we have, as I said, been pulling data out of of these these systems, and we've also, as a a BI function because of our position in the business, we are able to bridge gaps where the systems have them, which I don't think is is uncommon because it's oftentimes easier to to write something in your data warehouse to bridge a gap, in logic than it is to develop a system to to to meet that need. So over time, there's sort of some, as you say, some legacy technical debt built up both in terms of working with the restrictions of the system and building business logic to to meet the needs of the business as it's evolved over time more quickly than the software has. Mhmm.
Joe:What that led to is, I find about they call it a house of cards of of of of a of a data platform, because we've had, as you can imagine, and I'm sure this is is the same for a lot of a lot of people. People kind of go. They all have their own way of doing things. So what we've ended up with is a sort of sedimentary, BI stack where there's some logic, and then someone's added a bit of logic on top and so on and so forth. And over time that becomes quite complicated to manage.
Joe:The opportunity that was spotted, with working with you guys was to take all of that history, all that legacy, actually work out what's really relevant for the business today because a lot of it is just nuance, that's accumulated over time. And to create a much more streamlined, VI stack, with just the logic that's relevant in as few steps as possible to deliver what the business needs today. But on top of that, to build in the maturity around that stack that that prevents us sort of having the same problem build up again in the future.
Aaron:Yeah. Yeah. And I think probably, you know, an uncommon problem that, I guess, people are dealing with in a in a similar situation to ours, this legacy infrastructure, this legacy, stack is working to some extent from the outside world. So the management team, the operations teams, they they ask you for data, and they ask you to do some analysis, and they and they get an answer. So from that point of view, it's kinda working.
Aaron:Whereas, you know, the reality on the other side, you know, for you and your team, it's it's hard. Like, it's really difficult, and it's becoming increasingly difficult was was part of what we observed when when we started the project. So I think, you know, one of the things I saw us doing for the business was bringing, okay, obviously, you know, infrastructure cleanup or, you know, kind of a a better system to work on. And therefore, you know, people free up their time. They then become more effective.
Aaron:How from your point of view, how do you go about communicating the benefits of that to the business? Because from their point of view, it was working, and then there's this project to make it better. How do they see the benefits? How do you think they might see the benefits in a in a project of this scale?
Joe:You have to find something, the tangible benefit that they can latch onto is my experience. So just rebuilding the platform, with the same outputs is a tough sell because, well, because it might make my life easier, but they're, you know, that's not, that's not the aim of my dollar to make my life easier. It is it is so long as it supports the business. But just for the sake of it is hot. To get buy in from stakeholders in the business, they have to have a tangible benefit.
Joe:And for us, one of the tangible benefits was, moving towards one version of truth. So one of the challenges that we've always struggled with in BI is being able to deliver a set of numbers that is always the same no matter which report you look at it. And because there's always legacy logic in a stack, it tends to generate forks in that beta where there's one view written for one set of people in a slightly different view with slightly different requirements written for another set. And all of a sudden they're looking at reports and when they compare, they don't match. So story is all this based time, isn't it?
Aaron:Yeah. And it's and it's the one that damages the data relationship the most.
Joe:I They analyze credibility entirely, doesn't it?
Aaron:The whole thing's wrong. Yeah.
Joe:Yeah. Exactly. They look at one number, and then none of them match. And then all of a sudden the whole of it's wrong. It often not the case.
Joe:It's often not actually a particularly big issue that that that drove the difference in the number. But, you know, 99 percent it's the devil's in the detail. You can get to 99% relatively easy, and then getting it that last one percent is always the the the hardest part, isn't it? Mhmm. So that was one of the the benefits.
Joe:The other one was a fundamental change in our ability to report in smaller time windows, which is a real that was a a real boost in in selling it because that's a business requirement. And, actually, in order to deliver that business requirement, we can argue very legitimately that we need to overhaul the way it works and review and and redesign some Yeah. Logic there.
Aaron:And and I recall we we talked a lot with, you know, some of the stakeholders about the kind of house and cards or the spaghetti or the the stack and and and what it was. And some some listeners will be technical people. So what are the kinds of technologies that CitySprint had as legacy? And I even go sort of a bit further back in time to my first start because there was my recollection to, BI visualization tools, multiple ETL type tools involved. A you know, it already consolidated on Snowflake as a as a step forward.
Aaron:So maybe just give everyone a little overview of what is this tech stack or what's the tech stack?
Joe:Yeah. So it's been quite a journey, really, even to where we are. So when I joined the business, we had, as you say, we hit Snowflake in play. We also had, an on prem SQL Server SQL Server 2012, I think, for Windows 2012 running c so 2008 or something like that. And so as you can imagine, that created a whole host of challenges to try and keep the 2 data warehouses, in sync.
Joe:And, actually, they weren't in sync because they received data on different schedules from from our main system. So there was a window of about 2 hours a day when they actually match and they'd be asking. One one of those data warehouses was funding, QlikView at one point. The other one was funding and and SSRS, and the other one was funding Looker. So you've then got multiples of variations in each of those systems, not to mention the 2 systems are fundamentally running on different data anyway.
Joe:Mhmm. So, we so so same during my tenure here. We've we've shut down the the SQL Server, which is more complicated than it would perhaps sound because the way it'd been structured was that we were receiving data from our management system into the SQL Server and then replicating it onto Snowflake. Mhmm. So you would then dependent on dependent on the SQL server to process as part of the pipeline for snowflake.
Joe:Yeah. So you have to rewrite the whole process for getting the data out of the system, out of the management system in snowflake. It wound up Qlik as part of that. We wound up SSRS as part of that. And we migrated from Looker to Power BI because, basically, Looker was a little bit left field for us.
Joe:The business tends to run most of its services through Microsoft. Mhmm. And then almost randomly, but certainly from a strategic point of view, it would appear random that you had Looker, which is up to your Google company, providing our visual platform. So so, yeah, we we moved to, you know, to Power BI, and and now we're with your help, you know, rewriting the whole stack, consolidating everything, using Snowflake and Power BI to to to deliver it.
Aaron:And, of course, yeah, we're we're in the process of replacing Matillion and building DBT models and adding tests and data quality reporting and a whole ton of awesome stuff. So you can see a nice, clean infrastructure to build on. And probably, one of the things we talk about is that's then able to power AI and new things. Is there any excitement about AI in the business and what the potential might be? It seems to come up all the time, almost everywhere.
Aaron:I don't wanna AI is about the word everywhere,
Joe:isn't it? Perspective is. Probably in the biggest marketing phrase of of the last 2 or 3 years, isn't it? Everything's got AI in and out. There is excitement in business.
Joe:Yeah. I Aside from what we're doing in in BI, IT are using AI to improve the the career onboarding process, for example. So they're using, image recognition models to validate private licenses so that we don't have to go through the manual checks.
Aaron:Sure.
Joe:Yeah. Private license, that kind of stuff. But, from our side, we're we're quite keen to support our our customer service functions by using AI to model, sentiment from customer transactions, so cases and and and calls,
Aaron:that sort of thing. Because now technology to land in the BI team? Are you expecting that to to come down to, you know, your, scope of work when when that analysis is required, or is that is that elsewhere?
Joe:Ultimately, I think it should fall under the eye. Yeah. And it's my idea to have it included under the eye. Organization, within the organization, there's there's been a slightly fractured approach to some of the data processes. So BI have been central around the the central management systems data, but the CRM was implemented by the former commercial director, and that was sort of her baby.
Joe:So she would look after she'd look after commercials, the sales team, the customer services team, had an Salesforce to support all of that. And Yeah. And as such, it was a little bit out of our scope. But, actually, because we've got feeds into Salesforce where a lot of that data is captured, the the customer service and customer sentiment type data, that's captured in there. We have feeds into it.
Joe:It makes a lot of sense to bring that data into Snowflake alongside the management system data so that you can look for trends between not just customer sentiment, but the jobs that they're doing. And it's I mean, there's it it is really exciting as it it gets me excited because I love the idea that you've got this massive knowledge base waiting to be tapped. Yeah. Yeah. Which I I think a lot of perhaps a lot of businesses struggle.
Joe:They've got data. I think most businesses got more data than they ought to do with. And because it's not organized and available in in a really Yeah. In a really make up the benefits of it. Way, they never get to it.
Aaron:Yeah. And the, the the CRM has actually become the customer master now effectively.
Joe:I mean, you know, that
Aaron:that's sort of an important contact to where the business looks after, you know, who the customer is and and what what's then attached to them. So that is all impossible. I I do think that probably AI is overhyped in data, but there will be some rewards. All this kind of work that we're talking about now is a prerequisite, though. You know?
Aaron:We we won't get to that promised land, AI and data and the the business benefits until you have it clean and and organized and actually usable. So, yeah, I'm very excited about this project doing doing that for for City, Sprint, and and and others.
Joe:I see. I can I just take a bit on on AI? My experience of using Bryte and Chativity Mhmm. Is that, actually, it's a very, very useful personal assistant tool Right. Rather than and I suspect this is how it will probably take hold, is rather than businesses coming, standing back and saying, oh, this is our AI strategy.
Joe:We're going to use Chat u c here to do this function,
Aaron:or we're gonna use
Joe:machine learning to do this. Actually, I think the likes of chat GPT and similar platforms, generative AI, people users, individual users start to adopt it as a go to for all sorts of things. So I use it daily to support with DAX coding, for example, and Power BI or, or even comparing I I used it the other day to compare 2 scripts that I've written just to see whether or not there was any difference between them. And it very instantly, in fact, told me that that there there's one change, which is the one change I was expecting. So that was great.
Joe:Yeah. But, yeah, when someone sends you a a table of data that was produced in Excel, but it's just an image, if you wanna do anything with that data in the old school where you do have to type out numbers, but now you can load it onto Power BI, and it will turn it into an ACE table for you.
Aaron:I think I agree. You know, it's it's an accelerator for individuals' work. You know, it might make let's say the extreme. It might make an individual 10 times more productive if they're using it effectively, and it's in integrated into their work. It's probably not the, like, global panacea for replace a whole bunch of things.
Aaron:It's it's, you know, do everything slightly better or or incrementally better.
Joe:So so in integration into business will probably be more of an organic process rather than the strategic step, where, actually, 30% of the workforce are now regularly using this platform. And Mhmm. Mhmm. Since they're all using it and it's collecting data about how they use it, Yep. If you're on right here, so you've got data privacy because, obviously, that's a bit concerned with these sort of systems.
Joe:But if by the right enterprise level, they they promise to keep everything private. But presumably, that means that the the interactions you have within a business can be shared with other individuals in that business and that then that could become quite powerful. Not to mention it. Yeah. So from that productivity point of view, but but also from a tactical quantity for the business.
Joe:Yeah. I I think that's how it'll grow.
Aaron:Yeah. And then so, obviously, that's that's quite future facing. What else is on the horizon for you and to sprint? Anything at at oh, it's a open question. What's what's on horizon?
Joe:Well, so, so CitySprint, I'm I'm actually in the process of of handing, over to a successor at CitySprint. So so it's exciting times for the VR team there because they've they've got new management coming in. I mean, CitySprint itself has had a new exec management team come in at the beginning of the year. So, actually, there's a lot of change happening in the business at the moment, as is the nature of when you change senior management. So there's there's a lot of changes for BI in that respect because, as as you're well aware, the requirements of the project have been sort of pressure tested with with a new view from from the new management team.
Joe:And they naturally want new reports. They want to look at things in ways that we have to haven't looked at things before. So that that creates a lot of, activity. But that's exciting because otherwise you just get stuck doing the same thing. Right?
Joe:So whenever that kind of thing happens, I see that as an opportunity to rethink and get out of your own way almost, and, and, and look at things with fresh set of eyes, which is often yeah. People be institutionalized into into what they do, and they just do it because that's what they've been doing. So And it's probably a good advice
Aaron:for anyone, isn't it? Like, yeah, anyone in any role is over time getting refreshed in some way. I and I wonder what advice you'd have to anyone find themselves in that position where their role is changing or their company is changing or you know, what what would be your words of wisdom for for what they should do in that situation?
Joe:If a company's changing, embrace it. Just embrace the change always. And keep an open mind, you know, I think it's very easy for people to feel threatened when someone comes along and changes what they do. Their immediate well, very, very often, the immediate response is that it's an attack on their ability Yeah. When someone asks why they do what they do.
Aaron:Whereas it's just change. Right? Like, just change is a little bit scary and change happens, and it's it's in fact, I kinda laugh that change is the only constant. You know, everything's always changing. So, yeah, I think I like that.
Joe:Embrace it. A brilliant quote, which I'm now kicking myself. I can't remember who the quote is from, but I'll have to look that up. But, the quote was changing your mind is the best way to prove that you still have one. I love it.
Joe:And I thought that was a brilliant quote. Change it is great. Yeah. It's really important. Businesses, especially in this, you know, in this, trading environment, the world's trading so fast now.
Joe:It's, it's especially in technology, 6 months and, and things are all hot. So you've gotta be prepared to change. Can you just get left behind? And that's true at every level. You know?
Joe:That's that's from a business strategic point of view about what they're trying to achieve, the products they offer, the services they offer, but actually not for an individual person because people don't stay in the same job very long anymore, not like they used to. I mean, actually, in CitySprint, we're a bit unusual that we've got people I've got someone in my team that's been in a company one form another for 30 odd years. And that's Yeah. Not common anymore. You know?
Joe:I think the average is 2 or 3 years, and then people move on, which means there's always gonna be someone else coming along who's hungry to to make their mark.
Aaron:Make a difference.
Joe:Yeah. Make a difference. Exactly. And and if you're not prepared to embrace change, you're just gonna well, you you can either ride it or you're just being swept along by it, and and that means you're not in control. You're letting someone else control your own your your life and your career.
Joe:You need to get out in front and take the ball by the horns.
Aaron:Yeah. That's, that's that's great great advice, I think, for anyone. It you know, I've heard people say things like, you know, it might help, you know, from their perspective, help some other people not feel alone or helps them deal with that change or yeah. So I think, you know, sharing that advice of embrace it and, you know, this is perfectly normal as this idea. That that is what I would hope to get out of this podcast for anyone listening.
Aaron:So
Joe:Yeah. Yeah. I it is important in there to have to not I think it's very easy to fill as you said, I think it's quite easy to fill alone when you do data work because you don't it's important to have relationships with people, clearly. You do need to be able to communicate with people and the importance of stakeholder relationships because Yeah. Without those, trying to sell anything around data to people who don't live in brief data is hard enough as it is.
Aaron:Let's let's dive into that. So, you know, you you just mentioned the importance of building stakeholder relationships and and how that that affects data. What what does that look like from your perspective? How do you go about talking to people and and getting getting that relationship built so that you can make a difference?
Joe:I think being patient and respectful of everybody's point of view Mhmm. Is really important. So my assessment of of of the world of data and how it fits into businesses are two counts. There's there's at one end of the scale, you've got very, very technical people who are not stereotypically, perhaps, a bit introverted. They like being in their own space.
Joe:They like their relationship with a computer because it doesn't have an opinion, and and it just does as it's told. And it's very logical, and they can they can work at their own within their own framework, within their own mental framework with a computer. And then at the other end of the scale, you've got business managers who very often do not care about how the data came to be. It's the information the data unlocks they're interested in. They don't Yeah.
Joe:You know, they don't need to know whether or not there's SQL code or a DAX query or whatever. They're they're not interested. Right. And there's not many times where those two camps overlap, and so there's a bit of a a gulf between them. Mhmm.
Joe:And that will be an area where people struggle because you've got business managers with their high level vision of what is required for the business so they can run it more effectively, and they've got to somehow translate that to people who really don't care how the business runs or what the data what the information the data unlocks is. Their interest about how the data came to be and where it came from and what transformations were and so on.
Aaron:Yeah.
Joe:So sitting in my opinion, my career is built on the idea that I can sit somewhere in the middle. And I think that's true of probably most data managers. Yeah. That they can they can have the high level conversation and see the business perspective, but they can also get into the weeds of the detail and understand the technical piece. So if they can't actually do the coding necessarily, but they know enough about it to write to ask the right questions and have the right dialogue with their technical people and be respectful of both points of view.
Joe:Yeah. Neither one is more important than the other. They're both you you need both to make it work.
Aaron:Yeah. And I've heard that lots and said lots and lots of times, lots and lots of different industries. I I think that is the vital piece that, you know, the the successful people, I'm telling them the story of of their success, are straddling the 2 camps that don't necessarily talk to each other very well. And in in in this context, you know, business and data and being able to show both is is key. And that that's the sort of secret sauce if there is one.
Joe:Yeah. Like I agree. And, you know, actually, that's a real recipe for success, I think, isn't it? Because you you wouldn't want ever I mean, great if everyone could do everything, but you probably would have a diluted skill set at each end of the sale if if everyone can do everything. You want you want some people who can really get into the weeds of the detailed technical staff
Aaron:specialist,
Joe:and you need people who can gloss over that because they've got other decisions about the future of the company to make. And they they they're gonna spend their time building relationships with suppliers and and and and vendors and customers.
Aaron:Yeah. They
Joe:haven't got the time.
Aaron:Relationships in different different needs of the company. Yeah.
Joe:Exactly. So so, you know, that's that's the essence of a team, isn't it? It's different strengths and coming together and, and making it all work as a well oiled machine. That's a company, actually.
Aaron:That that's it. That is exactly what we're doing. So we were, we were just chatting there about how, the infrastructure work makes a difference to the company or an invisible difference to the company. Making that more tangible for the technical, folks, what's what's what's that look like? What what does kind of good look like, in in a new world, modern modern stack, modern software, modern technology?
Aaron:What's what's what's the state of the art?
Joe:I think the the the the measure of success here is that we move away from investigating challenges and and and issues within the data. We move the conversation in the meetings away from, oh, my number looks different to your number. Why why is this number the number it is? And we get to a point where everybody just trusts the number to be correct and therefore goes into the business to find out why it is what it is. So Yeah.
Joe:If there's if there's a data point that they don't they're not comfortable with or falls out of acceptable, parameters, the first question they ask is, what happened in the real world rather than, is this right?
Aaron:Yeah. I recall in in our work together, at one point, you you you were pretty excited about the kind of reporting we do around what's statistically normal, the kind of test reporting, that kind of test driven development and reporting around the data was was just something you'd not seen before. Like, so we're we're doing something quite innovative there.
Joe:You are absolutely right. I have worked, so my my route into city, CitySprint was as a part of a, consult a business management consultant team. So I I did the data, support for them. So we come into a business. We the tires on on whatever the business wanted us to look at, and I would support them by gathering data and and helping them present it back.
Joe:Therefore, I've had the the privilege of working in a number of different businesses in a number of different sectors. And I can honestly say, in none of them, was there anything that tested the quality or or, consistency of the data itself. So I've seen platforms that will test whether or not a job runs. Mhmm. A pipeline runs, and then it will alert you if it doesn't.
Joe:Yeah. But I've never seen one that says, actually, it ran, but we didn't get as many jobs through or we didn't get as many records through as we normally do or or the, series of values in a column fall outside an acceptable range. Never seen it. And for me, that's a real game changer in terms of data quality because it enables us to get out in front of the curve instead of waiting for some end user to receive a report that's not showing half of the data it's supposed to be showing, we'll get an alert before it ever goes to them, which Yeah. Just moves the needle massively in terms of credibility, efficiency, and it gets hopefully, when those challenges start to die down because we're on it, you know, proactively looking into all these issues and then keeping a knowledge base of what went wrong and and and tidying it.
Joe:Got version control on our scripts so that we don't, you know, implement implement something that then causes another knock on effect, which is also another common Yeah.
Aaron:Just awareness of the change. Yeah.
Joe:Correct. Yeah. So when you take all of that out, the data becomes more robust. It's more validated. People trust it more.
Joe:And, actually, the role of the BI team, certainly in Citispring, which the BI team in Citispring is a is quite a wide scope. So I know in other businesses, BI is quite a narrow scope in so much that they take data provided from IT and they analyze it or write reports on it, but that's it. They don't get involved in the architecture. Whereas in City Spring, we do. We do everything from out of the back of the system, the the management system, right through to writing reports and managing the platforms on which they run.
Joe:So as a result, we spend out my team spends quite a lot of time its time investigating problems, trying to architect data to meet a business need, test it, user acceptance test it. And, actually, what we'd like to be able to do is get to a point where we have a really robust, development pipeline for our data. A really therefore, a really robust live production environment with the right procedures around it so that nothing gets changed unless it's promoted correctly through user acceptance testing correctly. And they then start to spend more of their time business partnering and actually getting away from the the relationship just them and the keyboard that Yeah. Talking to people and understanding how the business that they produce fits into the context of the real world in business and why it's meaningful and what it means to the stakeholders, which Yeah.
Joe:And I think that's that's for everyone. I appreciate that.
Aaron:Yeah. That's right.
Joe:To my mind, that's the juicy part of the work.
Aaron:And I think that's that, to me, is gonna be the, yeah, the the gift that keeps on giving from, you know, our project together is that, you know, once we've got that change promotion with Mediatek in place, we've got the testing, we've got the data quality reporting, you know, the the data team can then build those business relationships and maintain the trust because, like you said, they will be going to them with potential issues or identifying them before they hit the user. And and it kinda strikes me that one of the biggest things that we're having to deal with actually is that we've we've got the software, we've got the technologies, we've got the the new way, but there's a a process change to the team that needs to happen. They need to get used to the idea of looking at the data quality each day, you know, thinking that that's actually part of their role is to care about this will affect the report. We we'd better talk to that group of users. And and and that's the kind of the change that that always takes a little bit longer is getting new people used to the new way.
Aaron:It's that's that's hard. You know, we get the technology in place, but still gotta change the people.
Joe:You're quite right. Yeah. The data and technology is half of of the data story, actually, isn't it, in any business? To draw on your point, is you can't so, again, my experience of coming into businesses fresh and trying to understand what's going on is you gather the data, and that will that will that will give you a pattern. Yeah.
Joe:But unless you talk to people and get the anecdotal side of it, that you don't get the context around the data. So you've got to have the conversation, the relationship with people, the stories behind what's going on in the real world, and then you quantify what they think anecdotally is happening with the actual data and the technology. So, you're right. It is definitely a an underrated and highly important aspect of data management is not the data. It's the people management.
Joe:Right?
Aaron:I love that. It's it's data and people. I mean It is. It really is.
Joe:I I we've we've we've done some KPI projects in the past. I've done a couple of KPI projects where you go into a business and look at what's going on. And the broad process for a KPI project, the way I run it, is is, interviewing people around the business to identify what the key measures are, what actually are the leverage points within the system, how do you measure those leverage points, and then you build the platform, you design the measures, you build a a a report that plays them back. But the bit that often gets missed is but then what are you actually gonna do with that? What's the business management process that comes out of the back of those KPIs?
Aaron:Who looks at the KPI? Who gets incentivized on it? Who cares about it?
Joe:Yeah. Exactly that. What are you gonna do when they're not in the the the threshold they're supposed to be? And then the last step that people often overlook is, how are you gonna make sure that those KPIs remain current? Sorry.
Joe:I just
Aaron:didn't convince these. Thank you. Cut that out. That's my 5 minutes building up.
Joe:Yeah. But how how do you how do you make sure those those those KPIs remain current? Because oftentimes, you find that KPI packs get longer and longer and longer. And as they get longer, the the the relevance of them diminishes to a point.
Aaron:Becomes that information overload. Like, you measure if you're measuring absolutely everything, what what matters? Yeah. So it's it's a bit zooming back to the data team and their project and, you know, their their kind of ways of working. So, you know, we've introduced new technology that is game changing and fundamentally better and freeing up their time.
Aaron:Their focus, therefore, is shifting to something else. So, you know, they're having to get used to where that new focus is. And, you know, that that's, you know, sort of always with people, softly softly, you know, sort of gently, you know, going towards a a new a new process in my view. You know, obviously, you could take different approaches. You could, you know, be kind of a sledgehammer about it, but, you know, you won't won't necessarily get a positive result.
Aaron:So, you know, I think that's that's another thing to think about with the with a project like this in general is if you're making big change, expect the people side to take a little while.
Joe:Yeah. Data is sort of the nervous system of a business, isn't it? And, consequently, it it not only is a a tool for understanding what's happening in the world of the business or anything for that matter, but it's also a tool for shaping behavior and culture in a business. Exactly. And I think it's very easy for people to see it as a tool for reporting and less so for a tool for driving change in culture.
Aaron:Yeah. That's interesting. Yeah. That's really, it's really good point of view, actually. Yeah.
Aaron:Because you're you want, you you hear it kind of all the time. You know, once you start to track a number, yeah, that's that's when you start to start to, in fact, to change. Yeah. That's that's Yep. That's that's the that's the sort of nature of it.
Aaron:So, yeah, I guess the the message there might be be careful about what you're tracking. And, yeah, it's like when you list looking at a big list of KPIs, you choose wisely.
Joe:Oh, absolutely. Yeah. And it doesn't matter what what KPI you choose. It'll have an impact on the behavior of the person being monitored by it, and they will respond accordingly whether or not that's the way you expected them to respond or not. So designing KPIs is is probably as much an art form as it is a science.
Aaron:Yeah. I I agree. You heard that story about the factory where they, adjusted the lights, so it created a a brighter or dimmer atmosphere. And so when they adjusted it to a brighter lighting situation, the productivity output increased. Then when they adjusted it to a dimmer lighting situation, the productivity also increased.
Aaron:So it's just, like, the actual act of observing was increasing productivity, not the lights.
Joe:Not the condition.
Aaron:Classic. Yeah. Well, thanks, Joe. That has been a, you know, pretty great.
Joe:Yeah. No. No problem at all. Thanks for having me.