Drill to Detail - Drill to Detail Ep.62 'Enterprise analytics, DW Modernization and the rediscovery of Data Governance' With Special Guest Mike Ferguson

Episode Date: April 3, 2019

Mark Rittman is joined in this episode by Mike Ferguson, long-term analyst, consultant and Managing Director of Intelligent Business Strategies to talk about data warehouse modernization, analytics and big data project success within enterprise customers and the re-emergence of interest in data governance and master data management within the industry.

Welcome back to another episode of the Drill to Detail podcast, the series about the people and products driving innovation in the data and analytics industry. So I'm your host, Mark Whitman. And this episode, I'm joined by Mark Ferguson, Managing Director of Intelligent Business Strategies and Analyst, Consultant and Fellow Brit, who seems to be traveling around Europe and speaking at events even more than I do. So welcome to the show, Mike, and why don't you introduce yourself to the audience and tell us a little bit about what you do and your company. Okay, thanks, Mark. So, yeah, what I am really is an IT industry analyst and a consultant. I specialize in data management, BI, and analytics.

I've pretty well been there all my career. We're a small boutique sort of industry analyst firm that does three major things. We do research. we do education a series of educational courses that we teach all around the world and then we have of course the consulting that we work with various customers at strategic on tech and technical levels in the areas that I mentioned, you know, data, analytics, BI, and any related peripheral activities around it,

like architecture and that kind of thing. Okay, okay. And what interests me particularly about speaking to you was, first of all, I've seen that you've been presenting a lot and running seminars and so on, but you've been particularly focusing on, I suppose, the enterprise side of analytics. And I think there's been been a lot of talk you know about kind of talking industry about so it's democratizing analytics and analytics at a small scale and all that kind of stuff

but you know you you've been really looking at it from an enterprise level and looking at some of the I suppose some of the kind of the really important um sort of difficult knotty tasks around this around things like data management and master data and so on. I mean, tell us a bit about how you got into the industry. What was your kind of entry into doing this sort of thing? Well, my initial entry point, I guess, was as a database administrator way back in the 80s,

in the very early 80s. And I just started as a regular database administrator. And a lot of people were very green behind the ears, I guess, in those days around this topic. And a new technology was emerging. I was very fortunate in my career to get the chance at quite a young age as a young researcher to co-found a company called Cod and Date with two relatively giants of the computing industry, Dr. Ted Cod and Chris Date.

Ted, of course, invented the relational model and got the Turing Award. Two of them are, of course, both British, but lived in the US. And of course, the relational model is a mathematical underpinning to every relational database that ever emerged in the industry, whether that be popular products like Oracle and Microsoft SQL Server and IBM DB2 and Teradata and Exadata, all of them. I mean, they all are underpinned by what Ted invented. And, of course, if you ever were a student of computer science

anywhere on the planet, you probably got an introduction to database by C.J. Date, who wrote all the books in the 80s and 90s around relational and popularized it, if you like. And so, yeah, that was an amazing experience for me. It was an incredible learning curve working with both of them. It was a very small boutique firm, and we operated pretty worldwide and remarkably in an era when there was no internet so when we uh turned up at the beginning of the relational era to talk about this stuff you know

we were literally playing to thousands of people in huge conferences so yeah and then after that i i had the chance to switch to the other side of the fence to go to the vendor side. And I was fortunate enough to be headhunted by a relatively young, small startup database company in Los Angeles at the time, which was Teradata. And they asked me to join. And I joined the UK organization initially, but within a pretty short few months, I was moved to report directly into Los Angeles and became a chief architect on the product strategy and architecture team there.

So that was tremendous. I mean, it was the first database really that pioneered massively parallel database computing aimed at the analytics market. And I suppose in a way, it was Teradata that took me down the kind of analytical road. Whereas before that, I was looking at relational in a general purpose, both operational and analytical workloads but and then obviously having worked there for I guess about nearly four years again I then had an old friend of mine

who's now retired Colin White again a British guy who was an independent industry analyst in the U.S., and Colin approached me, and we formed a company called Database Associates, which had a small number of people in the U.S. and initially just me in Europe, again aimed at the kind of data management, BI market, data warehousing market in those days at the end of the 80s, early 90s. And I started that and all the others have retired.

So, yeah, I'm kind of still going. And obviously, I've got a number of other associates that I work with. And I live in Europe, but I work pretty worldwide. But I think it's fair to say, you know, most of my life is spent in and around mainland Europe and UK. And I'm in the US probably about five or six times a year, I guess. So occasionally get further afield than that too. Excellent. And I think, you know, again, what interested me speaking to you was,

you know, there's a lot of people that I know in the industry industry kind of all in our 50s early 50s you know uh and and and you know and but one thing what's interesting about you is is you still seem to be very kind of relevant to to to what's going on now i mean it's you know there's a lot of people around who contributed a lot into the industry um and who aren't maybe so you know relevant now or certainly maybe sort of semi retired or whatever but you know you still seem to be kind of like finger on the pulse really of what's going on uh in the industry now uh and and relevant to the fact you're running kind of seminars and so on and actually consulting for companies it's not it's not just because you're a celebrity it's you obviously put value into places really i mean it's um it must be yeah i mean well try i mean i think there's no question

about i mean i'm chairman of big data london you know the fastest growing conference in europe i think now we had seven and a half thousand people in london last year and growing hopefully bigger again this year but but yeah i mean i think the challenge for me is is you know the area that i chosen for most of my career data uh bi BI, and analytics, has been pretty evergreen. I've had my business now for 25 years as an independent. But I think there's no question it's still evolving. It's moving. And that's exciting.

I mean, you get the chance to learn about and work with new technologies and figure out what value you can get from it. And occasionally when you've been, I suppose, in there as long as I have, you often see the whole cycle go around again. I mean, some of the things that are emerging now, I've definitely been there the first time around and so a lot of the things you learn first time around

are still relevant, believe it or not. But yeah, technology is changing and a big push obviously on, I suppose, what's new now is just a sheer complexity around data, the number of data sources, the number of data stores on multiple clouds in the data center and managing and governing the whole thing, as well as trying to, you know, use modern day new analytical algorithms and whatnot to deliver value from it and add it to the kind of traditional data warehousing and bi stuff that's been around

and underpinning decisions for so so long so so you know you've chosen to sort of focus your career on the kind of enterprise end of of kind of it and you know what within i suppose um within that world within the kind of the the world of corporate enterprises you know, within, I suppose, within that world, within the kind of the world of corporate enterprises, you know, what's the number one issue around analytics and data management that CIOs and CDOs and so on are asking you about? What's the kind of top thing on their mind at the moment? I think the top thing on their mind is kind of a dilemma.

The dilemma is how do you get governance and remove garbage in, garbage out on the one side? And then on the other side is, as the Americans would say, how do you turn on a dime? I mean, how do you deliver value quicker, transform, become more agile, and really drive competitive advantage by leveraging data and analytics. So I kind of think, you know, when I started,

there's no question that kind of data warehousing and BI were kind of off to one side and transactional systems were kind of dominating to a large extent. I think nowadays what's become pretty clear is that data and analytics have moved to the middle of the enterprise and every operational system and every process wants it. And so, you know, there's an enormous thirst enterprise wide for data and analytics. So I think the problem is, on the one side, is how do you integrate all the stuff that's going on out there in this space, you know, because it's gone way beyond data warehousing, you know, we've got all kinds of analytical workloads. And whether it's streaming, or whether it's, you know we've got all kinds of analytical workloads and whether it's streaming or whether it's you know graph analytics or analyzing huge amounts of click stream or

unstructured data in in big data platforms or you know any of that in addition to the kind of classic warehousing setup and so there's a kind of you know we've moved to a world of workload optimized systems but I think what's pretty clear now is how do you integrate it and deliver value from it I mean I think it's it's it's clear that culture is causing a major issue in companies that are striving towards um or wanting to become data driven okay i mean there's a lot there's a lot in what you said there's a lot of things to kind of unpack i mean just start going back through them and and try and tackle them one by

one i mean so so you know you you've yeah i think i've heard you talk in the past about modernizing the data warehouse you know and you've talked I suppose, the thing you said there was about turning on a dime and retaining the agility. I mean, what's the kind of the business problem that the organizations are facing there? And what's the impact, I suppose, of not addressing this? I think the problem is that data warehouses on their own are not doing it anymore in the sense that that structured data, it's very well bedded down. It's very well understood. It's very mature. And, yes, there are new things that have come along,

perhaps more modern data modeling techniques that are more change management friendly, like Data Vault and things like that. But I think generally speaking, that's been an area that's pretty well understood. Probably we're over 30 years since Barry Devlin wrote the first white paper on data warehousing now, and probably about 28 since Bill Inman's book. So, you know, that's a mature space. But I think the other thing that's happened is just a sheer thirst for new data

that is not all structured, you know, that can be semi-structured like JSON or XML. It can be very high-velocity data like clickstream or IoT data coming in or even huge amounts of unstructured data, whether it's text from social media or review websites or even internal email coming in from outside of customers. I think there's a big push around customer data way beyond just what they buy. People want to, you know, particularly at the exec level, they're pretty sacrosanct about customer data and investing in protecting and retaining customers. But I think, you know, the issue is because there are so many projects

that are happening, if you like, outside of IT in the area of analytics now, there's a real requirement to say, given that all, you know, we've gone from a data warehouse to what I would call an analytical ecosystem of multiple underlying types of platform that are optimized for specific kinds of analytics. And I think the requirement now is how do you bring those together and integrate it with, you know, the data warehouses that we've put in place. And at the same time, I think there are some question marks around agility on traditional data warehousing

and how do you therefore speed up the ability to change it and add new capabilities to your data warehouse that you couldn't do before. For example, if you have to make structural changes quickly, what happens if you've got a data warehouse that you couldn't do before. For example, you know, if you have to make structural changes quickly, you know, what happens if you've got a data warehouse and, I don't know, let's say in the European Union you've got sales analytics and you've got 27 or 28 or soon to be 27 data marts, one for each country. If you change the data warehouse, the potential impact is that 27 data marts

may also have to change and all the ETL jobs may have to change supporting it. And so that means that a simple change to a data warehouse could have a dramatic domino impact and then cause a very long amount of time before change is fully implemented across the board. And I think, you know, the problem is that with those systems being in production, companies now want, you know, to be able to change that far more rapidly

and quickly get things implemented and move on, you know. So I think there's that. I think there's the need to lower the latency of data to be able to integrate data warehouses with streaming, for example, to be able to get data and also to be able to join data and data warehouses with data in, say, other systems like big data systems, whether that be Spark clusters running on the cloud or whether it be Hadoop systems in the data center with Spark running on it or something like that. And so there's, you know, because I think what people want is they don't want to have to figure out how to access

multiple different analytical systems with different APIs. They just want it to be easy to access integrated data that may actually end up being in multiple places under the covers. And so in that sense, the question is, how do you produce related types of insights in different platforms, but still make it really easy to access those insights from traditional BI tools and applications. So, I mean, that's a great summary, I think, of what a lot of customers are kind of saying, and certainly it's the lead-in often to a kind of pitch from a vendor

like Oracle, for example, to say the answer is you buy everything from us and it all does this stuff. I mean, what are you finding is actually working or what would your advice be to customers with this kind of issue and how would they start to tackle this, do you think, in a way that's going to give them kind of results really? I think the issues are that over the last few years we've had a bottom-up approach to development where different parts of the business have had their own budget and gone off and done their own projects in a relatively autonomous fashion and so they've not really been tied together in any way and I think what I'm seeing now is that a lot of these projects

have failed in the sense that the value has been a lot less than expected. But I think what I'm seeing now is a real requirement for top-down where we're now seeing organizational appointments like chief data officers, which never existed five or six years ago. But these aren't IT-related appointments. They typically are a business executive who, in many cases in the UK at least, reports directly to chief executives these days.

And I think it's pressure from those executives who are accountable to deliver value for business from data and analytics that's really beginning to, let's say, bring to order all of the disparate projects that have been going on around there and trying to align them. I think align them around what you might say, well, there could be multiple different kinds of data and analytical projects associated with fraud, everything from streaming analytics to stop a fraudulent transaction in flight, which is a real-time requirement, all the way to graph analytics that are trying to find fraud rings, for example. And yet, you know, whether that's traditional data warehousing, big data, you know, SQL or streaming, there's a collection of projects that together would tackle the fraud problem. The same could be said for customer engagement or optimizing

business operations like a supply chain or manufacturing or something. And so there's a collection of data and analytical projects that I think need to be brought together and targeted at specific business problems. And I think, you know, what I'm seeing now is pressure from C-level executives to kind of say, look, you know, we've had an awful lot of new technology over the last five or six years. We've had a lot of people downloading and technology like data science is changing dramatically. New algorithms, new libraries, new technologies coming out all the time, and data scientists wanting to grab these and see if it makes any improvement

on what they're already doing. But I think the pressure from the C-level executives is say, look, don't forget we're working for a bank here, we're working for a retailer, we're working for an airline. We've got to deliver value and we've got to align it with business strategy. And I think it's pretty clear now that the pressure's on to organise much better than we have done.

It's not always a technology problem, it's a people problem as much as anything. I think there's this pressure now to to organize and try and fuel a much more data-driven culture you know and yeah for the first time well not the first time but I think it's only recently in the last few years that I think data and analytics in particular has has really resonated in the boardroom. And I think that's the thing that's happening. I mean, clearly companies that don't have an executive at the C-level accountable, I think, have a much more challenging problem

because they still have autonomy in their business units and in their departments, and they may be pulling in all directions with no sign of alignment across the organization so whilst I'm not saying that it should be centralized or anything like that I think the classic cycle has emerged you know we we went from centralized IT doing all of this to distributed sort of setup with business doing their own thing. And I think now we're calling for a kind of federated setup where there's some kind of overarching responsibility and alignment that's trying to reach across the enterprise and bring everybody into line. Okay. I mean, so do you find there's an appetite for kind of enterprise-style projects now with this kind of area?

What's the kind of typical time span of an initiative or a project now? And, you know, is there a budget century now for this sort of thing? Or how do you find that really? I find that a real problem, to be honest. I mean, I think there's not a lot of budget or a lot of time for so-called enterprise projects. People want quick win. They want delivery in a specific area and they want incremental growth, which is easier to say than do. I think the problem is that there's often a case that these projects are not sort of tied to a bigger picture

which is part of a front of a jigsaw box which is how do all these components come together to deliver the value you want. And I think for that reason there's no question at this point in time that the whole distributed development is still driving individual investments and area investments in projects. I think the appetite for enterprise-level projects has not come back, except for, I think, where it's beginning to come back, and that is in the area of data management.

That is to say, well, that's what we mean by that. We've now got data coming in, you know, all different types. It's being ingested into all kinds of data stores all over the enterprise. Everything from cloud storage to NoSQL databases like document databases and column family databases and whatnot, to still staging areas and data warehouses and big data Hadoop systems. The problem right now is nobody knows where any data is. And so there's no mechanism to find out what data is where, to understand the degree of duplication.

And so we're seeing a big interest right now in things like data catalogs in order to automatically discover where the data is. Because I think the frustration is that whilst all of these investments are going on in data science, we're still potentially getting garbage in, garbage out. I think privacy has been a really big deal recently with gdpr and that's been brought up you know made its way to the boardroom and all the other initiatives associated with it like ethics and all the rest

of it around data but i i think generally speaking the frustration is that if you spend all this money and everyone prepares their own data and produces insights from various analytical models and whatnot, the danger is that it's still inconsistent and incorrect. And I think the appetite is now there to to get rid of garbage in garbage out and for that reason i do believe that that data is getting um being looked at from an enterprise-wide level and and and admittedly i think there's still a big issue you know i think even if you have an enterprise executive like a CDO responsible and accountable for it. And primarily that's going to be data governance, but also I think driving competitive advantage from data. Then they are tasked with putting together a data strategy.

They're tasked with being offensive in the form of driving value from a data strategy and taking an offensive stance on that to make sure that that happens quickly and incrementally and continues to drive new value whilst at the same time having a defensive element in order to be able to govern and remain compliant and all of those good things and make sure that privacy and quality and all the rest of it are upheld. So I think this is one of the reasons why frustration with self-service data prep is happening,

and people are now looking to things like data lakes as a way to perhaps ingest data into a data lake environment and then drive data projects that don't do everything for everybody. But I mean, if you look at data scientists, what, say 80% of their time preparing data? You know, really? I mean, we need to get over that but i think i think the recognition is we have to try and accelerate that at the same time to produce i think data products or data assets that are trusted that are reusable and that accelerate these projects rather than force everybody back to raw data all the time just a couple of things

you met you touched on there. I mean, you mentioned about being data-driven and you mentioned like data lakes there as well and agility. I mean, data lakes. I mean, it's, yeah. I mean, in a way, it is kind of like having your cake and eating it. It's a bit like our Brexit strategy, isn't it? In that, you know, it's saying, well, let's land all this data here

and a purpose will arise for it. And the end users will take charge of kind of making sense of it. And yet logic would say that that is perhaps a bit of a kind of wishful thinking. What's your experience been of data lakes? And where kind of have you seen them show value and have impact on the business maybe? It's a good question. I mean, first of all, I think it's a terrible term.

I think it's been misinterpreted, often directly associated with a specific technology like a dupe, which means if you have a dupe, you've got a data leak. If you don't have a dupe, you haven't. I don't believe that whatsoever. You just look at cloud storage you know people storing large amounts of data on amazon s3 or azure um object storage or azure data lake or or google cloud or whatever um you know i think the idea behind it is is an

organized collection of data so it doesn't have to be centralized you know where you bring all your data to one physical place before you can process it i mean i think the misconception of a data lake is that is that all data has to be moved to one place before you can process it i don't agree with that i mean I think there's other configurations for a data lake, such as a logical data lake where, you know, if you take a very large bank or a very large pharmaceutical or a very large retailer or whatever, they can't physically bring all data to one data store. It's just not practical, you know,

especially when many of these huge companies operate in tens or even over 100 countries in in some cases but i think when you look at it like that then the recognition is well can you organize storage so that there are you know dedicated stores that are associated with ingesting data dedicated stores associated with staging during the curation process, and dedicated stores that only hold trusted, ready-made data. And then I think that's really beginning to organize it. And so you start to see the concept of zones

and that kind of thing emerge. But I also think the other thing that's emerging now is not only could you give me a data catalog to tell me what data I have, kind of thing emerge but I also think the other thing that's emerging now is not only could you give me a data catalog to tell me what data I have and where the personally identifiable data is and what quality it's in but can I also use this technology to create a marketplace where that marketplace is only showing consumers of information, trusted data assets that they can pick up and reuse. I mean, if I'm building a model to do with customer churn,

and I already have a trusted asset in customer master data, why on earth would I ask the data scientist to go all the way back to the raw data and build customer identity again by pulling together attributes from different stores to just say, this is what, here are my customers. I mean, if that already exists, why would you do it again? Of course, there's additional insight that you want to know about customers. But if the customer identity data is already integrated, then why do it again? So I think the idea here is to not go the whole distance from raw data to value, but go, you know, all the way to the last mile in a sense that give me pre-built, you know, ready-made data assets

that can be rolled out incrementally and then new projects that come along can leverage what's been invented before. But please don't let multiple projects go and do the same thing again and again and again, because inadvertently, they potentially will create multiple versions of the same thing that are all inconsistent. And then trust breaks down again. So I think the issue here is, how do you break this cycle? I mean, everyone says, well, we want to go back to the raw data.

Nobody's throwing away the raw data. But I think what we are saying is can we incrementally continue to accelerate projects and make them go faster and faster by producing these things, these data products or data assets, and making them reusable and making people know they exist through things like a catalog and the so-called data marketplace sort of idea. It's these kinds of things that I think people are now looking at,

and I suppose in the sense that the perception is that the data lake is the answer. Well, I think the data lake's the starting point, you know, but I'm going from data lake to data marketplace, you know, and so it's the marketplace where you really want to focus. That's the thing that's going to get you the value. In fact, that's the topic of my presentation in TDWI in Germany this year. It will be about marketplaces. But I really see now lots of companies trying to say, look, is there a better way or are we going to have hundreds of business use business analysts and and data scientists all with self-service data prep all building building their own individual silos and everybody integrating

data and nobody's sharing anything that they create which would be a disaster i mean that would be that would be even worse i mean that would be you know loads of different versions of the same thing and no one wants to trust any what anyone else has created so i think we've got to somehow find a way out of this but uh you know that that to me is the the big challenge right now so so are you finding i mean on a sort of similar or kind of related note you know um a result of resurgence of interesting things like master data management and i suppose customer data platforms and so on. Is that a topic that's now kind of relevant again to people you speak to?

Very much so. Ironically, the BI survey, the largest one out there that Bark does, a good friend of mine, Karsten Banger, that I do an event with every year. His survey came up with, for the first time I've seen that back end of 2018, it said that master data was number one in BI. I mean, that's amazing. But what does that tell me? That just tells me, you know, data quality and MDM are now fundamentally important because everybody's fed up of garbage in, garbage out. So I've been teaching master data management since 2007 in lots of different countries around the world.

And I have to tell you, in the last 18 months, the numbers on my MDM classes have just skyrocketed. I think GDPR has helped with two types of master data that's a customer and employee um but uh without a doubt big big interest in master data now whether it's customers or products or suppliers or materials or whatever it is um and i and i think it's all down to governance and trust. You know, the people want this data is master data for me is probably the most widely shared data of all.

You know, you use it in operational transaction systems and you use it in analytics, you know, whether it's data science or whether it's traditional data warehousing or whether it's graph analytics, you're going to need master data to put context on what you're looking at. So, yeah, I do see a huge resurgence in that and to some extent, no SQL databases even getting in on the graph databases and MDM being looked at.

The other thing you mentioned, I think, there was customer data platforms, which is another area, which for me is kind of like an evolution from customer master data, which is really where you're combining that customer master data with insights coming out of data warehouses and out of big data to give you a kind of holistic view of everything you want to know about a customer and that you can potentially feed that insight into particularly front office applications, whether that's marketing platforms that are doing multi-channel marketing or sales applications such as Salesforce and whatnot, or any CRM application, whether they're contact centers and whatnot.

So I kind of see customer data platform is high priority because I think it's always been the case that in the boardroom that customer data is considered highly valuable. And I think it's not surprising that there's a demand for customer data platforms. And we've seen a few MDM vendors evolving into offering these customer data platforms. But I think, you know, in order to get the value out of them, you clearly need to leverage those in in all the front office you know applications whether that be customer sales or service so so really mean kind of you know i suppose rounding off the conversation the the you you know you you talked

about kind of going into a lot of organizations and helping with these kind of strategies and so on what what are the kind of the i suppose the challenge with a lot of these questions is, where do you start? And how do you actually kind of start to get the organization to do something and to, I suppose, get some momentum in this area and so on? I mean, what would you say to someone in an enterprise that was trying to get a lot of these things moving that you're talking about?

What's the kind of baby steps they can take to get these things going and to show some value maybe at the start i think the first thing that has to happen is is anyone who wants to show value had better understand their own company's business strategy you know it's always been the case that business strategy alignment for me is number one successful you know i mean if i am the assumption well what it means is like if you if you have a business strategy it's typically got here are my strategic business objectives you know these are the you know this this is what we are trying to achieve as a business um it it would also have some kind of KPI, key performance indicator,

that would tell you or allow you to measure whether you're on track to achieving that objective. And it probably has some kind of targets set for that KPI, such as it has to be at a certain level by a particular timeframe. And then there are people who are accountable for the business achieving those goals. And then those people will decide on which projects and how much budget's being divided up into, you know, these projects and associated with these projects in order to invest in and make things happen in order to spend the money, if you like, in order to help the organization achieve

those goals. And so at a business strategy level, you will find that in every business. And if you don't find it, then there are much bigger problems. But very clearly, a data strategy and an analytic strategy need to align with that. So if it is a strategic objective of reduce fraud, then what data assets are you producing? And what analytical models and what BI reports and what BI dashboards and what prescriptive models that drive automatic decisions

are you putting in place to reduce fraud? And are some of them running in real time or some of them running in batch or some of them running in a warehouse or some of them running in a big data platform like Hadoop on the data center or Spark in the cloud, you know, or some of them running in a NoSQL data. What's the deal? And how does that family of projects come together to achieve that business goal? And then the same would be, again, like improving customer engagement.

Then again, same thing, you know know what data projects are you producing what data are you producing what um analytical models are you producing what um bi and reports and and prescriptive models and whatnot are being put together to improve customer engagement and who's doing them and how are being put together to improve customer engagement and who's doing them and how are they coming together to achieve that and then you know are we able to measure the effect of those models that have been produced by data science to say are we actually in do you know doing that or not are we retaining our customers are we driving new revenue as a result? I think the fact that you have a business strategy really helps organize a data and analytical project, whether they happen to be front office customer related or back office, supply chain, or financially related, and, you know,

whether it's everyday mainstream business operations, you know, if you're driving increasing revenue, if you're reducing cost, or if you're very fortunate to do both, then you're widening profit margin and driving value. And to me, you know, that is really what we have to do here. I mean, I think there's misconceptions of stuff it all in a data lake and we hire all these people and they're all just going to find golden nuggets and we're just going to disrupt everybody. I think that's really very short-sighted

and I think erroneous to go down that road. I mean, we've been down that road already. We did that with data warehousing 28 years ago. Everyone talked exactly the same thing, single customer view. Everyone talked about, oh, we'll throw it on the data warehouse and we'll just deliver value. We know we have to design this stuff. We have to know what we're going after.

And I think that's still the case. That's why I firmly believe that we need business strategy alignment to make it succeed. That said, the next question is going to be, well, what projects and then what are you going to do these things excellent excellent well look mike thank you very much coming on the show it's been really interesting hearing your your take on sort of where enterprise enterprise analytics is going and uh yeah i just really enjoy speaking to you and um

yeah thank you very much and it's uh yeah great to speak to you Thank you.