Good morning, good afternoon, good evening, good whatever time it is, wherever you are in the world.

I’m Malcolm Hawker. I’m the host of the CDO Matters podcast, and I am glad that you are taking time out of your day or your evening to join us today and have a conversation about all things cloud. Is it feeling cloudy where you are? Maybe it should be because that’s what we’re gonna be talking about today on the podcast. I know that’s a very general broad discussion.

Of course, we’re gonna focus on data. We’re gonna focus on data infrastructure, data ecosystems. What do CDOs and other data leaders need to be thinking about from the perspective of the cloud?

Since we’re talking about the cloud, I am glad to be joined by an expert in the cloud. Ayman Hussain, thanks for joining us today.

It’s great to be here. Thank you for having me.

Awesome. Ayman, why don’t you share a little bit of insight on your background or is particularly the last couple of positions you’ve had, but I know you’ve got a a long consulting background as well. But, just share a little bit of the of of your background and your expertise so people know where you’re coming from, the perspective of your cloud expertise.

Absolutely. I’ll start with a little bit of, storytelling here. I had a mentor who had said, hey. If you wanna figure out how this sauce is made, you should find out all the moving pieces of the puzzle and start working upstream.

So with my journey to where I am today, I’m with Google today. And prior to Google, I was with Microsoft. And prior to Microsoft was I was at a strategy consulting company called Slalom. And if you think of why I’ve chosen these careers in this journey effectively is to figure out where decisions are made for corporations and enterprises that influence technology adoption and what directionally they’re gonna do.

So my career is essentially a stepping stone of roles and jobs where I wanted to figure out who’s making those decisions, where are they coming from, and working upstream to that space, what gives me the ability to see where these decisions are made. And along the way, you have to pause for a little bit, understand who what these roles and responsibilities are. For example, you’re a CDO interfacing with other CDOs and other organizations, understanding what their initiatives are. And so Journeypaths required me to be a strategy consultant where I actually understood how strategy implementation solutions and consulting influences this decision making.

And then I picked a couple of brands. In this case, I picked Microsoft for a while and now I’m at Google to understand how technology software companies and telecom companies are creating solutions and capabilities to address those strategies that are being placed to ultimately serve a customer’s outcome and need for whatever the business and industry vertical they are in. So that’s where I am coming from. So today at Google, I’m a customer engineering leader.

My role and responsibility is to help customers do more with the technology but for the right purpose. So my experts and I get engaged to understand what the business solution outcome you’re looking for and see if we have the right LEGO blocks and pieces to make it happen.

Well, so that’s an interesting kind of lens for from a career progression. I mean, understanding decision makers and understanding what weight makes them tick. I think if you were to ask me to tell a similar story, I would have gone from the view of more problem solving. Right?

I’m I’m a natural problem solver, and it’s not like I don’t care who the decision maker is. I certainly do. Of course, I work for a software company. We we care about these things, but it’s it’s an interesting perspective.

So so then who is the decision maker in the cloud, or is this is this mostly a CIO decision where the CDO is being seen more as a beneficiary or user? Or are you seeing CDO’s influence or even own decisions about cloud migrations?

It’s a great way to say we’re right. It’s an evolutionary, nonlinear way to think of it. There was a time when CIOs, CTOs would make decisions in spaces like this. The role of a CDO, as you know, is relatively new in the context of, leaderships in these organizations or in, solution providing organizations.

And what I’ve observed is the supporting roles, CDO, CTI, CIO, CMO, you name it, CRO, all of these are there to achieve a broad functional differentiator for the enterprise. If you are in the in business of profiting, you want to have low cost, high revenue. And if you think of that mindset, you want to build solutions using technology to have the right approach to solve that problem. Now, the decision makers are really more decision order, givers.

They want something to happen, a board level conversation, a very high level investor conversation. And then the roles and responsibilities of the C suite will now step in and take care of it. For a CDO, it’s all about data. For a CTO, it might be all about the hardware and the technology.

CIO would be about the information that data has and how to manage that. And so the decision making, depending on the solution space, really, now has segregation, not because of a need to have it. It’s getting complex, it’s getting big, and therefore, these roles are separating from the original C suite role of a CTO or CIO that just did everything. Now we’re segregating it based on the fact that these swim lanes are getting very complex.

Some companies even have a cloud officer, chief cloud, officer to understand how optimized cloud rollouts are going to take place and achieve success in that space. So it is clearing the, role differentiator, but it’s helping vendors and providers like yourself and myself to be very niche and, focus on those capabilities that will help enhance that outcome ultimately.

So to paraphrase, you know, CIO is is driving from the perspective of infrastructure and scalability. CDO is driving from the perspective of data insights, analytics, data science, AI.

Those things need to come together and be kind of lockstep in order to make sure that whatever you’re doing from a technology perspective aligns from a strategy perspective. Okay. Makes total sense. So this may sound like a maybe a bit of a dumb question, but I ask a lot of dumb questions.

Are are is it there’s still a lot of companies that are still just now, you know, working on cloud migrations. Correct? I mean, this is not a done deal.

It is, it hasn’t, aged out yet. It’s still relevant for many organizations.

What what a ballpark on percentage? Like, what do you where do you think or maybe you have stats on this.

On this segment. So you won’t be surprised. Large enterprise and corporations that are dynastically large have a very low cloud adoption percentage because of the fact that first of all, the entry cost is so high that they cannot accommodate the entire organization moving to cloud. And the other part of it is they have a lot of capitalized cost.

The way they did accounting was different. What we see is greater success in medium, small, and even startups that are cloud first, mobile first mindset. And they are winning that conversation because they already have baked into their strategies. For example, if you’re a startup and you’re going to go raise funds through investors, you better tell them where your money is going to go.

So you have an idea. This is where it’s going to product development, software development, cloud costs. You break it out right away So there’s no surprises. But if you’re a large company, you’re not talking with the revenue officer or marketing officer or CFO because kind of gets involved in a lot of the capitalized costs.

And there’s a ping pong effect. A lot of people have pivoted hard into the cloud And then now they’re coming back because some of those cost optimizations requires that you understand it in a way that makes sense. I use the car analogy. A lot of people buy cars every three, four years.

Some people will keep it for ten, twenty years. It’s all about what you’re doing and what purpose it is. If you’re a fleet manager and your job is delivering products using cars, you probably change over your vehicles faster and quicker because that’s your moneymaker. You can’t afford to have challenges with maintenance and costs otherwise.

But if you’re just using it for transport or even using it not even in the core business of your, revenue generation, you may have a capitalized cost saying we’re gonna keep these for ten, twelve years. So a lot of the IT expense is kinda wrapped up around that, and therefore, the adoption in the higher enterprise is a lot less than you think it is.

Well, that’s interesting. It doesn’t surprise me at all that companies are kind of addicted to CapEx and being being able to depreciate, infrastructure spends because I’ve certainly been in that boat, and been the beneficiary of kind of user lose from an OpEx perspective.

So so that’s interesting. Now is there particularly I know you are very close to the oil and gas, but but where from, like, a segment perspective do do people kind of line up? Is it you you had mentioned midsize tends to be the ones that are moving a little bit more. Are there any industries that are maybe lagging a little bit behind others?

Absolutely. Yeah. So industries that tend to have a focus on regulation both from a compliance perspective or safety perspective. You mentioned oil and gas.

Yes. I do live in Houston, Texas. A lot of the oil and gas companies are, concerned of the impact for safety at both environmental and people. Right?

So they wanna make sure that they’re not compromising something that may have a reputation of harm or have challenges in the future from a litigation perspective. And so they are a little bit apprehensive to decide. So some of the dynastic industries that have compliance and regulation heavy, health care being one of them as well, tend not to do aggressive cloud movement. But then their industries, for example, retail has pivoted hard into cloud capability because their business model requires them to be available all the time, ready available in the context of servicing.

And then their cost of infrastructure is so marginally, slim that they have to not be able to have cost that is sunk in a CapEx. Like, open a store, set up cloud, the store doesn’t perform well, shut down the store, the cost goes with it, from a cloud perspective. So retail tends to be very aggressively cloud oriented, and, sort of doing very well in that space.

So I’m a CDO, and I’m I’m at one of these companies that has not yet kind of made the jump. What what are the business pains that I would be experiencing that would make me want to start seriously investigating a cloud migration? Is it a lack of scalability, lack of flexibility? If I’m a CDO, I may not necessarily have a lot of accountability from just a pure operations perspective. Like, I don’t own the data center. So as a CDO, what are the things that I that I should be looking at from a cloud perspective to like, what are the problems that I’m gonna be solving with a move to the cloud?

Absolutely. So I’m gonna use a story that’s happened for real. I’m not gonna use the names of the companies. And this is about fifteen years ago, twenty years almost, depending on how you look at it.

There was a chief executive of this very large dynasty company. He kind of had a spidey sense that something was not looking right in the world. And he called his CFO and said, close the books. I said, go ahead.

Close the books. I wanna see where we stand. They were publicly traded. And that time, in that juncture, depending on the technology because they were aged, a few years ago, it took him eighteen days to close the books end to end.

And on day twelve, a part of the world in the way of financial markets had a challenge and, hiccuped which cost this company a lot of money. And this C level officer was significantly peeved because now you have a challenge of analysts and, you know, stock market and shareholders said we need to close the books faster. At that time, and I’m talking almost ten, fifteen years ago, when they pivoted and adopted solutions that were both cloud and, more modern, they were able to close the books in eight hours. And now I think they do it even faster.

So think of it that way. Why would a CDO be engaged? If you have, something that requires you to have the data to make an analytical decision insights and you’re not able to provide that, you may lose out in context of today’s world. For example, the pandemic.

No company ever had a disaster and business continuity planned for a pandemic. They had versions of it in smaller pockets. They did not have it in that context.

So imagine the the CDO officer now has to step up and do something that gives you the ability to have pivot around it. A ship goes sideways in the Suez Canal and supply chain logistics take a hiccup. You have to plan for these contingencies and all of that comes down to, do I know what I have? Can I run analytics again instead?

Now that may be a scale conversation. So you can say, hey, I need to scale, so I’m gonna use cloud for storing the data as well as use the analytics. Or it could be an internal on prem conversation for you saying, hey, you know what? I already have the scale.

I just need to invest in the right solutions and capability.

In oil and gas, one of the things that is very data intensive is seismic processing. And you’ll be surprised of how many of these companies still do seismic processing in house for a couple of reasons. The scale, but also the cost of these high performance compute clusters and GPU clusters is so cost prohibitive in the cloud in the context of what they need agility for. Sometimes they still maintain their own Cray clusters and these high HPC clusters because they say, look, you know, we’re gonna run a a simulation that will take twenty eight days to run.

That cost in the cloud may not may bankrupt me, but I I already have the sunk cost. I can do this efficiently and build these models and algorithms. You know, think of, like, weather forecasters or climate forecasters, whatever it is. You know, air pattern, sea patterns, whatever they wanna do, they will use a traditional method to do that, but the role still applies.

You need to know what the data is doing, how it’s doing it to you as a CDO also coming from the, employer that you are with. Does that product make sense? Because just knowing what the data is is not good enough. You need to classify and to understand and make sure that the data is appropriate for the decision making you’re gonna make.

So when looking more at a migration from just a a pure kind of road map and execution perspective, when I’m looking at putting together some sort of a plan, right, I’m I’m a CDO, and I need to understand kind of the before and after.

What what do those plans typically look like? I mean, if I’m building software, I’ve I’ve got a plan to build the software. Right? I’ve got I’ve got requirements to build the software.

But if I’m making one of these large scale migrations and I’m thinking really about that scalable data infrastructure that you that you were just talking about, is there such a thing as, like, a cloud migration plan? What do I need to be thinking about?

Absolutely. There absolutely is a cloud migration plan.

I I’ll use another analogy and example. Let’s say you’re a home builder and you have a lot of property that’s vacant and you’re gonna build homes on it. Unless you’re, you know, uber rich, you’re probably taking a loan from a bank and the bank loan is gonna have an interest payment that’s due every month. So your job is to build that property fast and quick.

Now you can’t build a house on a property just because you want to, you have to get contractors, the lumber, all the supplies that needs to go. So you have to sequence it in a way that when you start engaging building, you’re ready to go because you cannot have a couple of months of gaps where you’re paying the bank but you don’t have a reward because you can’t sell the house. Same thing applies to cloud and cloud foundations for the journey to the cloud. If you’re not ready and you fire up the cloud and you move one petabyte of data and then you sit around for three months because you’re not using that data, that petabyte of data is gonna cost you money in the cloud.

So you have to have a framework. So almost all hyperscalers have a framework of the journey into the cloud. And what what the providers like yourself and us are focusing on is like, you don’t have to do all of it in day one. Let’s figure out how to decouple the things that make sense.

Marketing data, maybe that goes to the cloud. Maybe payroll data doesn’t go yet. Maybe ERP and some of the complexity of those stay for a while. But you don’t wanna say, I’m gonna need, you know, twenty petabytes of storage in the cloud.

Let’s go buy it right now, but not use it. Your, you know, your cost to those hyperscalers is still going to stay. You’re going to have to pay them. So you want to have that journey map, foundational map understood.

And the one of the catalysts of this foundational journey is security and governance. So if you haven’t done that bit of homework, then you know what? Pause, don’t go to the cloud, do governance and security because guess what, it applies to your internal data center as well. We just call it cloud which means somebody else is taking care of a lot of the management and operational part of it.

But if you run your own data center, you should treat it like your own internal private cloud and should have the same level of security governance and capability. You can’t just take shortcuts anymore.

So that that seems like there’s a bit of a tension there.

I don’t know. I’d I’d love to pick your brain on this. In that in that it seems like there you could make a reasonable argument for big bang, but we all know that big bangs are high risk. Right?

It seems but it seems like you could make a reasonable argument for a big bang. But what you’re saying is is is do incremental, take it bit by bit. Where’s where’s the balance there? I mean, of the is there anybody trying to do a big bang?

I assume Oh, there are there are mitigating circumstances that will make you.

For example, let’s say you’re a you use a colo and your lease is due, and it’s gonna be done in twelve months. You have an option to release renew the lease, but, you know, these people are not gonna renew it by the month. They’re gonna say sign up for another five, ten year contract. Right?

You may look at it and say, you know what? This is not the right way to spend my money. So you may have a mitigating circumstance that says I need to exit the colo, or get rid of my building data center because I’m moving buildings and I’m not going to renew the list. Whatever it is, you may have to do that.

And so it’s not an overnight change. You might have, like, six to twelve months to do this. So you have to exit by a certain amount of time or you have penalty at that point where you’re actually paying the, some kind of a lease, problem or you haven’t gotten all systems out. So yes, there are people that will do the big bang approach.

But the big bang approach is still over a duration of time. But it is in the context of linearity very fast and quick.

The people that do small approaches to the Cloud are the ones that are actually coming from the perspective of we need to do things differently. Let’s try it out in a way of acquisitions, mergers, let’s look at new product development. So a lot of the software based organizations that are really releasing new product scalability, trying to have a cloud SaaS mindset will probably do it as a net new capability and not not, immediately divest. Great example of this is banking and finance.

Right? They have legacy systems that are very efficient. They’re not just going to pivot out of it overnight, but they will start adding features and functionality that will scale to the cloud for the fact that as those things age out, they will not have the same problems. Where the focus becomes it’s not the cloud.

The focus becomes on your operations and management and sustainability of those things. For example, if you’re a dev shop, if you’re a software developer, if you don’t have good mindset of developing an agile framework, then you need to spend a lot of time on that because you may not now need to have multiple court sources that need to be published and launched and released. And if you don’t have the ability to move a bit fast, you’re going to be in a challenge. Think of, you know, not too long ago in the news, a major, cybersecurity company had a hiccup and it put a lot of people out of work for a while in the context of the systems of operating systems.

Well, you could have had cloud, you could have scaled for the cloud, but you you would have been impacted the same way regardless if it was on prem or in the cloud. But if you had agility and you had a framework that had decoupled a lot of these moving pieces, yes, you may have still had impact, but you wouldn’t have been disastrously as bad in that context. But that framework is not cloud. It’s not a cloud provider and a hyperscaler that’s gonna walk in the door and say, we solved it for you.

You actually have to invest in solution provider, partners, strategists, and build that culture, which changes that way of approaching cloud adoption.

So putting on your your consultant hat and and and looking back at some of the larger projects you’ve been involved in that maybe had a a couple hiccups. I’m sure everything wasn’t always perfect and not everything is was a hundred percent executed. What what are, like, two or three of the biggest kind of lessons learned, gotchas that you would that you would say to your clients? Right? All else being equal, avoid these two or three big bear traps. What would they be?

Well, there’s a a contextual bear trap that I always ask my customers not to get as accustomed to. Cloud is not necessarily cheap.

Mhmm. So don’t assume just because you’re on the cloud is gonna be cheaper. Alright. With the ramp, with the hockey stick, still not cheap.

What the cloud gives you are the things that you probably cannot do today, so fixate on that. Agility, fast speed, internet, access anywhere, redundancy, reliability, resilience, all of those things are things that you don’t want to get tripped up on. So when you think of technology itself, there are certain type of apps that are not very good candidates for the cloud, so they get tripped up on it. For example, if it’s a very data intensive chatty app, you probably don’t wanna go to the cloud just yet if your user base cannot do the same level conversation and capability for access.

That’s why a lot of ERP systems tend not to go to the cloud because they have not really engineered the product to take out the chattiness of it. Right now, we live in a mobile first world. A smart smartphone and capability is a big deal. If you’re writing apps, even for the cloud, that are not mobile friendly, adoption to probably doesn’t happen, and then you still have challenges so that that cost, mindset becomes a challenge.

So when you think of the biggest hiccups we’ve had is understanding the cost footprint of going to the cloud. It’s not apples to apples. You have ten servers on prem, ten servers of the cloud, great. Here’s your cost saving, not so.

And there are a lot of things that you need to understand. What if out of those ten servers on prem, only two are production and eight of them are development boxes? Guess what? The cloud is production a hundred percent.

We provide the service so that it’s always available to you. If you did not classify that workload properly, now you’re spending money for things that you don’t use. And so we have to optimize that. So those are the biggest roadblocks hiccups we’ve run into, especially on large projects where you take years to get there.

And like year one looks great, year two looks bad, and another, the C suite is wants to kill the project and kind of bash it entirely because they’re not happy with the cost savings or the lack of it. And so you are not in a very precarious problem. You have half in the cloud, half not, and may you, Ford, decided to not fund it anymore or decided you to do differently. It it is a challenge.

So those are the hiccups we will look out for. What is the outcome you want and what is the purpose you want? I had a real example of this, when I was in consulting. We had a customer that says, hey.

I wanna go to the cloud. And being the technology is how it’s all about it. I provided the best example of how to get the cloud. We implemented it.

And then few months later, when we were trying to follow-up to figure out if they were satisfied, the the person was extremely dissatisfied because their cost per portfolio went up by twenty percent. And they’re like, hey. You know what? It might it cost me more to run-in the cloud.

And one of the things that we failed to ask is why did you wanna go to the cloud? It turned out this organization, this small organization, had an exit strategy. They were looking to be bought, and they were trying to show them that they had good cloud mindset and capability because they wanted to put a footprint that looked better than usual. Had I understood the outcome, if that’s what they wanna do, we would have pivoted in a different way for the cloud journey.

You know, if the cost savings, reducing the EBITDA, for example, was the way they wanted to shop around for a buyer, you should have told us that because, you know, technology is not cheap always. So understanding the outcome is paramount to successful cloud journeys.

So a lot of what you just said suggests of that maybe you don’t understand your internal cost drivers. Right? There there could be some technical debt, systems debt. Who knows? You’ve got processes running, consuming compute resource that you didn’t even know necessarily. Is that is that something that you see?

Absolutely correct. We see that all the time. And we even see it at the granular level where you, on your day job, are very involved in data. Just because you have data and you don’t know what it is, it doesn’t mean it goes to the cloud.

I’m looking. You can go hundred percent the cloud, but you’re wasting space in the cloud. So you know what your data does, know what it is, know what is classifications or use the appropriate cloud service. And every hyperscaler has a tiered capability in this data space to give you the best bang for your buck.

So if you don’t know what’s going on on prem, it’s not gonna help you in the cloud. So, yes, yes, not knowing the data state itself, forget the asset, forget the inventory of actual physical, you know, servers and whatnot.

You have to have your asset knowledge of what the data footprint looks like. You have to have the asset knowledge of what the apps look like It just fails. We have horror stories of customers that says our entire x y z to the cloud only to forget that there was a major component that was integrated on prem with a provider that we know and knew about because it was an API or a connector and now everything’s broken and likely a most business critical function. This actually happened a few years ago where the payroll system was migrated to, Equivalent Thoughts system but they forgot that their integration to the bank to cut the checks for the employment employees, paycheck delivery was old school dial up that still was there.

And one no one realized was that bank was archaic technology, and they were using a modem to dial and send the zeros and ones over a regular telephone line so that that payroll file was uploaded and the, you know, transactions happened. So things like that, you forget that there might be pieces that you didn’t evaluate. Now there’s a lot of vendors that will do kind of a automated way to figure out your dependency of your architecture, dependency of data, dependency of solutions, and whatnot, but it still requires a human element to go in there and do it. So if you were starting from day zero, document everything, create flowcharts, architecture diagrams, have comments in your code, otherwise, you will always suffer from that.

And these are just evolutionary technology challenges we’ve had for decades. Like, commenting code an app, you know, making sure you have a good inventory system for your software, that’d be GitHub type solutions, or knowing what your data state looks like, security and governance around that. There’s so many things that we don’t do. I mean, ITIL lives for a reason.

They have all these, you know, configuration database mindsets for just that purpose.

Yep. So do you find any any of your clients doing a, like, post migration, cleanup? Right? Where they where where they where they where they migrate, and then they find all of this stuff is like, oh my god.

We had no idea that there was two hundred terabytes of whatever. Day data data that was connected to an application that is no longer in production. Who who knows? Right?

This could be a thousand examples of this, but do you find with the, like, all of a sudden going do. Okay. Alright.

Yeah. So the one of the things we notice is they do it as it’s happening. It’s not after the fact. Somebody will be like, hey.

We gotta move a petabyte of data, and somebody in the decision making authority will be like, petabyte. I’m like, I thought this app only required, like, five hundred gigs. I’m like, where’s the petabyte coming from? And then you realize that there’s data that was there that you just assumed it was part of the database.

Oh, yes. Storage, backup, multiple iterations of copies of this data environment that you did for whatever purpose. Well, don’t move that. So yes, sometimes they will say, okay, we have a petabyte for associating this app, but as as the app goes to the cloud, we’ll just go get it with the five hundred gigs and the rest will be on prem and we’ll have some kind of archiving solution that will now pivot back.

So they’re almost like the saying goes, building the plane as they’re flying. So a lot of the migrations will actually change as they’re happening. And if a good consulting organization, good integrator gets in there, they should capture that capability ahead of time for you to realize. We also have on the fly code changes.

For example, you may have a version of a software that thought it should have been updated but you didn’t. So minor change, update it as you go on the fly, do some UAT and then you’re ready to go. That is also a very uncommon way of doing it, but, you know, there’s inherent risk on it. And some of this risk can be challenging that requires a little bit of foresight and planning.

That’s why, you know, consulting organization integrators are, sufficiently busy because a lot of these good decision makers will say, we’re not gonna go to the cloud just by the, you know, pulse check. We need to figure out what that looks like by knowing what we have. And you spend time and money there, then you have a good footprint of how you journey to the cloud.

Awesome.

So I assume that there will often be some motivators for people to okay. Well, wait a minute. We gotta take stock before we do this migration.

Sounds like, again, it’s a bit of a balancing act between how much is the cleaning up do you do in advance and how much do you do it after the fact. Sounds like a decent amount of the after the fact stuff just happens because you find things you didn’t know about.

But it sounds to me like as as as a CDO, you need to be prepared for post migration to be doing at least some sort of house cleaning is probably a natural thing.

Absolutely. And one of the things you got to understand is, look, if if if a hyperscaler or any organization that’s in the business does it right, they should have, almost monthly, reviews or quarterly reviews to say, hey, this month you spent x thousands of dollars. Nothing has changed when we’ve seen an uptick.

Therefore, you’re adding data or you’re taking data out or whatever the conversation is. We should always be striving for optimization and that’s a value add a CDO, CIO, CTO always wants. Like, hey, you, let’s talk about, hypervisor to VMware type products. Just because you had a hundred servers on prem, you may not need a hundred servers in the Cloud. So yes, let’s move a hundred, but immediately let’s start optimizing and figuring out maybe we can bundle all these into one massive machine, therefore your lesser footprint, I guess what it saves money, cost, licensing, everything, right? So we definitely, work in this mindset of optimizing as we go, either in a frequency or annually, depending on where you are. And there’s products already out there that will take a look at it for you and tell you this is where you’re spending your money and you need to.

Yep. So let let’s let’s talk specifically data and AI.

What are some of the top trends that you’re seeing from the perspective of a focus on analytics, a focus on AI? Are a lot of people still excited about using the cloud specifically for AI? I assume the answer is yes. A year ago, it seemed like everybody was talk getting on the AI bandwagon. Maybe that that that has cooled a little bit.

But what are some of the trends you’re seeing more in the data in data analytics?

Trends we’re seeing is, there was a trend that said we need to do analytics and insights in the cloud or anywhere, and we the cloud seemed like the place to do it. And so that maturity level has reached a peak where they’re like, hey. You know what? We’re doing very good with it, but what else can we do?

Where else what else should we add? I’m like, I use the pandemic, and I use weather as a conversation. What if your business is depending on that? Do you have analytics that is tied into a weather database or data capability that gives you that, unstructured data for you to go do those things?

So what is happening right now, the trend is, like, we can do analytics fairly well. We can also visualize it fairly well. But we need to add what if scenarios in this. So what if scenarios is not a, hypothetical conversations with data scientists or business analysts.

And so now what’s happening, the trend is saying, hey, we want to use analytics in the cloud with this kind of an AI mindset or a generative AI mindset, but we can’t do it unless we have all the data there. So that’s the pumping of the brakes is the fact that a lot of the commercial enterprise data is already in the cloud or is fairly well established, but we need to add other components to it. So that that’s the shift we’re seeing. A lot of people say, hey, AI works.

We get it. You know, we’re sold in some data platform in the Cloud. We love it, but we need to add more context to it. So what we are seeing as a business to business ecosystem propping up where there’s data from different sources that you’re connected to and providing a level of success from that perspective as well.

Airlines is a great example. Imagine if you are in travel agency having airline data and traffic pattern data and delays of weather pattern. If you put it all together, you can do better what if scenario planning if you’re trying to book reservations otherwise. Today, you can just book a hotel, not realize the hotel’s under, you know, water because there’s a flood.

Imagine if you had that information, you’d have better insights, but you would use generative AI or AI to help you guide that. So a lot of the the data leaders are asking, hey, we wanna do data and AI in the cloud for a couple of reasons, but at the same time, we wanna do what if scenarios by adding other pieces that are not really core to our business, but this is the best way to start doing it and start believing in the fact that we can have insights that will be differentiating them in the marketplace.

So so I’ll I’ll make an assertion, and then you you can either support or refute it. But I’ll I’ll tell you something that kind of high level observation that I’ve seen over the last couple of years. There’s always been, at least in my world, there’s always been a little bit of a distinction between analytics and data. Right?

Analytics, lakes, warehouses, BI tools. I I would argue to a certain degree that from a cloud perspective, there’s many people kind of already get it. Right? They get they get the scalability.

They get a lot of the flexibility.

On the data side of the house, which leans more towards data management, things like data quality, MDM, data integration, data governance, data cataloging, those types of capabilities. I would argue we’re kind of playing catch up. But what I’m seeing is a lot of the hyperscalers moving towards that world of of of data management. We’re certainly seeing on the Microsoft side. Are you seeing this as well?

Yes. We are. And the reason you see both high all the hyperscares thinking about is because of the massive gap. So think of the conversations I just referred to.

We will take you to the cloud. You’ll be happy with it, but then you come back and say my insights are bogus. So where is the problem? Is it my analytical engine?

Is it my, superior database engine? No. It is the fact that the data was incomplete or data did not have what you expected it to be because you thought it did, but it really wasn’t. So data governance is a gap in the cloud.

People understanding what data they have, classification, catalogs, all these are very relevant. So, organizations like yourself, your employer, you you’re in a good place. But the reason these hyperscalers are stepping in is not because they wanna take your lunch money. It’s because there’s a gap.

And there’s no one’s figured it out. They need to solve it. Now what happens with the hyperscalers that are uniquely brand oriented, they were gonna do it only for their little patch of it. Right?

They’re not gonna do it for all the data that is out there coming from different sources with different capabilities, especially classified data that requires a classification that is based on a very unique identifier. For example, you and I, spent a lot of time here in oil and gas, capability. There is a unique identifier for a wellhead that is gotta be ubiquitous across all the different platforms and industry, including state government agencies. So if you have classified a classification of data not tied to some kind of index or some key, something that really represents that, you’re gonna have a challenge identifying the data or may have replicated data that need not be there because you did not do that.

So you spend a lot of time in ETL. You’re spending a lot of time transforming the data, conforming it in a way that ultimately does not give you the end result that you want. So that governance need is paramount.

Somebody that can do it for many clouds or different clouds or diversity of ecosystems is gonna win that conversation. And I I comfortably can say not all hyperscalers will solve it unless they were in the business of that prior to cloud becoming a thing. Like if they had a portfolio of software that does it, therefore they’re just going to enhance and make more of it. Most of them are reinventing the wheel from the ground up and some of them are just trying to figure out for their own purpose.

For example, a customer may want to know how they are getting charged for their cloud consumption without the classification of data and the governance and MDM. You may have an invoice that says data. How about wouldn’t it be nice if we could tell you what the data was and what it is so that you in your invoice could see where it’s coming from? That requires level of classification and, management that the the hyperscaler might need to have to do just to be relevant to the industry and business.

Yeah. I see it’s gonna be an interesting few years to see how things kind of shake out on the on the data management side and where where things are going because I I certainly see the hyperscalers getting more active in in that space. We we we have partnerships with with all of the hyperscalers.

But like I said, it’ll be interesting to see how how much of the market for data management software the hyperscalers themselves, tend to tend to subsume. I don’t I don’t have any magic answers there, but there are there are you know, I just look at, for example, a Purview by Microsoft. They’re investing a lot in Purview as as a data catalog, and that makes sense. Right?

Like, to to get the full view of everything, it gets sense it makes sense from a management perspective, but even just from a billing perspective, like you were talking about, like, understanding what’s out there. Right? So Yeah. It it it makes sense.

It doesn’t doesn’t mean that there can’t be others that are also building best of breed solutions that’ll work across all hyperscalers. And I think that there’s certainly value there. Well well so let’s talk about all of the hyperscalers, and let’s talk about kind of multi cloud. Right?

How many how many of your clients are running on multiple clouds, and how do they address that?

Almost every one of them. Right. Well, let let’s pivot into the fact that I work for Google right now. Right?

Google is, the third entry into the cloud space in AWS at the drop. Azure caught up. Google is a third. All my customers have cloud that is not Google.

Either it’s AWS, Azure or both. So all the customers are multi cloud and so the multi cloud strategy is not going away. It’s going to stay. There is not a compelling cost differentiator that says I need to have one cloud solution.

Most of it is tied to some level of hyperscaler discount that is so amazing that they’ll say, you know what? It makes no sense for us to move over. And I and the kind of pivoting back a little bit to the conversation you have with Microsoft Purview. Think of it this way, Microsoft is charging you for that Purview use.

Of course.

Most of our customers will say, I already have my data in the cloud. Give it to me for no cost. No one’s gonna do that. But if they’re gonna spend money on it, they might as well get the best of breed.

Right? They may not just say, I don’t need Purview from Azure. I need a that keeps scrub everything and multiple systems. And then there’s another thing that’s, evolutionary wise in the cloud space is becoming more relevant today in the global footprint.

Sovereign clouds. Lot of countries and municipalities are coming up with rules that says you have to have data in a certain place or location. Well, not every hyperscaler has a data center in every city for you to go consume their service. Sometimes there’s a drop with Microsoft, sometimes there’s a drop with AWS, sometimes Google, maybe Oracle, maybe IBM.

Yeah. We’re multi cloud solution now just to have that sovereignty conversation enabled. And therefore, you need a tool that is going to be able to talk across all these capabilities in a way that achieves that data governance, clarity, visibility that you need because you are tied to your geography, not because of a cloud provider because of some rules of, sovereignty that dictates that. Furthermore, add, sensitive data, health care data, social security number, you know, financial data that is very unique.

And your CSO might have a position saying that I need this data protected in a way that we don’t get caught, you know, unaware if we get hacked or compromised. So they want to know where the data is, what the data was so that if there’s a hack, they know right away what that looks like. And so that’s happening as well.

Okay. So it seems like there’s an interesting opportunity there, for for companies like us to be a little more multi cloud. I mean, we we we already are.

But for things like MDM and data quality or data catalogs, I mean, you have to host your infrastructure somewhere. But at the same time, you could be building connectors that have seamless integrations into anything. I would argue maybe that’s kinda what Fabric looks like. Are you familiar with with Fabric?

I am familiar with Fabric was, like, did come over from Microsoft recently. Yeah.

It is familiar in the context of what you are trying to do, but think of it for the purpose of the data itself. Right. Yeah. A lot of industry data are getting, standardized.

They’re joining forums or some organization that says, let’s standardize this data so it’s easy for us to share it and use it. When you get to that point, what you’re now talking about is speed of data transfer, use, and analytics suggested. You don’t have to reset the data in multiple clouds or use it in that context fabric. It’s kind of allowing you to do that where you don’t have to have the data in the same place, but now you’re compromising speed and agility depending on what you’re looking at.

So you, as a vendor of a product, should develop it in a SaaS model that works in all clouds, now have a magic sauce behind the scenes that says you can be in any cloud and I can host it in your cloud or other cloud, and we can work across, obviously, the data governance as uniquely, class classification that allows you for that transport of capability. Otherwise, you do the transformation. But, yes, it is, a fabric from Microsoft is a compelling solution, a capable solution, but it does not really cover the entirety of it because you have to step up to it. Right?

You can’t just say I’m gonna have old school SQL on prem and I just don’t want fabric. There is a level of challenge and modernization that has to be done. And guess what? Someone like you will be very busy because if you have on prem data, you still didn’t know what it is.

If you should tell them to go to Fabric.

Yeah.

So one last question.

There seems to me to be opportunities here within each of these hyperscaler clouds to allow for levels of of intercompany data sharing that isn’t really happening today that I’m that or at least I’m I’m not aware of it. When I was at Gartner, I talked a lot about data sharing and and opportunities for data sharing and even things like shared governance and shared management of data. Because when you when you look at companies’ data, they if you’re b two b, for example, one company has a record for Verizon. Another company has a record for Verizon.

Another company has a record for Verizon. And you know what? They all look largely the same, and they’re all managed the same. They’re all stewarded the same.

The good data quality rules are generally all the same. So I’m I’m curious to get your to get your thoughts on do you see the use of of or do you see more adoption of or at least people talking or asking about data sharing in these instances?

They are absolutely. There’s two reasons for it. I don’t wanna own data that is gonna compromise me in the security footprint. So either that you keep your data where it is, and we’ll just share a little bit of it as needed.

Think of the federated security model, right? Many portals, on the web and apps require you to sign in with an identity. A lot of these people don’t want to save that identity on their own system, so they only use federated systems like Microsoft or Google or Facebook as the identity capability and then use business to business connector to make it happen. So you will find vendors that will say, look, I I don’t need that data because then now I have to own that data from a security perspective.

Let’s not let’s just connect. Let’s share the data in a way that’s common. Therefore, you invest in those organizational changes that make the data, reasonably shareable in the context you want. But there’s also an IP story.

What I mean by that there’s some data you don’t wanna share because you wanna have a competitive edge, and then sometimes that will just be as it is and never be crossed, and there’s also a litigication problem. There’s a lot of people should be feel that if they share data that they did not comprehend and now they’re litigiously tied because something may have some fine print and some uvula that says you can’t do that. So there’s always that conversation that has to be resolved, but it is trending in the right direction. For example, with all the general AI programming and learning models, a lot of people are complaining, oh, you took my book and wrote it or you use my YouTube content to create it.

You you will reach a place where even the sharing of the data needs to be compartmentalized into what it is and what what it is not. And therefore, that journey will be a little slower, but it is happening. It’s getting rapidly, capable, in a way that even some data shared between competition is useful. And I I there are only a half a dozen players in the world that can monitor the world for cyber threats. Guess what? When one organization gets impacted, they actually call their competition and say, hey, did you guys see that same thing happening?

Yep.

To verify that there’s attack footprint happening. And a lot of these are denial service attacks that happens around the world from enemy nations, those security providers that are in competition in the enterprise space will actually talk to each other to figure out if there is a pattern happening. There’s gonna be a level of data sharing also in that space that you will do it for the better good.

It’s interesting you should mention that because when you were talking, that’s exactly what I was thinking about. And when I was at Gartner, I actually used the network intrusion use case, like the DDoS use case because there is widespread sharing of of the footprints of each each of these intrusion attempts across major providers because there’s benefits to to all from from that. But I I still I I still think that that’s we’re gonna see more and more of that because I think there’s a value add there when it comes to just basic stuff like data stewardship. Right? Like, updating records. Right? Like, fair fairly low value add things that a lot of companies are spending a lot of money on.

And if you could put in a mechanism for compensating one company for maybe updating a record where everybody benefits from that, I don’t know.

I I think there could be something There there is there’s a very in my mind, there’s a variation of this already happening with the credit rating agencies.

Right?

They’re Exactly.

Right?

You you get a bank loan from one that’s, got the rating from one, they will share it with all three of them because it’s gonna be on your, shared record.

So somebody’s updating it collectively.

Better governance, better capability will help us do all those things like your home address, your, you know, email address, subscription for mail magazine, whatever it is. Things that you forget that, if one person changes, they should share it, in a ubiquitous way. I’m like, the USPS mail system does that. Like, if you forward mail, they’ll actually tell the sender, this person has moved, and they have digital ways of sharing that data now so that the sender now updates their database as well, saying that this person has moved with this new address they reported in.

The credit use case is an extremely valid one. Yeah. It’s it’s a give to get model where a lot of companies are literally sharing their AR. Typically, they’re doing it through Dun and Bradstreet, not exclusively, but what they get back are the credit ratings.

So another great example. I’m in thank you so much for the conversation today. This is this is fascinating. I learned a lot, about what to avoid from the perspective of a cloud migration, what to expect from a cloud migration.

Where where can people get in touch with you, learn more about you?

I am, big on LinkedIn. I live on that as well. And so if you, are able to share through the content of my LinkedIn address, you can take it, reach out to me. I speak to anybody that has time for me as well as vice versa. I’ll make time for you.

Get get a hold of me through that. I fairly live on that in the context of social broadcast as well, so I do have point of view. Sometimes I will share. So please, share that information. I’ll be available to anybody that cares to learn more, talk more, and understand, how things are happening.

Awesome. I knew you were active on LinkedIn. That’s why that’s why I mentioned it. So anyway, to our listeners, to our subscribers, to our followers, thank you so much for being a part of this growing community of data practitioners and leaders.

I’m Malcolm Hawker. I’m thrilled you were able to join us today. Stay tuned for another episode of the CDO Matters podcast sometime very soon. Ayman, thanks again.

Thanks, everybody.

Absolutely. Great. Thank you.