The Power of Data Podcast
Episode 69: Dealing With The Different Dimensions Of Data
Guest: Irfan Khan the President HANA Database & Analytics, SAP
Interviewer: Sam Tidswell-Norrish, Advisor, Dun & Bradstreet
Hi, welcome back to The Power of Data Podcast. Today I am joined by Irfan Khan the President HANA Database & Analytics at SAP, it's great to have you on the podcast today Irfan, how are you?
Sam I’m doing very well and it's a pleasure to be here today.
Fantastic and if I’m not much mistaken you are based just on the outskirts of London, in a small town of High Wycombe which at Dun & Bradstreet we know well, just next door to the former HQ before we moved to London, am I right?
That's absolutely right Sam, we're practically neighbors or were practically neighbors I should say.
Yeah, we were although nowadays everyone is in their home office, somewhat irrelevant where the office is. Although we are looking forward to moving back to Paddington. Irfan to kick off let's talk about you for a moment. You've been in the technology space for over 25 years and many years at SAP. You've done a number of different roles within that technology space including sort of leaning into the sales space as well and you've done it in a number of regions around the world. Can you give us a little whistle-stop tour of your career so far?
Yeah certainly Sam, happy to do that and maybe not the most conventional career path I must confess. The one claim to fame that I have actually only given one job interview in my entire working life. I joined a company called Sybase that was actually based in Maidenhead not too far away from Marlow and from where the Dun & Bradstreet headquarters used to be. I joined Sybase in the early 90s straight out of university and predominantly in my first I would say 10 or so years working there was really in development type of roles, so software development software engineering type of roles. That essentially culminated to a point of the first maybe 12 or 15 years where I started getting into much more of the scope around evangelism and then becoming more part of the strategy part of SAP or rather Sybase at the time and then eventually became the Chief Technology Officer. And that was also short lived because in 2010 SAP acquired Sybase so hence my sort of one interview in my career because whilst I’m at SAP and it was through the acquisition part. So in a short time having arrived at SAP, realizing that this is a huge company you know something in the order of 100,000 employees compared to the 6,000 that I just come from and that really gave me an opportunity to maybe accelerate or diversify a career path. I was always interested in maybe pursuing more on the on the sales side but I hadn't necessarily held a sales quota before which is a little bit odd. I’d never really run a sales region but I was thrust into a global role in sales so I was actually responsible for the database portfolio initially for SAP then that sort of extended into a broadening of the portfolio into analytics and then for the last two years I was responsible as the president of what was called platform and technologies for SAP which is essentially all technology sales including database and data management and analytics. So, a very unconventional path but nonetheless I’m now back into development again so we're happy to be back in product again.
First off congratulations on the new role that was just last month if I’m not mistaken?
Absolutely it was first of February that was my official start date on the new role in development organization again, correct.
Fantastic and you moved from the technical side to sales so now you're back to the core dev part of the organization and what a time to be in the core development side and particularly data and analytics. How’s your perception of data shifted over time and particularly as it relates to where we are today and this sort of digital singularity that we find ourselves in over the last 12 months?
You know Sam, I mean if I look back over the last probably 10 years or so at my time at SAP, I mean the one sort of major substantial movement of SAP has been through a substantial number of acquisitions that they've performed, we've gone from having somewhat of a homogenous foundation and applications to becoming highly heterogeneous. So that's pretty much in line with a majority of customers are I mean if we take a look at the journey of every customer right now they're dealing with data on so many different dimensions meaning data that lives within of course their different environments whether that's on prem or in the cloud but that's the obvious thing. But what we're also now seeing is that the rise of data from unstructured sources coming from the edge as well, meaning that you have a lot more richer environments where that's not just driven by IoT, but even in terms of looking at classical things that you would have just assumed would have had a larger data sort of stockpile over the years like historical data in a financial services environment what we now find that data is now being pushed to the edge. You have a lot more value association with data in the more of the broader sense and in more of maybe in the horizontal sense as well rather than in one centralized store. At least from an SAP perspective I mean the last 10 years SAP spent something close to £70 billion acquiring probably a dozen or so SAS vendors right and those vendors include companies like a rebirth success factors many others that are part of that that organization that now is SAP, so therefore we see a huge proliferation of data and we don't just see it in one environment we see in a highly distributed landscapes across the globe.
Well, we know SAP Ariba well and our partnership with Dun & Bradstreet and extends to all sorts of areas of the SAP business including the SAP partner edge build program and an SAP Ariba partnership as well. But in research ahead of meeting you today I was looking through past talks that you've done and documents that you've written and you're often talking about the data value formula. Can you tell us a little bit about that and why it's important?
Yes, absolutely. I mean, the data value formula is not necessarily something which is new. However, the imperative, as you said earlier on the imperative to make sure that we have a far better management of data as a starting point is incredibly important now. So we look at data and first and foremost, if we take a look at the data that is being developed in, you know, maybe through the support of machine learning type of foundations, that's incredibly important. But then there's human creativity that also needs to be weighed up into this as well, because machines themselves are learning against models that we create. And therefore, the creativity element needs to also be somewhat, you know, described there as well. We also take a look at the majority of organizations now and the richness of data that exists, we also anticipate that they would want to have a free-flowing experimental sort of desire across all manner of different data sets out there as well. It's important that they can look at different scenarios, whether they want to look at simulating different outcomes, or perhaps even hypothesizing around what they could actually even see or infer from data coming from different environments and sources as well. So if you take a look at the data value formula, first and foremost, it's looking at the breadth of data. So the volumes of data. It’s looking at trying to create a level of quality and value associated with that data. So this is now not just through the human creativity part, but also now supplemented by the machine learning and the AI that may be out there as well. But net of it really is, is that most customers want to see a much more accelerated path to data to value. They don't want to be going down a path where they spend significant amount of, you know, capital expenditure building out data centers, only define themselves that they see first a cost and not first a value. So this is why this data value formula that we have been using, effectively takes into consideration the volume of data, the desired quality that you wish to derive. And then of course, the different usage scenarios that are also there as well. And if you multiply these vectors out, then ultimately that will give you a value association with the underlying data that you're building out. So in simple terms, it's a formula given the volume times quantity, times the usage, and it will give you the value of what you look to choose.
This is music to our ears, because obviously, this, as one of the world's oldest commercial data businesses, is something that we speak to our customers about every single day. How do we derive actionable insights through our solutions using our core data set? And the reality is data has been logged in every single aspect of our lives and ultimate data sources are prevalent in every corner of the world. Now, everything we're doing is being logged, but it's all data useful?
Well, interesting you should ask about is it useful? I'm sure it has a usage. But whether it's useful or not, at that point in time, it is sometimes debatable. You know, I read a study by JP Morgan Chase recently and they used an airplane analogy. So airline passenger analogy to be precise. And if you think about, you know, passengers on an airline, they typically fall into categories of either economy class, business class, or first class. And if you take a look at the accumulation of passengers over time, whilst the airlines have made substantial revenues from the first and premium class passengers, they make a substantial level of value, and of course, profit from the economy class passengers, because in aggregate, you know, they are of course, a very substantial source of revenue. And if you apply that same analogy to data, there is an abundance of economy class data, maybe that's the data that is sitting inside of maybe file systems, which could be point of sale data, it could be CDR data, in a telco environment, so customer billing type of data. So in aggregate, that data has a substantial value, if you can start looking at maybe trends within the data. So economy class data in itself is valuable only if you look at it from the trends perspective. But similarly, if you look at business and the first class type of data, this is where you'd want to run your operational systems, maybe a very substantial part of your transactional core of your business. So the transactions that have been performed, the sentiment about your install base and your customers, you know, that is all valuable insight. So really getting to the core of your question, okay, I mean, data has value. But of course, it needs to be brought into the relative sort of situational value of that data, and dependent upon how you will look to then process and manage that data, it should be able to give you implicit value, but you now go to put a lot more emphasis in terms of curation, the management of that data, making sure that as the right qualities, and therefore this now puts a lot more burden onto what would have been traditional it, but that, of course, is now shifting dramatically towards the cloud environments as well, so data in the cloud.
Yeah, there's a lot to talk about there. Firstly, you're absolutely right. The data does need to be brought in, process managed and Dun & Bradstreet obviously, one of our core proprietary assets is the D‑U‑N‑S® Number. The D‑U‑N‑S® Number is a unique identifier that is applied to every organization on the planet. And with that we have large master data management programs and data enrichment capabilities because you're right, you can't just have data, you've got to make sure that it's integrated and harmonized. And that's a massive theme in our industry at the moment data integration. Why is data harmonization so important at the moment?
Well, I mean, data harmonization, or let's call it synthesis of data, is so situational and so contextual to applications now, if you if you connect to any data source, and it's just free form, and maybe its schemaless, as a term would be technical term, you can sort of interpret that data in a variety of different ways, which is great. But if you harmonize our data, and you create some kind of a semantic layer around that and have it far more, I guess, from an intrinsic point of view or value, you make it a lot more situational to the user, than the likelihood is that the customer will see immediate value. So Dun & Bradstreet, so you've done a tremendous job and of course, from an Ariba perspective, you mentioned this earlier on, we of course, utilize quite a lot of the data assets coming out of Dun & Bradstreet. Specifically on the Ariba side, we use your Dun & Bradstreet foundation in terms of the world based files that you provide. So this gives us access to 300 plus million businesses that are out there. And that gives us context. So harmonization really goes hand in glove now with usability. And if you think about the user experience around data, people want to have much more data served up in time, which is pretty much contextual to the ask that they have. And making sure that it gives them more of a real time value, as opposed to maybe a legacy historical value. Because whilst that may be important, it's much more important to be informed in the decision making process that they're in right now.
I completely agree with you. And I want to come to some of the work that we're doing with SAP in just a moment. But can you give us a little insight into specifically what SAP are doing around integration and data harmonization? And what are you seeing from your customers as it relates to use cases?
Yeah, so data integration is one of those topics, which is dominating our CEO agenda and, of course, entire SAP board is firmly behind this right now. And to give you a bit of perspective, here, maybe if I may just take a slightly bigger run up to answering your question here. I mean, as I mentioned earlier, we've got approximately 70 billion worth of assets that have been brought into SAP in the last decade. So you can imagine that that has a high degree of challenge in terms of being able to then start integrating all of these different technologies and different applications and different course foundations together. And if you look at the amount of data, we probably manage about 300 petabytes worth of data under these various SAS solutions that SAP runs today. And if you take a look at the physical amount of data that lives across multiple data centers, that in itself becomes even a bigger challenge. So it's really about the integration value comes from making sure that you can drive a semantic layer, which we essentially are now referring to as a data plane. So bringing all of the data having one single, if you want to call it a harmonized layer, a single one domain model that allows you to be able to start interpreting the data in a variety of different ways. But having that context that I mentioned, right, the context is incredibly important. So if you look at it from for example, a point of view of an operational environment and say, manufacturing as an example, and you're taking a look at the current production run, in a manufacturing run that's taking place, you'd want to make sure that your quality assurance is being preserved. So therefore, integrating with quality systems, and taking a look at the legacy value of what would have been done maybe a year or three years ago, to ensure that you're not compromising that performance is equally important as well. So from an SAP agenda perspective, integration is by far one of the biggest drivers right now. And we're driving integration for not just for the SAP application suite, and of course, the application value, but for our ecosystem as well. I mean, Dun & Bradstreet are absolutely going to be one of the consumers of the data plane as an example of being able to abstract all of the different data sets, and being able to then start enriching those services using Dun & Bradstreet core services, you know, around those capabilities as well.
And when you talk about the ecosystem, are you specifically referring to the SAP Store?
Well the SAP Store is really the marketplace that allows the exchange to take place between SAP and the ecosystem. But if you think about the way, the majority of partners that have been servicing SAP customers, for the decades that SAP has been around, what they tend to do is look at the data in the context of whatever the application might be. Maybe it's in the ERP, maybe it's in the Concur travel expense environment, maybe it's in the success factors environment, looking at it from a talent management point of view. So what we're now finding is that the ecosystem can actually act as a glue that can maybe develop completely new applications, or maybe new services that will be able to bridge the value of the missed some of the gaps that may exist. Like for instance, if you've got a e-commerce catalog in Ariba, and you wanted to tie it to maybe employees and being able to make sure once they're working from home through this pandemic, that they can have acquisition of new office equipment, for example. So there's a link that you have to create between maybe employee central within success factors, and then perhaps the e-commerce catalog within Ariba. So, we see that the ecosystem could potentially fulfill some of those requirements. So it's almost like a natural extension of developing core functionality and more value rather than just servicing the data from the marketplace in isolation.
Got it. Well, let's now go into the Dun & Bradstreet partnership a little bit. And it's always fascinating to hear from such a valued partner like SAP, the type of value you're seeing from this partnership, albeit a slightly dicey forum to ask on a recorded podcast. But how do you see Dun & Bradstreet’s role in providing data value to your customers?
I mean, first and foremost, you're a very substantial partner of SAP. And if you take a look at the spotlight partner agreement that we signed in Q1 of this year, and we take a look at the natural evolution along the solution sets around Ariba in particular, we take a look at things like spend analytics, we take a look at the extensions that we have going on through our CX Sales and Marketing Cloud. I mean Dun & Bradstreet is playing a pivotal role within this and I would probably highlight the core hallmarks of what you bring to the table because you have such a high degree of pedigree in and around managing data and content and information and when you start bringing that level of prestige and value to SAP install base and our broader ecosystem over the past five years, it's fair to say that we you know, we've had an embed OEM type of relationship with Dun & Bradstreet which in itself is fueled some of the growth that we have, but equally so we're now looking to build out even more strength and value along the lines of the data value formula that I mentioned earlier on. And we see that Dun & Bradstreet core value add in that formula is going to be also be able to deliver even greater value to those end customers as well our mutual customers so to speak.
Thank you for that Irfan. When it comes to interesting new technologies I want to before we close out the podcast, I want to use this opportunity to pick an expert's brain. What are some of the more interesting technologies that you're saying impacting businesses like SAP and Dun & Bradstreet in our space?
It's an interesting question, because if you look at a lot of the innovation that's going on, I mean, aside from the fact that data is such a hot topic, right Now. We've got to be a little bit mindful about sometimes assuming that we're solving for tomorrow's problems right now by merely just adopting maybe one of the cloud hyper scalars. And, you know, moving a lot of the data at scale and at mass into those environments, because arguably, what we find the compute and the storage of data, the storage is more than likely going to be sitting in those hyper scalar environments. So some of the new technologies that we're seeing are probably technologies that are going to be more providing value at the edge. And a good example of this is think about maybe if you're a smartphone user, and you use Apple, right, you have an iPhone, I mean, the amount of logic that sits on the device itself now, I mean, if you have property, you're storing lots of lots of different photographs and memories on your phone, it's actually the machine learning libraries on the smartphone that generate your memories for you, right, they give you a little video, they give you the little montage of different video clips that they create for you using content that sits on your device. So machine learning is actually occurring at the edge. A good example also will be Tesla, right. I mean, Tesla in itself is, is effectively a computer on wheels. And it's using a stateful repository of data that comes from maybe a cloud environment for machine learning. So we're seeing data being pushed out to the edge. And having an integration point, which is more of a connection point as opposed to a pure physical integration point, we're likely to see a lot of technologies that will be evolving along those lines. I think we'll also start seeing probably some compiler specific technologies that will allow to optimize running on those specific type of endpoints as well. So in summary, I would say that we are seeing a lot of opportunities now for whether it's going to be upstarts or new organizations to be able to provide value, which comes from data at the edge, as opposed to just servicing data in a centralized store, which in itself will be very, very helpful for storing data, but not necessarily performing all the compute because as you can see, and probably gauge already, the majority of the compute that we're doing on our day to days is very much abstract and disconnected from a centralized store.
If we take a step away from maybe our day to day workspace, but think about some of the pandemic driven technologies or crazes that we're seeing at the moment, have you got any perspectives on perhaps the NFT boom, or the retail trading trends that we saw with GameStop recently?
You know, because the market is just so eager to want to firstly look for opportunities for expansion of scope, I mean, tech in itself, you look at the tech sector over the last probably 24 months, 18 months to be more precise, we've seen a significant valuation of the tech companies. And ultimately, what are they doing? Well, they're not necessarily doing anything new, but what they've done is they've crystallized their value because the pandemic is accelerated home working. If you take a look at, you know, e-commerce, you see the rise in costs, this is where Dun & Bradstreet is helping SAP on the CX cloud side, right, the Marketing Cloud, we're seeing a lot more affinity between partnerships being formed. So you know, if you look at it, really the sort of the big changes that we're seeing on the value chain, we're seeing, number one, we're seeing the rise of course of technology in general, but that's coming through alliances and partnerships that are becoming much more driven by formalized, go to market type of agreements and relationships as well, why it's such a privilege to be here with you today, Sam and talking to you through, you know, from a lens of SAP and Dun & Bradstreet. And we equally take a look at, you know, some of the larger organizations that are out there, the historical organizations that are out there, I mean, they're also having to now evolve. I mean, John Lewis just last week, so that they're going to have to incur a lot of unfortunate store closures, because the market is shifting, and people don't necessarily either feel safe, or even feel the urge or the need to go to a physical bricks and mortar organization, in a high street or elsewhere anymore. So I think that the pandemic has merely accelerated what was already going on in the market. And the likelihood is that we're going to see the formation of even greater partnerships. And we're going to see a strengthening of the go to market that goes on between organizations as well.
I completely agree. And that's part of the driving force for the launch last year of Dun & Bradstreet Accelerate, the partnership program. We really do see a world where partnerships and ecosystem economics really prevail and open up a number of opportunities that otherwise wouldn't be achievable on your own. Before we finish the podcast, I would like to know particularly in your in your new role, and as a senior leader, SAP, how are you thinking about team engagement and team dynamics, and trying to create some sort of normality in this new working world that we find ourselves in?
You know, there's a fine line between working virtually and of course, there is the incumbency, now, right to use video conferencing, as an example. But that sort of drives people to just sit at their desks all day. I mean, I for one, have got, you know, a number of direct reports that we end up having to do a number of one on ones on and one of those just subtle changes that I recently introduced was, rather than using Teams or whatever it's going to be Zoom or whatever, is just pick up the old fashioned, you know, cellular phone and call somebody on it. And the reason I do that is because it means the least you can stretch your legs and don't feel, you know, feel forced to have to sit at your desk all day. And I think a lot of people are feeling the pressure of the work life balance as well: how do you sort of segregate one from the other. And one of the things that we're trialing right now is a Friday-free-meetings. So that will give people the ability to at least have the capacity before they go into the weekend, to sort of clear their desk, have the opportunity to be able to wind down and ensure that they can have a demarcation between now I finished my work and I'm now ready to start my weekend. So Friday-free meetings is a good example of just giving some capacity back. So quality of work, and at the same time, the ability to really have the completeness of activities taking place as well. So there are a couple of examples. And of course, we're forever empathetic, right people having to look after elderly parents or family members or neighbors, you know, keeping an eye out for them. It's just opened up the world in a way that I don't think anybody quite anticipated that we will have to be as connected as we are but perhaps as disconnected as we feel right now.
I wholeheartedly agree with you Irfan. And yeah, it's true. I've never felt so close to many of my colleagues, as I do today, despite the fact that many of them I haven't seen for 12 months, and we are today that 12 month mark where offices around the world, and particularly in the UK shut, my fingers are crossed that we're not in this situation for too much longer. And most importantly that you and I have a chance to meet over a beer or coffee and yeah, to continue this conversation.
Certainly, and I'll hold you to that.
Thank you and thank you so much for your time today, I know how busy you must be in the new role. Best of luck. We are extremely grateful for your partnership and look forward to continuing to support as much as we can.
My pleasure, Sam, thank you very much and all the best for you as well.