Jonas Christensen 2:27
Tom Davenport, welcome to Leaders of Analytics. It is fantastic to have you on the show.
Tom Davenport 2:35
Thanks. Happy to be here, Jonas. Thanks for having me.
Jonas Christensen 2:38
Yeah, Tom. And this is special because when I started in analytics 15 years ago, you were already at that time considered one of the main, today we call it, influencers. I don't even know if that word existed back then, in the sense that we use it now. But you were one of the main influencers in the field of analytics. And I remember reading your book, competing on analytics back in 2007. And since then, you've written many more books, articles, research papers, and generally contributed so much to the profession. So, thank you for that on behalf of myself and anyone else who works in the field. Now, I've given a little bit of an introduction to you already. But could you tell us in your own words, a bit about yourself, your career background and what you do?
Tom Davenport 3:22
Sure. So, I'm a professor at Babson College, which is only a business school in the Boston area, known primarily for entrepreneurship, which is not my primary focus, but I'm a visiting professor at Oxford SaĂŻd Business School for several years now. I'm a fellow of the MIT Initiative on the Digital Economy, and I'm a senior adviser to Deloitte analytics and AI practice. And I've worked with a number of companies as an advisor. And I try to write or do a lot of research and writing on analytics and AI and related subjects.
Jonas Christensen 4:06
So generally very busy in the field of analytics, data and AI. How did you get into this space in the first place? Because when you did that, when you started in this field, you really were someone driving it rather than taking over the reins from those who've been there before you, if that makes sense.
Tom Davenport 4:23
Yeah, I didn't have much competition to start with, which was great. But in graduate school, I had done a lot of analytics work and I mostly paid my way through graduate school, even a little bit of my undergraduate training, doing statistical consulting, mostly to social scientists at universities. And I became a consultant and then a business school academic and I didn't really do much in analytics, but worked in other areas related to Information Technology management, and I was doing a lot of work in the Knowledge Management area. This was probably in the mid 1990s or so. And around the turn of the century, I guess, I concluded that Knowledge Management was great. But people were primarily focused on textual knowledge or in some cases, implicit knowledge in people's heads. But they didn't really pay much attention to knowledge derived from data. And so, I decided I'd start doing some work in that space and I did one research project. By this point, I was running a research centre for Accenture and I did one research project on this called ''Data to Knowledge to Results: Building an Analytic Capability'' and publish it. Nobody paid any attention whatsoever. And then a few years later, I started this project with SAS, primarily SAS, which was the leading proprietary vendor of analytics software at the time. It wasn't really called analytics so much. It was called statistics. But that project was to look at how companies were using business intelligence as it was primarily being referred to. And I concluded that some companies were really quite aggressive about it, and were really competing on it, and was trying to figure out a word for it. And I ended up calling it Analytics. And I remember the Head of Marketing for SAS said, ''Why did you call our analytics? Nobody uses that term''. But it seemed like, to me a broader term than just statistics. So anyway, it caught on. And I think people like hearing about what leading companies were doing, even though most of them were not leading at the time, and have been. I might have moved on to other topics, if it hadn't been so popular and successful. Usually, you know, if a topic loses interest on the part of the business public, I move on. But obviously, analytics and big data and AI have not lost interest yet. So I'm still with it 20 years later.
Jonas Christensen 7:12
I can tell you that one of the things that appeals to me in these areas also that is nascent and so evolving all the time, and we kind of have to create it as we go rather than work on something that's been around for a long time. It's part of the fun to make people see what we can see that this can do in the future, if we just get the right building blocks together.
Tom Davenport 7:31
Yeah, it's certainly a rapidly changing field, which I think puts a lot of burdens on people in the field, but also it makes it quite stimulating.
Jonas Christensen 7:41
Absolutely. And we're going to talk a little bit more about that, because today's two topics are sort of the history of analytics and how it's evolved. And then the leadership that's necessarily required to continue to evolve through that. And Tom, you have this term that you use or description, which is the Four Eras of Analytics. Could you talk about what they are and perhaps give us a bit of a description of the details underneath these four eras?
Tom Davenport 8:12
Sure. Yeah, so the first era, I really call it the Artisanal Era, because it's very slow and labour intensive and all human-based and involve small data and was kind of a back office activity, not very prominent. That was what prevailed when I started doing my own work in analytics and in graduate school. It continued for quite a long time. I'd say up until the turn of the century, maybe around 2000, when you started to see these companies in Silicon Valley, doing things with much, much more data than any of us had ever envisioned. And so that became the Big Data Era. And it was mostly practised by Silicon Valley startups. And it didn't really involve a lot of new analytics approaches. It was mostly data management approaches. That's when we started thinking about Hadoop and Pig and Hive and Python, and all these kinds of open source approaches to managing data and analytics. Not so innovative, I would say at that point. Although there was a fair amount of new data types that people started to use initially, just sort of Clickstream Data, but eventually, other types of relatively unstructured data. And then around - You know, I mostly work with big companies. Then around 2012 or 2013 or so, I started to notice that those big companies were basically using a lot of the same approaches that the Silicon Valley companies were but they were also doing more traditional analytics, decision support based analytics. Only doing it on a much larger scale, a more industrialised basis for it. And so I call that the Data Economy Era, where almost every company started to think about how you analyse data, if you're going to be successful as a business. As I say, big data, small data, the more traditional approaches plus the more highly industrialised approaches, starting in some companies to get into machine learning. Because machine learning and predictive analytics are at a certain level synonymous. Machine learning often involves more complex algorithms and so on. But anyway, that all led around, I don't know, 2017/2018, obviously the cycle times are shortening, into the AI Era where much more use of Machine learning, more complex algorithms, more automation of model development, just many, many more models with much larger volumes of data. And I think less interest in traditional analytics just because of the power of machine learning. But as I say, it's a pretty, I think, smooth continuum across era three to era four, and most of the era three companies have now moved pretty substantially in era four. Is that enough of a description or...?
Jonas Christensen 11:35
Yeah, it is. I find it interesting, the observation of the cycle times and how it's speeding up the jump between one to two to three to four, and so on. And I think, especially the last five years, I've found the industry to be evolving very fast in terms of tools, techniques, and also the skill sets that are necessarily required of the people inside doing the analytic, but also the organisations consuming this stuff.
Tom Davenport 12:00
Can I just say one thing about that? I agree with you and one of the really challenging things for analytics oriented people is the skills that were necessary in the first era never go away. The second era skills never go away. They just kind of cumulate. So I think it's one of the reasons why this idea that one type of person, one person can do at all, is no longer really feasible. This data scientist unicorn that can do everything you need to do in order to be successful with analytics, and AI is just impossible. There are just too many different skills involved that we've accumulated over the different eras of of analytics.
Jonas Christensen 12:46
Thank you for calling that out Tom, and to listeners out there who are sitting there feeling very scared with all the stuff you have to learn. Here's Tom Davenport telling you you can learn it all. Make sure you don't feel overwhelmed, even though it can be like that. Tom, so you touched on something there, which is we're all talking about analytics 4.0. But there's still a lot of the 1.0, 2.0 activity going on, the back office activity. I know that we definitely in my team do that on a regular basis and do have to manage some of that stuff. If you were to put a rough percentage on companies, organisations, maybe across industries if you want to pick particular industries, that do analytics 1.0, 2.0, 3.0, 4.0 respectively? What would that look like? I suppose what I'm talking about really is where they've matured to, rather than whether they're doing all or one.
Tom Davenport 13:37
Yeah, well, I think now, the vast majority of at least medium to large companies are doing analytics 1.0. You know, maybe they're doing it in Excel or something like that. But you have to do reporting as a public company, at least. And you have to do it fairly well or you go to jail if you're the Chief Financial Officer. So that's pretty important. Era two, I think, I view that the remaining aspects of Era Two, I think are mostly now web analytics. And most companies, I think, do some of that. You can do it for free in Google Analytics, if you want to say not terribly sophisticated, it's mostly counting things, but it's quite common. Era three, I kind of feel like about half of large companies are doing that in addition to the first two things. And then Era four mean that data suggests that, data from surveys suggest that, around 40 - 50% of companies have some AI activity going on. But most of it is not deployed into production. So, if you're looking at really aggressive use of AI, I think you're in the 5% or less category and I'm quite interested in this because I am just finishing a book now. It's not going to be called Competing on AI but it could be. It's basically, like competing on analytics. Only it's about AI. And it's about companies that are really aggressive in their use of AI and building their business strategies around it. And it's a pretty small percentage.
Jonas Christensen 15:23
Yeah, interesting. The percentage of 5% is obviously growing, but it shows you how hard it is to get to that level at this point in time at least for most companies, regardless of scale or volume of data that they might have sitting there. It's an organisational maturity. So, could you give us some examples, perhaps of some of these companies that have managed to get to that analytics 4.0 and what are the elements that make these organisations more advanced than others across strategy and vision, but also the technical and operational ability to bring these things to life?
Tom Davenport 16:01
Sure. Well, I'll give you some names of companies first, and then I'll tell you what they do. I didn't really find any in Australia, but doesn't mean there aren't any. But probably the closest geographically was Ping An in China or DBS bank in Singapore, both of whom I classified as kind of AI fueled, I was calling them, or AI first or all in on AI. I think the book is going to be called. And then if you look in other parts of the world, in the US, biggest largest grocery chain Kroger falls into that category. Capital One, which was one of my banking, competing on analytics company is now really aggressive on AI. Outside of the US shell, Airbus doing some really interesting work. Unilever, and in Canada, a fair number of companies. I talked about Scotiabank and Loblaw, which is the leading retailer and Manulife, leading insurance company. So what do they have in common? Well, they have strong leadership that really cares about this, which has always been important. I wrote a little piece about Puyish Gupta, the CEO of DBS, and how he's really played a very strong role, as a leader, of their AI capabilities. And same is true of most of the other companies. They are investing and building capability broadly. Alot of different technologies, a lot of different use cases, all around the organisation and many of them in production, not just experimental piloting approach. They are addressing the issue of their people from the standpoint of A) Do we have a cadre of people who can do data science oriented work? In many cases, also empowering the sort of citizen analysts or the citizen data scientists to do that kind of work with things like automated machine learning, and so on. They are also thinking about, 'What does it mean for the people who are going to be working with AI and how do we upskill them ?'' and some of them, including Unilever, has a very good approach to that, I think. They are really building their strategies around AI. So that might be new products and services. In the book draft which I'm just finishing about Morgan Stanley doing the next best action programme for its customers, that really changes the way they propose investments to them. It could be Ping An. It's so incredible in terms of their use of these ecosystem models, where they align with various companies in five different areas. I think it's banking, insurance, automobiles, smart cities and healthcare. And in aligning with those companies, they also end up getting more customers. And if they get more customers, they get more data and they get more data, their models get better. It's this kind of incredible, virtuous circle. And a number of companies I found are doing these sorts of ecosystems enabled by AI, which it's not too different from what Google and Facebook and so on were doing and Uber and Airbnb and social media. But in this book, I'm really focused on the legacy companies and how do you create a AI based competitor out of a company that is not a startup. It's much harder to change then. And, I don't know, I think those are the primary factors that - Obviously they have to have a lot of data. They are building new technology environments to support all of that data. Almost all of them are using data lakes of some type or other, many of them have moved into the cloud, not all of them. And I think it's possible to do this work on premise. But many companies have said it's made it somewhat easier for them to do this kind of work, integrating data and so on in a cloud.
Jonas Christensen 20:21
The ability for all companies to move into this is often a really challenging thing. So I'm finding it really nice to hear. And on behalf of the audience, also nice to hear that it can be done. And a lot of so called old companies have done it. You don't have to be a Silicon Valley startup to get there. And you're making me think of at the time when I worked in banking, we visited a data bank. We were sort of doing some knowledge sharing with them. And they showed us their technology platform, including also how they're capturing data and how they wanted to use it. And we said to them, ''Ah, look, we're so jealous of what you've built here. This is fantastic''. And they said, ''Well, you know what? We're jealous of you because you have customers. We don't have any customers actually going through this''. So there's a huge chicken and egg problem. Lots of them throughout all of these areas. So, Tom, you talked a lot of that banking here, actually, and financial services in general. You also used insurance examples. And typically, when you talk about very advanced analytics and machine learning AI, the examples that you hear are Netflix, Amazon, Facebook. All those companies that have really risen from nothing to start them in the last quarter, 20 years. But there are banks out there, more traditional organisations that do succeed with this stuff. Why are they different to all their other incumbent competitors and how did those other incumbents follow suit?
Tom Davenport 21:48
Yeah, well, you know, it's interesting, I thought a lot about that relative to DBS. And I think DBS was a not very impressive bank in terms of its customer service levels a couple of decades ago. And people refer to it as '' damn bloody slow'', and it was almost like a government agency. It was created by the Singapore government initially. I think it just has a lot to do with a) leadership. Piyush Gupta had a lot to do with it. He surrounded himself with some really smart people. I initially started working on AI with his head of technology and operations, a guy named Dave Gledhill, who's now joined Lloyds, and in the UK. He's British to begin with. But a lot of really smart people, at least in in DBS's case, a really strong orientation to competing with everybody. They view their competitors, not as traditional banks but as startups, even though they're not yet competitors in most banking areas. They want to have the same kind of rapid product development approaches and rapid innovation and so on, that those companies do and they continually refer to those digital natives as their real competitors. And then dissatisfaction with the status quo. Always improving. I mean, I wrote a little piece about their use of a chatbot for a digital bank that they started in India. All digital, no branches, and it's now in various several countries in Asia, including Indonesia. But just constant dissatisfaction with any customer need to call the call centre, trying to figure out why did they feel they needed to do that and why can't we put that capability in the chat bot and so on. So just constant improvement using technology and process changes, as well. He saw the same thing in any money laundering application at DBS, where they had a system a rules based system, as most banks do, but they added machine learning to it and they added network graph to it and they put in a new platform to deliver more data to it and you know, just kind of constant improvement.
Jonas Christensen 24:08
Leadership here is obviously a very important part, if not the most important part at least to get started. So mindsets division and so on. How do you see that being very different in your example here with DBS vs the other organisations? Is it all driven from the top or is there a different organisational culture as well that consumes this stuff?
Tom Davenport 24:31
Well, yeah, I thought a lot about that. I'm a sociologist by academic training. So the kind of organisation and leadership side is very interesting to me. And I have certainly seen examples where the organisation can do pretty well with analytics and AI with kind of mid to senior managers sort of leading the charge and eventually they persuade other people to do it and so on. But I don't think you make really fast progress without senior executive like the CEO leading it. And certainly that's true in DBS and Ping An. I would say some of the less dramatic successes in AI, the senior executives support it, the CEO supports it, they talk about it, some to the outside world, but they're just not quite as passionate about it. And they don't understand technology as much. I mean, I found it really interesting. When I talked to Piyush Gupta at DBS, I said, ''How did this come about?'' and he said, ''Well, you know, I was a protege of John Reed, at Citigroup. And John Reed, I knew from early in my career, was probably the first banker to really identify information technology as a competitive weapon. And Citi, now I don't think Citi has that lead. But for a while, you know, they were the first in the United States to do automated teller machines. And they were just very aggressive. They had a separate business unit to do technology based innovation. And so, he learned from him. I think, ultimately, you have to be a pretty good technologist, if you're going to believe in this stuff. I wrote about another company that serves the insurance industry and is doing these applications to do AI based identification of collision damage and estimations. And he made these long term bets that someday you'd be able to take a high quality photo on your smartphone of your car's damage. Someday, you'd be able to analyse that with deep learning models. They started to collect data on on images of collisions or they had and they really made it a huge asset. So it takes some of these long term bets. And unless you have some comfort with how the technology is evolving, that's difficult to do.
Jonas Christensen 27:02
Yeah, and long term bets require, as you said, a lot of support from the top, right from the top and ability to see that vision with you when you're building so that you don't get caught up in quarterly reporting cycles and all the rest.
Tom Davenport 27:17
Yeah, long term bets require long term investment. And CEOs tend to become aware of that investment. If they're not bought in, they probably are not going to be happy with it.
Jonas Christensen 27:27
So, Tom, we've talked now about this maturity through analytics 1.0 all the way up to 4.0. And with that, the people who are leading the analytics functions across organisations have to mature through that as well, because we talked about not having any unicorns anywhere. But I think the leaders that sort of manage these teams or functional areas, often have to know enough about all of it to be able to orchestrate that stuff. And that actually is very hard and a very difficult job because it's evolving so fast. But we as analytics leaders needs to mature and evolve with that evolution. So to get to 4.0, for most organisations in, say, the next 5 - 10 years, what are some of the do's and don'ts of analytics leaders who necessarily have to push this agenda?
Tom Davenport 28:22
Well, yeah, and it's not easy, because there aren't that many people that understand analytics and can have some of the other leadership traits that you need to manage those groups well, so I don't think you necessarily have to be a data scientists to lead those groups. But clearly, you have to understand what's possible with analytics and AI. And that requires a pretty good understanding of the technology. You have to be a good evangelist for this stuff. Because the vast majority of organisations, they don't necessarily understand what the value is, and how they can pursue it in terms of various applications and use cases. So I would say almost all of the good analytics and AI leaders I've seen spend a lot of their time evangelising. That is less true if you go to a digital native firm where you don't need to evangelise as much. Everybody sort of gets it. But in traditional firms, I think that's very important. And I like this term - Mark Shafer, who is basically the head of analytics at Disney, he started out as heading revenue management applications, but now his group pretty much does everything. And he calls it ''Evangelitics''. Really evangelising about analytics and they make it a practice to do that and they have kind of conferences where they can teach people about what other companies are doing and so on. Anyway, it's a good term. I think the ability to build trust and to have good relationships with senior executives, not necessarily something that data scientists are born with. So, if you're purely technical in these roles, I think it can be quite challenging to adopt some of these other attributes of evangelism and understanding the business well, and communicating effectively with business leaders. So I think you're probably better off with - you can get by with a little less analytical sophistication and you need some more business sense and good relationship skills, and so on. Obviously, you need to know something about information technology and data. In many, in many cases now, the Chief Analytics Officer is also the Chief Data Officer. So you have to think quite carefully about ''How we're going to transition our data environment nd how do we add value?''. I think it's fortunate how we say Chief Data Officer without Analytics is a really tough job, because it's hard to show short term value with data management. You can show short term value with analytics. But obviously, analytics depends on data. So in many cases, I think it does make sense to combine those two jobs. So, those are just a few of the things.
Jonas Christensen 31:17
Yeah, so to the listeners out there, it's important that you practice your communication skills as much as your coding skills. Absolutely. If you look at the evolution of different coding languages over the period of era 1 - 4, they've changed many times, but the thing that hasn't changed is the language that comes out of our mouths. So, you need to master that first and foremost.
Tom Davenport 31:39
Well, yeah, and related to that, I think more and more of the coding can be automated. Whether it's kind of point and click interfaces, or low-code/no-code tools or automated machine learning. More and more of the computer code necessary to create an analytical model can be automated. The evangelism and trust and relationship building and stakeholder management: No automation there, sorry to say.
Jonas Christensen 32:09
So Tom, these organisations that have really evolved in this space, it sounds like they often have a Chief Data and Analytics Officer. So they have that executive clout. They have support from the CEO, or modern support. They have a absolute clear remit and an agenda that they need to push. They have the technology that they need. They have the skill sets in the organisation to do the work. Is there anything else that these organisations are doing that is, I suppose, invisible glue that you can't add up like that?
Tom Davenport 32:41
Yeah, you have to have a pretty good sense. I mean, it's true of any successful CEO. You also have to have a pretty good sense of where the industry is going and how data and analytics and AI might contribute to that. So an ability to sort of see the big picture, at least, the future, and you have to have an ability to sort of bring the organisation along with you. I remember when I first started working with this guy, Gary Loveman, he was a Harvard Business School professor. Friend of mine. He became CEO of Harrah's and then that became Caesars. For a while it was the largest gaming or gambling company in the world. But he said, you know, ''The one big mistake I made is that I didn't get rid of people who couldn't follow the kind of database analytics based approach that we were taking quickly enough. I just kind of assumed that they would come around and they didn't, and eventually had to fire them anyway''. So I think bringing the rest of the organisation with you is really important. And even, even there, he was there for, I don't know, 10 years or so. When he left, the organisation slipped back into some of its old, you know, less analytical database decision practices. So it's really hard to get an entire organisation to change its thinking.
Jonas Christensen 34:15
So, you're describing here the risk and benefit of having that really strong evangelist, that one person who drives that agenda and makes a lot of stuff happen. How do we get organisations to consume all this stuff at scale, and sort of embed the cultural aspect into the organisation, so that we don't have the one person who is either there or not, and everything happens accordingly.
Tom Davenport 34:43
Yeah, well, you know, more and more organisations now are doing these kinds of data literacy programmes. I think that's a part of what you need. You have to understand what types of data are available and what you can do with them in terms of your business. Some of that can be kind of generic within a company. But some of it probably needs to be related to this specific function that you're in. How do you do data driven marketing? How do you do data driven supply chain work and so on? It comes down in part to how people are evaluated and compensated. If you can move to an environment where you're rewarding people for the database decisions that they made, even if the decisions don't necessarily work out well always, if people use the right approach to decision making, sometimes luck is going to get in the way, good or bad. But that's a big factor and that motivates people. Hiring people in the first place who care about this sort of thing. You know, Capital One, at one point, you had to pass a quantitative test before you could be hired in the company. I still remember talking to the Head of Human Resources, he said, ''Thank God, my company was acquired by Capital One. I could have never passed that test''. But I think those are all factors in having an ongoing long term culture that uses this set of approaches to deciding and acting.
Jonas Christensen 36:15
So one of the ways that I described to people that are not so aware of what analytics is and what it does right now is I compared to IT in the early 90s. Because I think that's kind of where we're at both in the level of maturity, relative to where it will go, and also in the organisational ability to consume analytical output. So in the early 90s, we sort of starting to get PCs into the office, and people are learning how to move the typewriting to computerised version. So we're all sitting there doing and WordPerfect and whatever else it was called back then. But all of a sudden, there's an expectation that you're just computer literate, when you walk into an organisation. No one's asking you, whether you know Microsoft Office today, when you start in most jobs. And we also saw in that period of time Chief Information Officers or Chief Technology Officers, whatever you want to call them, really come up from, from the back office that you've described already, in terms of analytics, right to the forefront and being this strategic role and leadership in the organisation. And you can envision the same for this space. So we take our long term goggles on here and look to maybe 10 - 20 years out, do you think that every organisation will have a Chief Data and Analytics Officer there and will they be a really important part of that executive leadership in most organisations? Or are there transitional roles that will morph into something else as we get there and the organisation is ''analytics literate''?
Tom Davenport 37:51
That's a very good question. I mean, I find it hard to understand how you could do without that role in any sort of large organisation. But I think the first Chief Data Officer was appointed in 1992 by Capital One, interestingly. Every year, I worked with a little - one of the companies that I advise is a consulting boutique called NewVantage Partners, and they do a big survey or a survey of about 100, large, mostly Chief Data and Analytics officers. One question that's always asked is, ''Is the Chief Data and/or Analytics officer job well established and thriving or is it still nascent and evolving or, have we seen lots and lots of turnover in it?'' and the numbers are getting better, but still only 40% say that it's established and successful. So it's turns out to be really hard to make a data environment better. And I think that's part of the problem. But I think we will definitely have Chief Data and Analytics officers or some variation upon them for the foreseeable future for most large organisations. And my guess is that they'll also, since ultimately a lot of analytics and AI is intended to make decision making better, I think they'll probably also end up being responsible for decision making in their company and how it's done. We still haven't really done much with decision making in the vast majority of organisations and if you ask your company, ''What are your top five most important decisions?'', they can hardly ever tell you. They don't keep track of who made what decision and how they did it. So we have no accountability. So, I'm hoping that that decision making component gets added although decision making can be very political. Many CEOs say, ''I'm responsible for decision making'' but in terms of really embedding better decision approaches into organisations, I think we haven't really seen much of that yet. And I'm hoping that it becomes the responsibility of some role and Chief Data and Analytics Officer would be a reasonable one to add it to.
Jonas Christensen 40:14
So you're describing here, an element of analytics maturing from back office to front office and also analytics being embedded in the products and services that organisations provide to their customers. So, products become not just digital products, but data products. And that requires, of course, a lot more of analytics leaders to produce products that are user-friendly, have ease of use, are technically relevant, are relevant in the marketplace. So you're really starting to combine technology, analytics, customer experience, design, and all the things that really matter in product design. So, we're morphing here into actually being a product and experience designers at the same time as building models that are statistically significant and are relevant and don't have bias in them as well. The complexity goes up many fold. What the companies that do this well, doing to facilitates that connection between the functions across the organisation and how do they structure themselves for one? Where do the analysts that the data scientists sit in your organisation but also how do they foster that collaboration to create these products and services that necessarily require cross-functional skill sets to come about?
Tom Davenport 41:40
Yeah, well, you know, I think you hit on it. And we haven't talked about it thus far. But I think more and more companies are realising that they need this product management role in data and analytics and AI. And it's similar to the kind of product management roles we've seen for traditional products in that it's cross-functional. They don't necessarily have a lot of people working for them, but they kind of coordinate and collaborate across the organisation. They understand how analytics work, but they're also quite focused on ''How do we introduce it successfully into the marketplace?''. And by the way, I'm not talking just about external data products. I think you would take the same orientation to internal data products. A new customer attrition model that you use in marketing, I think could also involve a product manager. And that product management in general is a relatively new concept. You know, somebody training for product management. You can do it in business schools now. But it's still relatively new. Data, product management is really new. There are a few people trying to focus on what it means and running a few seminars about it, and so on. But I think we need a lot more of that. Since I think increasingly, almost all of the things that we introduced to customers will have a data and analytics and AI component to it. There'll be smart products in one way or another. And so every product manager is going to have to understand this kind of thinking.
Jonas Christensen 43:20
Yeah, very fascinating evolution that we're going through here. And this is something for companies to really think about. How they structure their organisations for the medium and long term to be able to one produce all this stuff and to consume it for themselves and their customers. Now, Tom, we're sort of coming towards the end here. I've got three more questions for you.
Tom Davenport 43:41
I'll answer quickly.
Jonas Christensen 43:44
Firstly, where do you see the biggest opportunities for analytics 4.0 to really take off in the next decade?
Tom Davenport 43:53
Well, it's certainly in the more consumer oriented industries. They're the ones who have the most data by far. They are trying to personalise their products and services. They are trying to really understand at a much more granular level their customer needs and desires. So, banking insurance to a somewhat lesser degree health care, I think, is probably one of the biggest growth areas for this kind of work. Consumer products, although in many cases they don't directly sell their products, so they don't have as much direct consumer data, but they're getting more of it. In industrial businesses, I think there are some interesting things happening with digital twins and predictive asset maintenance and so on, but it's still probably going to be less than in the consumer oriented businesses. And among functions, I think the fastest growing areas are still marketing, which has lots of access to data. Sales increasingly has a lot of data. Marketing and sales, I think we'll have to collaborate more in the use of analytics. And interestingly enough, having Human Resources is one of the big growth areas in terms of identifying people who would be high performers within the organisation, avoiding attrition, etc. Now, there's more legislation brewing about that category than any other area. So we may not be able to use some of these analytics that we have developed. But until now, it's been very fast growing over the past several years.
Jonas Christensen 45:35
Yeah, it really shows how many areas analytics can be used for, and it's just up to our imagination basically. Because we are collecting so much data and all these things, because we are in digitised world. Tom, you've mentioned that you have a book coming out. I'm very interested in reading this book. Do you have a timeline for when we should expect it to be on bookshelves around the world?
Tom Davenport 45:59
Yeah, well, I actually have three books coming out in 2022. One is the one that I mentioned is going to be called ''All in on AI'' and we're shooting for October on that one. I just finished one with a couple of co-authors on AI and healthcare and that one, I think, should be maybe summer, northern hemisphere summer. And then the third one is pretty much done, but also takes a long time. It will be called ''Working with AI'' and it's 29 case studies of people who work with AI on a day-to-day basis. All of them will be coming out, I would say summer, August, September, October timeframe.
Jonas Christensen 46:45
Brilliant. I've just published the book myself with a bunch of co-authors and I know how long that took for us to organise. So I'm impressed that you can do three in one year. That's good.
Tom Davenport 46:56
Just barely got them done.
Jonas Christensen 46:58
Yeah. Well, Tom, last couple of questions. So firstly, on leaders of analytics, we pay it forward, which means I ask you: Who would you like to see as the next guest on Leaders of Analytics and why?
Tom Davenport 47:12
I would probably turn to one of these organisations that I've been doing work with. I really liked Dan Jeavons at Shell because he combines really this process reengineering approach with AI. They're doing some really interesting work. There are several people at DBS who I think would be good candidates. There's Head of Data and Analytics, but also Head of Digital Transformation. A multitude of people to choose from. A guy name Jing Xiao at Ping An. You can get him to talk to you. He's really impressive. And he's their sort of Chief Scientist. And then in the US, I've been really impressed by this organisation at Kroger called 84.51°. It's a really interestingly named company. That's the longitude of Cincinnati. They're quite good. Andy Hill is really good at Unilever. So a lot of people to choose from.
Jonas Christensen 48:11
Wonderful. Thank you for those recommendations. They are definitely ones that I will be looking at. Tom, lastly, where can people find out more about you and get a hold of your content?
Tom Davenport 48:20
Well, I sort of never had a thought I didn't publish. Most of it is either on my website, www.tomdavenport.com. or you can connect with me or follow me on LinkedIn. Almost everything I do ends up being mentioned on LinkedIn in one way or another. You get access to it. I mostly write for Forbes, Harvard Business Review and MIT Sloan Management Review. But everything gets mentioned on LinkedIn and in many cases put onto my website as well.
Jonas Christensen 48:50
Yeah, and I would recommend that the audience do check out Tom's website. It's a treasure trove of very interesting analytics content. So please go and check it out. Tom Davenport, thank you so much for being our leaders of analytics. Really appreciate your time and all the best and enjoy your day.
Tom Davenport 49:07
My pleasure. Enjoyed the conversation, Jonas.
