One of the perennial issues in analytics and data science is leadership. There is now general agreement that data and analytics are essential for business success. However, using data and analytics—becoming data-driven—does not happen without analytics leadership. This may come from the top, driven by executives and senior managers. However, it is also important for any manager.   There are many questions about developing analytics leadership, including what areas need to be addressed, and the roles of various players, including universities, businesses, and partners.   The importance of words—and research Jack Phillips is CEO and co-founder of the International Institute for Analytics (IIA). The IIA provides access to a network of analytics practitioners, experts and academics to help companies to become data-driven. Much of the IIA’s work focuses on analytics leadership, which is also a strong focus for SAS, including through its analytics value training. He draws an interesting distinction in the words used to describe analytics leadership.   “We increasingly see a difference between analytical leadership and ‘leaders of analytics’. The first term applies to non-analytics professionals, while the second applies to the emerging class of (often c-level) leaders of the data, analytics and AI function.” He also notes a distinction between analytics and other functions in organizations in terms of the volume of research. There is very little information about what skills and behaviors are needed for and by emerging leaders of analytics. Jack sees this as a real problem. “We know that artificial intelligence, and data and analytics more generally are the future. Understanding what type of leader is needed is also crucial.”   Cutting through the hype The emphasis for future analytics leaders is very much on how analytics is consumed and deployed to generate business value. At the moment, much of the hype is focused on artificial intelligence (AI), and particularly generative AI like ChatGPT. There is also a lot of enthusiasm for machine learning. However, it is important to be able to ‘walk before you try to run’. The challenge for analytics teams is to ride the wave of enthusiasm in the business, while ensuring that any investment generates value. Jack agrees that this is important.   “There’s nothing like having something that has a whole lot of sizzles and gets the attention of every major journalist and media outlet in the world. That’s what has happened over the past eleven or twelve months with generative AI. The teams we work with are using this opportunity to drive attention toward data quality and traditional analytics as much as this new shiny object.”   Jack comments that the applications for generative AI are interesting, but also limited to certain jobs and activities. He suggests that it is important for analytics leaders to make clear that generative AI may seem new, but it is only an extension of what has been happening for decades. It therefore sits very firmly within analytics—and should not replace basic functions.   “I think you have to recognize that the opportunity for generative AI is very limited compared to the potential of broad data and analytics in most organizations. Basic business intelligence and reporting can take many enterprises a very long way.”   Investing in analytics leadership The next question is what qualities are needed from analytics leaders—and who should take responsibility for developing them? Jack’s views are simple.   “Our research has shown that analytics leaders need an enterprise mindset, evangelist mindset, change agent mindset, people mindset, growth mindset, and technical mindset.”   Paul Hansen, from the Norwegian Business School, states that business knowledge is key for working with analytics. Riaan de Jongh, who has previously joined us to talk about the data science skills gap, suggests that analytics students should be required to do projects in industry to develop that knowledge. Jack agrees that the responsibility for developing crucial skills must be shared between education and industry.   “Higher education plays a crucial role in training future leaders. Einstein said that information is not knowledge, and the only source of knowledge is experience. We should see the split in responsibilities as universities provide the information, but enterprises have to provide the experience over time to develop new leaders.”   There is very little doubt that the highest performing enterprises when it comes to analytics are heavily investing in talent development, particularly through partnerships with local universities. Enterprises that are not doing this are falling behind. Development of data science skills—and especially analytics leadership skills—must be a shared responsibility.   ... View more

One of the biggest issues with analytics model development is often seen as the final stage of the analytics lifecycle: deployment into production. Many models never make it this far, even if the prototype does exactly what was required. Organizations have generally found that the answer to this problem is to think about deployment during development.   In other words, even as the model is developed, the data scientists, business team, and IT team need to be thinking together how the model will be deployed. This will include consideration of the systems required to run it, whether it will need to be ‘translated’ to run on those systems, and how it will be monitored. This makes it easier to deploy the model and keep it running in the future.   Recently, however, companies that are more mature and higher performing in analytical terms have started taking this a step further. They have begun thinking beyond the analytics lifecycle to create data or analytics assets that can be used over and over again for different purposes. The key to this is data product management, and the role of the data product manager.   Introducing data product management Product management is a fairly standard discipline in many tech companies. It has been described as being at the interface between user experience, business and technology teams. It is broadly about delivering great products that meet customer needs. It is particularly important within agile product development, where the product manager is responsible for keeping the process on track and consistent with customer needs.   In organizations using data and analytics, the role is similar. In a recent article in Harvard Business Review, Thomas Davenport, a co-founder of the International Institute for Analytics (IIA), suggested that data product managers are responsible for two types of products: data products, or reusable datasets, and analytics products, or models that can be reused or repurposed. Data managers are responsible for managing the product development process with an eye to future use.   They therefore work closely with both data scientists and business teams, coordinating the process and communicating and translating between the two. Ultimately, they have the responsibility for ensuring that models and datasets are deployed into production. This means making sure that they are usable and valued by customers. Data product managers therefore need a very good understanding of user needs, and must be able to speak the language of both the business teams and the data scientists. They need to understand data, but also the business.   After deployment, data product managers are then also responsible for managing the ongoing life of those products. This means monitoring their use and deployment over time, to ensure the ongoing value of data and analytics assets.   Finding the balance: data product management in practice What kinds of people make good data product managers? The answer to that, interestingly, is very much NOT data scientists. They are usually far too interested in the details of the model, and refining it to make it fit for purpose. Interestingly, data scientists often want to take a model further than is strictly ideal for the product manager, who is likely to favour a more generalist and therefore reusable model. This points to an interesting tension at the heart of any data product development process. In practice, organizations often find that there are three roles that need to be balanced: data scientist, engineer and data product manager. These three need to be jointly responsible for negotiating the product development process, and compromising on their individual priorities.   A new name for an old problem? Interestingly, the need for this role is not new. We have been talking for some time about how organizations need someone who can translate between data science, IT and business teams. In the past, we have often seen this as the role of data scientists. We have discussed how they can and should develop a better understanding of the business. We have also supported the trend for business users to develop their analytics and data skills, and become citizen data scientists to bridge the gap.   It therefore makes sense that high performing analytics organizations have formalized this role of data product manager. The only question is how long it will take before this is seen as standard for success with analytics. There is still an early-mover advantage to be had here, and wise organizations are likely to adopt this product mindset rapidly.     ... View more

Becoming a data-driven organization is not necessarily easy. It requires a genuine culture change in the organization, and it needs everyone to have the skills required to use data and analytics—and to understand the insights that they provide. This is simple to describe but can be extremely difficult to achieve.   SAS has recognized this issue and put together a training program on analytics value and transformation. It is designed to help both individuals and groups from the same organization. The program lasts 8 to 12 months, to give plenty of time for participants to change their behavior, and develop the mindset required to use data and analytics. It includes virtual session days to provide the theory of the key models required for analytics success, and coaching and hands-on learning days. These are spread across the program period to cover the four elements of foundation, deployment, advanced and AI/use case.    The main purpose of the training is to give attendees the knowledge, skills, and abilities to handle data and analytics as part of their day-to-day work. More importantly, perhaps, it also aims to provide students with a solid mindset that is focused on the regular and routine use of data and analytics to understand customers, situations, and problems.   I recently attended the analytics value training program as a attendee, and have also spoken to a couple of customers who have attended the course. I therefore feel well placed to comment on what I’ve learned, and how it has changed my approach to analytics within organizations.   The unexpected lessons are often the most important Going into the training, I hoped to learn more about how analytics and data-based decision making operate within the insurance industry. More broadly, I also wanted to understand how we can help organizations to derive value from investments made in analytical software and in data science. That aim was definitely accomplished. Perhaps more interesting was the unexpected learning that I had not previously considered. For example, I saw clearly in the program that analytics value training can build really strong bridges between data science teams and frontline service teams. I think all organizations face this problem. They have highly experienced data scientists who are great with data and algorithms. However, they often have relatively low levels of business knowledge or industry experience. On the other side, organizations also have business teams with deep industry expertise and knowledge, but little or no analytical experience. Analytics value training is all about bringing these two groups together and helping them to collaborate more effectively. It gives them a common language, and a much greater sense of what each group brings to the table.   You will learn more if you are prepared to go outside your comfort zone The second lesson for me was an interesting one. A lot of the people in my program were data scientists. My general impression from all of them was that they would have liked more data science in the program. They wanted to roll up their sleeves and do more hands-on algorithmic thinking, and perfect their craft a bit more. The issue is that this is their core competence, and their comfort zone. It’s therefore completely natural that they wanted to focus on that—but it wasn’t really what they most needed from the course. Instead, they needed to understand more about how to collaborate effectively with the frontline teams and think about data availability from the customer’s perspective. The program forced them out of their comfort zones, and into those new areas. Sometimes you need to be forced into thinking about new issues, because it’s easier not to bother. However, when you are prepared to go there, you can learn a lot more. Similarly, for the frontline people, it is helpful to get more insight into how to leverage data effectively, and how it can be used to answer business questions.   One of the best possible outcomes is having the tools to answer your own questions Analytics value training gives organizations the tools to self-assess themselves and their analytical proficiency. They can look across seven critical data or analytical dimensions and rate their maturity. This framework can be used to help organizations to make decisions about where and how to invest in analytics, and how to prioritize various areas for development. Effectively, the training gives participants and their organizations the tools to reach analytical maturity by themselves—and that is enormously powerful.   More information Analytics Value Training AVT start   The program for team or organizations, Analytics Leadership Program ALP package    ... View more

Dr. Daniel Yar Hamidi (PhD) is an expert in corporate governance and digital transformation. His main base is the Department of Information Technology at the University of Borås, in Sweden.  We caught up with him to discuss the value of analytics in organisations, and why corporate governance and innovation make good partners.   Daniel, tell us a bit about your work. My research is about corporate governance, generally in the field of innovation. I have worked on entrepreneurship and small and medium-sized enterprises for a long time and more recently, this has moved into corporate governance, and particularly decision making for governance in organisations. Most of our students are undergraduates, but I also work on advanced courses for working professionals. Even working at the university, one of my main aims has always been to engage with working professionals and help them to take the next step. These are effectively postgraduates who are coming back to the university to upskill and sometimes reskill.   Governance is often about making sure that things are steady, but innovation has a reputation for being a bit chaotic and disruptive. How do you see the relationship between the two? Perhaps the most important characteristic of innovation is that it’s about change. What we want with innovation is to challenge and change, and the outcome is very unclear. And as you say, governance is all about creating stability by strengthening institutions and developing grounds for consensus among stakeholders. I am interested in looking at innovation from a governance perspective, which means that I constantly try to find processes that will nurture innovation by creating the stable background to enable us to take risks. Innovation is about risk-taking, and governance is about risk management. When we bring the two together, we can talk about the kinds of risks we want to take, and those that we want to avoid.   What have you observed about the role of data and analytics in innovation, especially in small and medium-sized enterprises? In the last few years, I have had a very concrete focus on service innovation. I think if you are going to find out about and offer any new value proposition and create new services, you really need to be data-driven. We need to have relevant data, and be able to analyse them. Most of the innovations in companies today are actually about services. All the products and technology we are using are promoting new services to be developed and offered to the marketplace. To do that, we need to be able to gather and analyse an enormous amount of data—and if we can do that, we get new opportunities to innovate new services. Data and data analytics is absolutely the centre of this process.   You talked about upskilling and reskilling. Where would you say SME managers and leaders are in their ability to understand and manage data and analytics to support better decisions? It’s a paradox. We all use a lot of services that are based on data and analytics. However, when it comes to small and medium-sized enterprises, the capability to use this kind of technology is still in its infancy. It’s very immature. Many small business leaders don’t really see the possibilities in their own business. They appreciate the value, and they understand the power it will offer them, but they have enormous difficulty in translating that to their own businesses.   What do you think is the answer to this? We need to raise the levels of data literacy, and help them to see how they can bring together their business acumen with data and analytics. I think we have a long way to go yet. We are talking about small and medium-sized enterprises but I also work with large multinational companies, and many of them are having a similar difficulty clearing out their processes. They spend more money, and they have much more knowledge. However, they struggle to create value from data. SMEs are probably struggling more because of lack of resources, lack of knowledge and lack of partnership with other institutions, companies, and organisations. That’s where programmes like SAS’s Analytics Value Training comes in. I’d like to see more collaboration between academia and companies like SAS. We need to help companies to combine academic knowledge, business knowledge and analytics knowledge. If we can do that, it will be a great contribution to society.   Dr Daniel Yar Hamidi is a part of the ALP and AVT ecosystem expertise, read more   ... View more

It has almost reached the point where there are certain issues about data science that are taken as read. First, organisations need data and analytics. Data-driven decisions are the only way to navigate a rapidly-changing world. Organisations that harness data and analytics first and most effectively will be in a stronger position. Second, there is a shortage of data scientists. Third, even if you can get the expertise you need, it is still extremely difficult to change an organisation and make it data-driven.   There are no easy answers to these issues. Training up new data scientists is happening, but slowly. Data scientists can also only do so much. Training business users is also important, including through various training programmes, such as those offered by SAS.   Developing the answers Christer Bodell is now Director of Automotive and Manufacturing at Capgemini Invent, and formerly worked at SAS in the Nordic Region. He was involved in developing the original analytics value training programme. Talking to him, it seems that some issues remain constant. “I saw that when we showcased analytics solutions, there was a lack of capability to receive them and turn data into action. It was frustrating to show people what was possible, and then have them say, but what am I going to do with this? The analytics value programme was a way to reach out and get beyond this.”   Christer highlights the importance of getting the business to take control of analytics, and certainly to “take a seat at the table”. His concern is that data scientists and IT are often seen as being in control of analytics. In a way this makes sense: it’s technical, involves computers and the language of analytics and analysts can sometimes seem impenetrable even to other analysts. However, only the business can identify and decide which questions most need answers. We must find a way to translate between the three.   Analytics solutions as products One suggestion is that every analytics solution needs to be treated as a product inside the company. In other words, before it is developed, it is important to assess the market demand for it. What problem are we really trying to solve here? What is the value if we succeed? Data scientists need to design the scope, prototype, get approval, produce, and then launch each solution. Without this, solutions may sink without trace simply because potential users do not know about them or have little idea what to do with them. Christer comments that he has constantly seen three very clear challenges to implementing analytics in practice.   “Number one is you have to sell it, and that means you have to get acceptance for it. Two, you have to build it. And three, you have to get it working in the environment. The first two of these has got a lot easier because everybody believes that they want AI. It is also a little easier to build. However, the last piece—to get it properly used—remains as hard as ever.”   Minding the gap There is clearly a gap between desire and ability to implement. This gap is usually a matter of many small details and difficulties, rather than one big issue. Christer notes that there is also a gulf in understanding that must be overcome.   “It gets complex when you go from showing what has happened in the past, which everybody can understand, to showing something predictive, like a forecast or a predictive model. There is a discomfort in that you can’t absolutely rely on the numbers, in dealing with probability. People need to become more comfortable with how to deal with this type of uncertainty.”   This chimes with my previous experience of providing training for analysts in creating indexes. There was no single number or even right answer. Instead it was about finding weightings that could be agreed. I always felt that getting people to talk about weightings and debate them was a win, because it meant that they were engaging with the issues. This remains a long way from the ideal of data democratisation, but it is a start. Christer remains positive about the potential to get analytics accepted more widely. He notes that there is always a challenge to get people to look far enough ahead to see what problems can be solved with new technology. It is a bit like Henry Ford saying that if he had asked people what they wanted, they would have said faster horses. The pace of technology advancement is sometimes faster than we can assimilate as humans—but that doesn’t mean we should stop trying to plug the assimilation gap with every means at our disposal.     Moving forward, have a look at Analytics Leadership Program: __all information here__         ... View more

Over the last few years, increasing numbers of organizations have taken the plunge and invested in analytics and data science. They have recognized that it is more and more important for business decisions to be data-driven, not based on ‘gut feeling’ or the whims of those in charge. However, it is also fair to say that many of these organizations have not seen the desired return on their investment—or rather, they have not seen the rapid return on investment that they were expecting. It has often taken considerably longer than anticipated to become data-driven, and to move from isolated pockets of analytics use to analytics being ‘how we do things around here’. It is worth considering why this might be the case, and perhaps more importantly, what we can do about it.   A complex world First, let’s unpick what we mean by analytics and data science, because it is a complex area. One of my colleagues recently wrote a very helpful article about the changing scope and nature of data science. This explains that ‘data science’ consists of about 15 different activities, from finding, validating and merging data through building models to operationalizing and deploying them, then model monitoring, governance and retraining. Linus also commented that few data scientists are able to carry out all of these activities—or at least, not well. Most prefer to specialize in one or two areas from the 15 activities, and focus, say, on developing models, or operationalizing them, rather than trying to do everything. Linus also talked about who within the business was best placed to do each task. He highlighted a ‘business–IT split’ in the work. He explained that some tasks need more IT input, and others need more input from business teams. The examples that he gave were that finding the right data was really a business team function, because they were better placed to know what data were available. However, developing analytical models was more of an IT function.  Why does this matter? Because it shows that it is not enough to simply decide that you want to start using analytics and data science, or even that you wish your decisions to be data-driven. You cannot just pay for software, recruit a data scientist, or even several, and simply leave them to get on with it. Similarly, it doesn’t work to provide your business teams with analytics software and hope that they will be able to use it to answer business questions. It’s just too complicated.   Operationalizing analytics Instead, you need to consider how you will operationalize the use of analytics and data science throughout the organization. You have to develop a culture of analytics and data use and enable teams to develop the skills needed to use them. Ultimately, only this will enable you to deliver value from your investments in analytics and data science. In practice, realizing value for money from an analytics investment is a matter of developing three areas across the organization: Knowledge, which means knowing when and how to use analytics most effectively, and understanding the capabilities that are available Skills, which means the ability to use analytics effectively, including advanced techniques; and Abilities, and particularly being able to define analytical problems clearly, communicate the results effectively to decision-makers, and demonstrate analytics value. Driving change through leadership Delivering a programme of change, such as digital transformation, or becoming data-driven, needs leadership. Over many years, we have seen that strong analytics leadership is key to accelerating the realization of business value from analytics investment. It enables cultural change and drives the necessary evolution of the business towards data-driven decision-making. It is crucial to making analytics ‘the way we do things round here’. However, there is more to analytics leadership than simply driving change. Strong analytics leadership also makes an organization more attractive to data scientists. Data science is a shortage area. Data scientists are in high demand, and they can therefore pick and choose their employers. They are more likely to want to stay in an organization that both values them and understands their needs and wants. In other words, talent retention depends on providing an environment where data scientists want to work, and strong analytics leadership is part of this. Put simply, analytics leadership is the key to accelerating the realization of business value from investments in either analytical software or people. There is no other way.   Join us to explore the evolving scope of analytics leadership Over the past five years I have had the privilege to work with a number of analytics leadership practitioners as we delivered analytics value training. I will be exploring with them the opportunity to unlock greater value through better analytics leadership. I hope you can join us.   ... View more