Essential Habits for Data-Focused Leaders

A good one, a great one. So I'm Rita Salam, and welcome to High Business Impact Habits of Top Leaders Who Have Data, Analytics, and AI Responsibility. So one of our most frequent inquiries in the data analytics and AI team, that's where I sit in Gartner, is many of our leaders ask, how can we become a data-centric... organization. A reasonable question, like how many of you aspire to be a data centric organization? Yeah, most of you and I think that's great but I hope that I convince you by the end of this 30 minutes that we not only want to be data centric but we want to be first and foremost value centric and the reason is that we see that that is a habit of top performing leaders. Let me start with An analysis that we did. So we had a hypothesis that companies treating data analytics and AI as strategically important to their organization, meaning it's aligned with the business, would be top financial performers. And so to test that hypothesis we analyzed the financial performance of the global S&P 1200 companies on six key financial metrics for the past nine years from 2014 through to mid like half of year 2023. And we compared that performance, again, financial and valuation metrics against earnings calls and mentions in earnings calls of data analytics and AI terms during that same time period. And the theory, the sort of hypothesis was that given that companies have very limited real estate on earnings calls, that mentions on those earnings calls would indicate strategic, you know, that this thing is strategic. And so what we found is that consistently high-performing organizations, meaning organizations that performed better than their industries on 80% of these financial and valuation metrics across that entire nine-and-a-half-year time frame, mentioned data and in analytics terms, particularly AI, 50% more than... what we would call low performing organizations, those that performed worse than their industry across 80% of these metrics for nine and a half years. So, you know, what do these leaders do? What do the leaders in high performing organizations do? That's what we'll talk about today. So data and analytics leaders, first and foremost, again, clearly from the analysis view data, analytics, AI as strategic to the organization's business, particularly in the world of AI. So We made a prediction to sort of represent that intent. By 2026, more than a quarter of global organizations will have at least one top-earning product that is based on data, analytics, and AI. And that's particularly in the world of AI. You're likely to have more and more sort of analytics or AI incorporated into your product or service. But in our analysis, of over 100 data and analytics and AI leaders, we found... that not all leaders are the same, right? Their focus, some function more as business drivers, as you can see, taking in customer, focusing on revenue, product. They often have a seat at the executive table. Others play more of an enabler role, focusing primarily on data foundation, on operational efficiencies. But the most impactful of them really do both, really do both. You really have to do both. So here are the top high impact habits of top technology leaders that focus or have as their portfolio data analytics and AI. Now you can either begin to develop them yourself, you may already be doing some of these which is great. Hopefully you'll get some new ideas to take the habit to the next level. Now for those of you who are in my session on Monday, This will not be new to you, that was sort of the mantra of the session on Monday, that you need to create a business strategy that's infused with data and analytics and AI, not some separate data strategy that's disconnected to the business. I mean, very clear. It's core to the organization's success, the value delivered is very clear, how the value is measured is very clear. Now what I'll focus on today, not so much that I didn't really focus on this on Monday, is the value story skill. So, you know, how do you do this, right? We talk a lot, again, about being data-centric, but the data-centric strategy by itself usually results... and you being perceived as an enabler rather than a driver, as a cost center rather than central to the business. And so we need to build that value-centric data native business strategy. And again, the strategy sort of answers the question, you know, why am I doing this? What am I doing? How is this going to impact the business? A lot of times when we talked to many, many of our clients about their data and analytics strategy and we say, well what's your strategy? They'll give a vendor name. It's AWS, it's Microsoft, it's Google. Or they often list a technology roadmap or a long list of initiatives. We're working on self-service, we're working on governance, we're working on master data management. And all of those things obviously are great, but how do they impact the business strategy is what we need to know. Whereas the operating model is how do we succeed? What are the capabilities that we need? What's the technology? What is the delivery model? What are the skill sets we need? to deliver on the initiatives laid out in the strategy. But high business impact leaders, importantly, excel at telling a value story around their strategy and operating model that make business leaders and their peers the heroes in their value story. They're skilled at connecting the initiatives that they're working on, we call them value enablers, to the specific stakeholder priorities of the organization, meaning... How is that stakeholder going to be successful at the end of the day? How are they going to measure their own success, their organization's success? How are they going to get paid their bonus? We need to know that and tie our initiatives to that thing and specifically then determine the specific outcomes in terms of impact on process, activity, decisions in the organization, lay that out very clearly, and then measure the return to the organization. identify the KPIs that actually measure those outcomes. Again, we need to be talking about the initiatives we're working on in data and analytics, in any technology. Actually, this applies to anything in terms that our business peers, our executives, our board can relate to in terms of how it enables their success. Now, for each of these examples or each of these habits, I'll give you an example. So the first one is urban shopping. They do a great job at this. This is a pseudonym for a retail chain in Europe and their strategic business objectives are to really optimize revenue from the customer journey. They use marketing analytics to do this to achieve that objective and so they had to figure out the business impact along that journey. So they mapped out the journey and they identified the impact of analytics at each part of that process and they identified KPIs to measure the impact on the stakeholders. So just if we whirl through this quickly, they used analytics to segment their customer base. They identified a specific persona, Ann is a budget shopper, she based on analytics, marketing analytics of historical behavior, her demographics, they present to her personalized promotions by email. We all get them, and so The more personalized, the more precise, the more accurate, the better our data quality is. Potentially, that impacts revenue per transaction. Anne visits the website. She completes her order. But based on that, we may serve up to her additional things that she may like to buy, again, based on her demographics, her purchasing history, et cetera. People who buy similar things, again, analytics impacting revenue per customer. She may receive her order at home. We don't map it out here, but maybe there are supply chain analytics that make sure she gets her package in two days that she paid for. And then, you know, she may visit the store. There are additional location analytics. We may use beacon data from smartphone locations to present to her promotions while she's in the store to help her navigate the store to find the things she wants. So, you know, very useful, great value story that's very clear on how what we're doing in analytics around marketing and sales impact the thing that this organization cares about. So impact number two, high performing leaders who focus on data analytics and AI prioritize talent, skills, and change management. We can deliver all the analytics in the world. If people aren't going to use that to impact how they change their behavior, how they do their work, how they make decisions, you're not going to realize value. So and these three things are interrelated. Again you can't get value out of data and analytics unless you leverage the analytics to change how people do things. And so high-performing organizations, they're typically data and AI literate, meaning everyone knows how to use the data and the insights. available to them to intentionally affect how they work and how they make decisions. They measure the results. They're accountable. They're analytical. They have critical thinking. They challenge the status quo. They think about using data and analytics and AI to innovate within the organization. They have communities of practice. right, where they impart knowledge, where they train, where they deliver data literacy, where they impart governance policies and practices. They're empathetic because they realize that hey change is hard and so how do you bring people along who may be resistant but who need to leverage those insights to impact the business. And so value-driven organizations really do think and act differently with all of these trades and they're constantly planning for new capabilities, new roles. You know, prompt engineering obviously is the biggie right now. Everybody in the business ultimately, even if we're rolling out co-pilots to just consumers, everybody in the business will need to have some additional skills to fully leverage, you know, generative AI or predictive AI or other analytics that we may deliver to them. Of course, developers will need a different level of skill than your CEO or your VPs and your executives. You know, and there'll be some roles that may go away. I wonder, you know, in the 60s, as you all know, there was a role called a computer, right? Before we had computers, there were people who calculated, you know, long equations to send people to the moon. We saw that in, you know, movies. But what role do you think will go away with generative AI or AI? I wonder. Maybe it's business analyst, right? If everyone has an AI chatbot to be able to analyze and... understand what's going on in the business, maybe we no longer need that. But we'll have other roles that emerge, data coach, data translator, AI monitor, AI ethicist, of course, data ethicist. We'll have other roles that we actually have predicted a neutral impact on jobs at Gartner. But there will be some roles that we need more than others, skills we have to develop, and others that we will need less over time. So high-performing leaders, keep an eye on that one. So we know no roles will be changing, we know you'll have to keep up to date on skills, but the one thing will be constant is the need to deliver insight. So we've got a couple case studies here. The first one is Intuit, and what I really like about this one is that they don't just give users data or reports. They don't just throw out that data catalog and say go at it, you know, do what you want, have fun. Even with rules, that's really not the best way to get impact. What they deliver is insights. You know, what should users be deriving from the content that's being delivered to them? How should they use it in their jobs? How should they use it to impact their jobs? So here, you know, they do quarterly reviews, so past, present, and future. They look at their monthly updates and then they do forecasts and they tell their consumers of this content what they should be doing with it. The other one that I really like is GFS security, Guangfa Security. And what they've done, I think they have what I would consider to be a best-in-class data literacy program where they make analytics a skill for everyone, but just they skill people differently based on their role. The program is designed to develop analytic capabilities that is very role-specific, but they also use it to identify top talent. And so it has great attributes. So they certainly, they have this in-house training. It's online, it's in classroom, and importantly, it's applied learning. Again, teaching people how they use the insights to impact their jobs. They use competitions, they use rewards. I'm a big fan of that to encourage participation. They focus on analysis. Again, how somebody uses this to impact the business. So they don't just teach tools, they teach real-world challenges from the business. they measure the results. They have certification as a way to make sure that people have the skills that they need. So I really like that one. But how do you reinforce change? We all know that change management is the rub, right? That is how you make progress. That's how you deliver value. And we all know that that's a monumental task in most organizations. So I was an expert on a panel. I was moderating an expert panel. a few years ago and somebody from the audience asked the panel, you know, how long does it take for change to be realized? And one of the other panelists was a CEO of a public services company and said, well, whenever the last person in the old regime dies. He literally said that. You know, of course I was thinking that's metaphorical, I don't think he's gonna be killing anybody, but it does indicate that, you know, change takes time. It's hard. We need to be patient. We need to be, again, empathetic. High-performing leaders here, you know, provide people, again, back to sort of the value stories to help them understand, you know, what's in it for them. They inspire them to want to make the change. Again, not easy. You know, one of the ways to bring people along is through stages to reinforce for them sort of the from to because narrative. You know again making it easy and possible for them to understand how they make that transition, reduce their friction, you know be clear that the new normal yet may be scary but it's going to be amazing for you. Reinforce it, show them how to how to make the way. And here's the big one, and this is not just DNA, have the bravery to shut stuff down when it's time. You know Get rid of the old stuff once for all. Clearly there'll be in some cases stuff that's just too risky, you know, to shut down. If you've got regulatory reporting that, you know, we understand that. But majority of things, you know, give it an end date. And don't give people the choice to remain in the past. And you know what? It's okay not to bring absolutely everyone along. There'll be folks that you just can't bring along and that's okay. You know, you can't please all the people all the time. But, you know, through these steps... Hopefully you'll get people along and again measure them. Measure them. So, best habit number or high performing habit number three. Drive business innovation, leveraging emerging trends, use fast teams, set out audacious goals. Now you've been inundated with trends this week, no doubt. And by the way, I think we've sort of given you the general guidance that playing it safe doesn't mean ignoring trends, right? Doing anything, even just to learn, you know, if it aligns to your business, can sometimes be better than nothing. So high-performing data and analytics teams, again, they drive innovation. Sorry, they are pretty quick to deliver value, and if something isn't working, they shut it, in terms of a proof of concept, they're pretty quick to shut it down. You know, over the past seven years, you know, AI has been, I would say, you know... pretty, I wouldn't say mainstream, but many forward-looking companies have started to do proof of concepts and implemented least predictive AI. But it's not uncommon to take six to eight to nine months to do a proof of concept on a model to try to say, oh, well, I have value. And then by the end of that long effort, you find that it has minimal impact. So I think some of the practices that we see, you saw it in Page Group, I'll talk a little bit about them in a second, 28-day cycle time. for assessing if you should move forward, doing a quick proof of value. Again, having big goals, not just, oh, we're going to improve by 5%, but we, our goal is to improve by 90%. And if we get halfway there, it's good. Now, here are our top trends for 2023. I am not going to go over them. Gareth Herschel delivered them yesterday. If you want to listen to that presentation, it's amazing. You can listen to it on replay. I think what you can see though is the trends align to the six habits and you know many of them are not so you wouldn't be surprising to you. AI risk, you know optimizing value, data sharing, sustainability, humans don't forget them, platform and ecosystems which we'll talk about in a second. So here's the example again I'll go in a little bit more detail about Page Group. You know they again had a huge really impactful initial I would say it's not really everyday AI because it's a core process. You know, recruiters, a key thing that they do is write job descriptions. But, you know, they were, it's really one of my favorite examples because, you know, even though it is everyday AI, they had a significant impact. You know, as just to refresh, they're a European job recruiter. They've got 8,000 recruiters. They were able to take the time it takes from writing a description from 90 to 5 minutes. 20 to 90 minutes to 5 minutes. But the real story about this, again, is how they did it. So they have a fast cycle innovation lab, they call it their data lab, where they assess new proof of concepts in 28 days. If they're not able to demonstrate value after 28 days, they move on. Now, in some cases, they'll say, we'll reassess later, like if it's a matter of a cost-benefit type of problem. But 28 days. If they have to make decisions and people aren't in the office or they're on vacation, their policy is to just make decisions with whomever is there on that day. So a very, very good process. And again, major goals. The big thing for them, the AI-ready data. So they had been working... Working on AI for the past seven years and they've been working on their data fabric for the past four and a half years so that they credit the data fabric with them being able to get to the five minutes so they were able to achieve a certain amount of productivity gain from the generative AI piece but because they had to pull data from four different systems that had never been integrated before to pre-populate these job descriptions they were really able to get to that five minutes. because of the AI-ready data, and they're going to be able to scale it. So high-performing habit number four, build data analytics and AI products, not projects. Now, there are three types of archetypes, if you will, for data products. There's data as a service, there's sort of value stream insights, and there are essentially information products and services that you might sell. external to your enterprise to create a revenue stream. Again, I think many of you will ultimately do that. Now, that doesn't mean you need to turn everything into a product, but identify those product opportunities by really being, again, customer-centric. Be that customer advocate. Understand requirements. Again, your customer may be internal. They may be external if you're over to the far right. Basically, you know, promote agile delivery practices as you would if you were a product manager creating a product that's going to have a SKU, that's going to have to be updated, that you're going to have to enhance over time, that you're going to have to maintain. The idea is you want to create reusable data products that, you know, have a lot of traction across the organization, not just one-offs. You know, ideally, if you can hire people with product management skills. That's helpful. And again, I don't think this just applies to data and analytics, but we see that high-performing data and analytics leaders do this thing. So case study. Again, most companies, if I'm being honest, aren't taking the data they have, the unique data they have, combining it potentially with other data that's external to create unique data sets that they can monetize. Either by selling the data by itself or putting analytics on top of it and creating some sort of analytic service. Of course, many use obviously data products to create value. Again, these reusable data sets, but it's the same process. ZF Group actually does this. I think they offer a best-in-class approach to data product creation that any company can adapt to sort of kick-start the process. So what they did is to monetize the data that's generated by its systems and first identified a business problem, right, where the data that they have could absolutely map to. And then they gathered the data and created these new products, making sure that the business understood how to market and sell these new products. And of course, they measure the impact and adjust and evolve the product. They enhance it. They maintain it over time. So you may see it looks pretty similar to the infinity loop, what they did, essentially delivering a product, measuring the impact, adjusting and evolving. over time they measure and have KPIs, OKRs, to really measure the results of the impact that this is having. So really best in class. This is actually a case study on Gartner.com if you want to learn more details about it. So high business impact number five. Basically building a composable data foundation, ecosystem that essentially enables fast response to business. and technology changes. Now, this is a representation of a data and analytics cloud ecosystem. And what we know is that data and analytics cloud ecosystems versus on-premises platforms in the past support more integrated workflows from data to analytics to AI and machine learning to applications. That's sort of the value proposition of having a lot of the hyperscalers are really touting. Versus stacks, on-premises stacks in the past, you could get the set of capabilities that you need, the portfolio of capabilities, but the integration was often very light, discrete products, sort of integrated at the invoice level often. But what you see now is that these sort of ecosystems of capabilities prioritize workflow integration, prioritize integration with ISVs to fill in gaps where they... They may be light or to integrate with your existing environment. And so, you know, I think what's important here is, again, AI-ready data. AI-ready analytics is a foundation. I think we're at a point where we really can't kick that data can down the road. So this becomes a very important foundation. Again, look at PageGroup. And many of our clients, I think, are really there. having made that sort of investment, even if it's parallel to as you're developing these AI use cases. I know many of you will say, well, I'm not going to just develop this all at once, take three years to develop my data-ready foundation. No, you have to deliver value now. So you might have to do it sort of in parallel to high-value use cases. Again, AI-ready data. This is a really, I love this slide, and we talk a lot about this in our team, where you have traditional data management practices. And those are amazing, but AI requires sort of additional capabilities, whether it's labeling and bias mitigation and different types of security, and even if you're moving into generative AI with enterprise search, different types of technology like embeddings and vector databases, etc., data, which is data enrichment, knowledge graphs. At the same time, people building out these models typically focus on model accuracy. How many models do I deploy in production? They need also to get a sense for data and the importance of AI-ready data. And so we have to have a much more integrated process. Definitely high-performing leaders put in place processes to scale because it's very different to do something in production versus scaling it and maintaining it over time, particularly when we're talking about AI. You need data ops and DevOps and AI ops and ML ops particularly. you know, processes that allow you to scale over time and particularly incorporate new technology as it evolves, which we'll certainly see with generative AI. Great example here, put in place a framework that allows you to, again, model and scale operations. Fidelity has done a great job here. They've managed to increase speed to production by, you know, double the speed, you know, essentially reduce the time by, you know, 100%. And they've been able to resolve product issues and reduce the time it takes by 80% by putting in place these kind of processes. So last one, I'll be really quick, governance. And unfortunately, this is probably one of the more important ones, right? Data and analytics leaders, AI leaders treat governance and risk management as an essential value driver, right? There's nothing more important here. You know, where you need to articulate foundational investments in terms of, in terms that business peers can understand. Now usually when we mention governance to the business, what happens? Have any idea? You probably have seen it. Eyes glaze over. They may roll in the back of their heads. They may even fall asleep. And they may just say, no, you're trying to control me. And so we don't want that. But there is a way to talk about how governance is so critical to most of most every use case that we're delivering and we need to tell that story. And by the way it doesn't mean that we have to set up this big upfront governance practice. It doesn't mean that we you know catalog everything before we can start day one. It doesn't mean that we enlist data stewards before we even you know think about analytics. It has to be a much more agile process. The way I view it and the way I think we have to sell it is governance as a team sport. It is essentially the enterprise trust initiative. So I almost don't use governance with business people. I say, you like to be trusted? You want to be trusted externally? You want to have an excellent reputation? Then you need to have governance. And that's how we need to sell. It is the enterprise trust initiative. Everyone wants to be trusted. So I'm going to just focus on this one, given we just have a little bit of time. So this one happens to be an AI one. It's a great example, again, AI governance. They are, in this case, Axon is a global leader in connected public services. They've assembled an independent AI board, which I think is really interesting. The AI Ethics Board really helps them in the development of their AI-powered devices. Their development team has access to them. Their ombudsman has access to them. And this board sort of over-represents civil liberties. privacy voices and concerns. It shares its feedback with Axon but also publicly. So Axon credits this board with really helping them with their AI strategy. They feel like it's a crucial part of their sort of AI first development strategy. So as a reminder we covered you know business and business build a value centric data and analytics and I native strategy. Prioritize talent, skills, change management as core competencies. Again, you can't deliver value if you don't have the right skills, if people don't know how to use the analytics that you're making available to them. Build, of course, drive innovation by leveraging emerging trends. One of the aspects of an excellent strategy is that you incorporate trends and drivers in that strategy. Many strategy documents that we review don't have that. Set audacious goals. Think about fast innovation cycle times. Build products, not projects. Build a scalable, composable, agile data analytics and AI ecosystem and foundation. Focus on AI-ready data. Focus on processes that allow you to scale, not just do proof of concepts. And then finally, treat governance and risk management as a key value driver as it is. So I'll just leave you with one question. Given all of this, maybe you can think about what type of leader will you be? Hopefully, you will be both a business driver and an enabler to help your organization shift from being data-centric to being value-centric. So thank you for listening. I'm really excited about you all being here with us this week. Hopefully, you have a safe trip home and enjoy the rest of the conference if you're going to stay for another session. Thank you.

Yeah, most of you and I think that's great but I hope that I convince you by the end of this 30 minutes that we not only want to be data centric but we want to be first and foremost value centric and the reason is that we see that that is a habit of top performing leaders. Let me start with An analysis that we did. So we had a hypothesis that companies treating data analytics and AI as strategically important to their organization, meaning it's aligned with the business, would be top financial performers.

And so to test that hypothesis we analyzed the financial performance of the global S&P 1200 companies on six key financial metrics for the past nine years from 2014 through to mid like half of year 2023. And we compared that performance, again, financial and valuation metrics against earnings calls and mentions in earnings calls of data analytics and AI terms during that same time period. And the theory, the sort of hypothesis was that given that companies have very limited real estate on earnings calls, that mentions on those earnings calls would indicate strategic, you know, that this thing is strategic. And so what we found is that consistently high-performing organizations, meaning organizations that performed better than their industries on 80% of these financial and valuation metrics across that entire nine-and-a-half-year time frame, mentioned data and in analytics terms, particularly AI, 50% more than...

what we would call low performing organizations, those that performed worse than their industry across 80% of these metrics for nine and a half years. So, you know, what do these leaders do? What do the leaders in high performing organizations do? That's what we'll talk about today.

So data and analytics leaders, first and foremost, again, clearly from the analysis view data, analytics, AI as strategic to the organization's business, particularly in the world of AI. So We made a prediction to sort of represent that intent. By 2026, more than a quarter of global organizations will have at least one top-earning product that is based on data, analytics, and AI. And that's particularly in the world of AI.

You're likely to have more and more sort of analytics or AI incorporated into your product or service. But in our analysis, of over 100 data and analytics and AI leaders, we found... that not all leaders are the same, right? Their focus, some function more as business drivers, as you can see, taking in customer, focusing on revenue, product.

They often have a seat at the executive table. Others play more of an enabler role, focusing primarily on data foundation, on operational efficiencies. But the most impactful of them really do both, really do both. You really have to do both.

So here are the top high impact habits of top technology leaders that focus or have as their portfolio data analytics and AI. Now you can either begin to develop them yourself, you may already be doing some of these which is great. Hopefully you'll get some new ideas to take the habit to the next level. Now for those of you who are in my session on Monday, This will not be new to you, that was sort of the mantra of the session on Monday, that you need to create a business strategy that's infused with data and analytics and AI, not some separate data strategy that's disconnected to the business.

I mean, very clear. It's core to the organization's success, the value delivered is very clear, how the value is measured is very clear. Now what I'll focus on today, not so much that I didn't really focus on this on Monday, is the value story skill.

So, you know, how do you do this, right? We talk a lot, again, about being data-centric, but the data-centric strategy by itself usually results... and you being perceived as an enabler rather than a driver, as a cost center rather than central to the business. And so we need to build that value-centric data native business strategy. And again, the strategy sort of answers the question, you know, why am I doing this?

What am I doing? How is this going to impact the business? A lot of times when we talked to many, many of our clients about their data and analytics strategy and we say, well what's your strategy? They'll give a vendor name.

It's AWS, it's Microsoft, it's Google. Or they often list a technology roadmap or a long list of initiatives. We're working on self-service, we're working on governance, we're working on master data management.

And all of those things obviously are great, but how do they impact the business strategy is what we need to know. Whereas the operating model is how do we succeed? What are the capabilities that we need?

What's the technology? What is the delivery model? What are the skill sets we need?

to deliver on the initiatives laid out in the strategy. But high business impact leaders, importantly, excel at telling a value story around their strategy and operating model that make business leaders and their peers the heroes in their value story. They're skilled at connecting the initiatives that they're working on, we call them value enablers, to the specific stakeholder priorities of the organization, meaning...

How is that stakeholder going to be successful at the end of the day? How are they going to measure their own success, their organization's success? How are they going to get paid their bonus?

We need to know that and tie our initiatives to that thing and specifically then determine the specific outcomes in terms of impact on process, activity, decisions in the organization, lay that out very clearly, and then measure the return to the organization. identify the KPIs that actually measure those outcomes. Again, we need to be talking about the initiatives we're working on in data and analytics, in any technology.

Actually, this applies to anything in terms that our business peers, our executives, our board can relate to in terms of how it enables their success. Now, for each of these examples or each of these habits, I'll give you an example. So the first one is urban shopping. They do a great job at this.

This is a pseudonym for a retail chain in Europe and their strategic business objectives are to really optimize revenue from the customer journey. They use marketing analytics to do this to achieve that objective and so they had to figure out the business impact along that journey. So they mapped out the journey and they identified the impact of analytics at each part of that process and they identified KPIs to measure the impact on the stakeholders. So just if we whirl through this quickly, they used analytics to segment their customer base.

They identified a specific persona, Ann is a budget shopper, she based on analytics, marketing analytics of historical behavior, her demographics, they present to her personalized promotions by email. We all get them, and so The more personalized, the more precise, the more accurate, the better our data quality is. Potentially, that impacts revenue per transaction. Anne visits the website.

She completes her order. But based on that, we may serve up to her additional things that she may like to buy, again, based on her demographics, her purchasing history, et cetera. People who buy similar things, again, analytics impacting revenue per customer.

She may receive her order at home. We don't map it out here, but maybe there are supply chain analytics that make sure she gets her package in two days that she paid for. And then, you know, she may visit the store. There are additional location analytics. We may use beacon data from smartphone locations to present to her promotions while she's in the store to help her navigate the store to find the things she wants.

So, you know, very useful, great value story that's very clear on how what we're doing in analytics around marketing and sales impact the thing that this organization cares about. So impact number two, high performing leaders who focus on data analytics and AI prioritize talent, skills, and change management. We can deliver all the analytics in the world. If people aren't going to use that to impact how they change their behavior, how they do their work, how they make decisions, you're not going to realize value. So and these three things are interrelated.

Again you can't get value out of data and analytics unless you leverage the analytics to change how people do things. And so high-performing organizations, they're typically data and AI literate, meaning everyone knows how to use the data and the insights. available to them to intentionally affect how they work and how they make decisions.

They measure the results. They're accountable. They're analytical.

They have critical thinking. They challenge the status quo. They think about using data and analytics and AI to innovate within the organization.

They have communities of practice. right, where they impart knowledge, where they train, where they deliver data literacy, where they impart governance policies and practices. They're empathetic because they realize that hey change is hard and so how do you bring people along who may be resistant but who need to leverage those insights to impact the business. And so value-driven organizations really do think and act differently with all of these trades and they're constantly planning for new capabilities, new roles. You know, prompt engineering obviously is the biggie right now.

Everybody in the business ultimately, even if we're rolling out co-pilots to just consumers, everybody in the business will need to have some additional skills to fully leverage, you know, generative AI or predictive AI or other analytics that we may deliver to them. Of course, developers will need a different level of skill than your CEO or your VPs and your executives. You know, and there'll be some roles that may go away.

I wonder, you know, in the 60s, as you all know, there was a role called a computer, right? Before we had computers, there were people who calculated, you know, long equations to send people to the moon. We saw that in, you know, movies. But what role do you think will go away with generative AI or AI? I wonder.

Maybe it's business analyst, right? If everyone has an AI chatbot to be able to analyze and... understand what's going on in the business, maybe we no longer need that. But we'll have other roles that emerge, data coach, data translator, AI monitor, AI ethicist, of course, data ethicist. We'll have other roles that we actually have predicted a neutral impact on jobs at Gartner.

But there will be some roles that we need more than others, skills we have to develop, and others that we will need less over time. So high-performing leaders, keep an eye on that one. So we know no roles will be changing, we know you'll have to keep up to date on skills, but the one thing will be constant is the need to deliver insight.

So we've got a couple case studies here. The first one is Intuit, and what I really like about this one is that they don't just give users data or reports. They don't just throw out that data catalog and say go at it, you know, do what you want, have fun.

Even with rules, that's really not the best way to get impact. What they deliver is insights. You know, what should users be deriving from the content that's being delivered to them?

How should they use it in their jobs? How should they use it to impact their jobs? So here, you know, they do quarterly reviews, so past, present, and future. They look at their monthly updates and then they do forecasts and they tell their consumers of this content what they should be doing with it.

The other one that I really like is GFS security, Guangfa Security. And what they've done, I think they have what I would consider to be a best-in-class data literacy program where they make analytics a skill for everyone, but just they skill people differently based on their role. The program is designed to develop analytic capabilities that is very role-specific, but they also use it to identify top talent. And so it has great attributes. So they certainly, they have this in-house training.

It's online, it's in classroom, and importantly, it's applied learning. Again, teaching people how they use the insights to impact their jobs. They use competitions, they use rewards. I'm a big fan of that to encourage participation. They focus on analysis.

Again, how somebody uses this to impact the business. So they don't just teach tools, they teach real-world challenges from the business. they measure the results.

They have certification as a way to make sure that people have the skills that they need. So I really like that one. But how do you reinforce change? We all know that change management is the rub, right?

That is how you make progress. That's how you deliver value. And we all know that that's a monumental task in most organizations.

So I was an expert on a panel. I was moderating an expert panel. a few years ago and somebody from the audience asked the panel, you know, how long does it take for change to be realized?

And one of the other panelists was a CEO of a public services company and said, well, whenever the last person in the old regime dies. He literally said that. You know, of course I was thinking that's metaphorical, I don't think he's gonna be killing anybody, but it does indicate that, you know, change takes time.

It's hard. We need to be patient. We need to be, again, empathetic. High-performing leaders here, you know, provide people, again, back to sort of the value stories to help them understand, you know, what's in it for them.

They inspire them to want to make the change. Again, not easy. You know, one of the ways to bring people along is through stages to reinforce for them sort of the from to because narrative.

You know again making it easy and possible for them to understand how they make that transition, reduce their friction, you know be clear that the new normal yet may be scary but it's going to be amazing for you. Reinforce it, show them how to how to make the way. And here's the big one, and this is not just DNA, have the bravery to shut stuff down when it's time. You know Get rid of the old stuff once for all. Clearly there'll be in some cases stuff that's just too risky, you know, to shut down.

If you've got regulatory reporting that, you know, we understand that. But majority of things, you know, give it an end date. And don't give people the choice to remain in the past.

And you know what? It's okay not to bring absolutely everyone along. There'll be folks that you just can't bring along and that's okay. You know, you can't please all the people all the time.

But, you know, through these steps... Hopefully you'll get people along and again measure them. Measure them.

So, best habit number or high performing habit number three. Drive business innovation, leveraging emerging trends, use fast teams, set out audacious goals. Now you've been inundated with trends this week, no doubt.

And by the way, I think we've sort of given you the general guidance that playing it safe doesn't mean ignoring trends, right? Doing anything, even just to learn, you know, if it aligns to your business, can sometimes be better than nothing. So high-performing data and analytics teams, again, they drive innovation. Sorry, they are pretty quick to deliver value, and if something isn't working, they shut it, in terms of a proof of concept, they're pretty quick to shut it down.

You know, over the past seven years, you know, AI has been, I would say, you know... pretty, I wouldn't say mainstream, but many forward-looking companies have started to do proof of concepts and implemented least predictive AI. But it's not uncommon to take six to eight to nine months to do a proof of concept on a model to try to say, oh, well, I have value. And then by the end of that long effort, you find that it has minimal impact. So I think some of the practices that we see, you saw it in Page Group, I'll talk a little bit about them in a second, 28-day cycle time.

for assessing if you should move forward, doing a quick proof of value. Again, having big goals, not just, oh, we're going to improve by 5%, but we, our goal is to improve by 90%. And if we get halfway there, it's good.

Now, here are our top trends for 2023. I am not going to go over them. Gareth Herschel delivered them yesterday. If you want to listen to that presentation, it's amazing. You can listen to it on replay. I think what you can see though is the trends align to the six habits and you know many of them are not so you wouldn't be surprising to you.

AI risk, you know optimizing value, data sharing, sustainability, humans don't forget them, platform and ecosystems which we'll talk about in a second. So here's the example again I'll go in a little bit more detail about Page Group. You know they again had a huge really impactful initial I would say it's not really everyday AI because it's a core process. You know, recruiters, a key thing that they do is write job descriptions. But, you know, they were, it's really one of my favorite examples because, you know, even though it is everyday AI, they had a significant impact.

You know, as just to refresh, they're a European job recruiter. They've got 8,000 recruiters. They were able to take the time it takes from writing a description from 90 to 5 minutes.

20 to 90 minutes to 5 minutes. But the real story about this, again, is how they did it. So they have a fast cycle innovation lab, they call it their data lab, where they assess new proof of concepts in 28 days. If they're not able to demonstrate value after 28 days, they move on. Now, in some cases, they'll say, we'll reassess later, like if it's a matter of a cost-benefit type of problem.

But 28 days. If they have to make decisions and people aren't in the office or they're on vacation, their policy is to just make decisions with whomever is there on that day. So a very, very good process.

And again, major goals. The big thing for them, the AI-ready data. So they had been working...

Working on AI for the past seven years and they've been working on their data fabric for the past four and a half years so that they credit the data fabric with them being able to get to the five minutes so they were able to achieve a certain amount of productivity gain from the generative AI piece but because they had to pull data from four different systems that had never been integrated before to pre-populate these job descriptions they were really able to get to that five minutes. because of the AI-ready data, and they're going to be able to scale it. So high-performing habit number four, build data analytics and AI products, not projects.

Now, there are three types of archetypes, if you will, for data products. There's data as a service, there's sort of value stream insights, and there are essentially information products and services that you might sell. external to your enterprise to create a revenue stream. Again, I think many of you will ultimately do that. Now, that doesn't mean you need to turn everything into a product, but identify those product opportunities by really being, again, customer-centric.

Be that customer advocate. Understand requirements. Again, your customer may be internal. They may be external if you're over to the far right. Basically, you know, promote agile delivery practices as you would if you were a product manager creating a product that's going to have a SKU, that's going to have to be updated, that you're going to have to enhance over time, that you're going to have to maintain.

The idea is you want to create reusable data products that, you know, have a lot of traction across the organization, not just one-offs. You know, ideally, if you can hire people with product management skills. That's helpful.

And again, I don't think this just applies to data and analytics, but we see that high-performing data and analytics leaders do this thing. So case study. Again, most companies, if I'm being honest, aren't taking the data they have, the unique data they have, combining it potentially with other data that's external to create unique data sets that they can monetize.

Either by selling the data by itself or putting analytics on top of it and creating some sort of analytic service. Of course, many use obviously data products to create value. Again, these reusable data sets, but it's the same process.

ZF Group actually does this. I think they offer a best-in-class approach to data product creation that any company can adapt to sort of kick-start the process. So what they did is to monetize the data that's generated by its systems and first identified a business problem, right, where the data that they have could absolutely map to.

And then they gathered the data and created these new products, making sure that the business understood how to market and sell these new products. And of course, they measure the impact and adjust and evolve the product. They enhance it. They maintain it over time. So you may see it looks pretty similar to the infinity loop, what they did, essentially delivering a product, measuring the impact, adjusting and evolving.

over time they measure and have KPIs, OKRs, to really measure the results of the impact that this is having. So really best in class. This is actually a case study on Gartner.com if you want to learn more details about it.

So high business impact number five. Basically building a composable data foundation, ecosystem that essentially enables fast response to business. and technology changes. Now, this is a representation of a data and analytics cloud ecosystem.

And what we know is that data and analytics cloud ecosystems versus on-premises platforms in the past support more integrated workflows from data to analytics to AI and machine learning to applications. That's sort of the value proposition of having a lot of the hyperscalers are really touting. Versus stacks, on-premises stacks in the past, you could get the set of capabilities that you need, the portfolio of capabilities, but the integration was often very light, discrete products, sort of integrated at the invoice level often. But what you see now is that these sort of ecosystems of capabilities prioritize workflow integration, prioritize integration with ISVs to fill in gaps where they...

They may be light or to integrate with your existing environment. And so, you know, I think what's important here is, again, AI-ready data. AI-ready analytics is a foundation.

I think we're at a point where we really can't kick that data can down the road. So this becomes a very important foundation. Again, look at PageGroup.

And many of our clients, I think, are really there. having made that sort of investment, even if it's parallel to as you're developing these AI use cases. I know many of you will say, well, I'm not going to just develop this all at once, take three years to develop my data-ready foundation. No, you have to deliver value now. So you might have to do it sort of in parallel to high-value use cases.

Again, AI-ready data. This is a really, I love this slide, and we talk a lot about this in our team, where you have traditional data management practices. And those are amazing, but AI requires sort of additional capabilities, whether it's labeling and bias mitigation and different types of security, and even if you're moving into generative AI with enterprise search, different types of technology like embeddings and vector databases, etc., data, which is data enrichment, knowledge graphs.

At the same time, people building out these models typically focus on model accuracy. How many models do I deploy in production? They need also to get a sense for data and the importance of AI-ready data.

And so we have to have a much more integrated process. Definitely high-performing leaders put in place processes to scale because it's very different to do something in production versus scaling it and maintaining it over time, particularly when we're talking about AI. You need data ops and DevOps and AI ops and ML ops particularly.

you know, processes that allow you to scale over time and particularly incorporate new technology as it evolves, which we'll certainly see with generative AI. Great example here, put in place a framework that allows you to, again, model and scale operations. Fidelity has done a great job here.

They've managed to increase speed to production by, you know, double the speed, you know, essentially reduce the time by, you know, 100%. And they've been able to resolve product issues and reduce the time it takes by 80% by putting in place these kind of processes. So last one, I'll be really quick, governance.

And unfortunately, this is probably one of the more important ones, right? Data and analytics leaders, AI leaders treat governance and risk management as an essential value driver, right? There's nothing more important here. You know, where you need to articulate foundational investments in terms of, in terms that business peers can understand. Now usually when we mention governance to the business, what happens?

Have any idea? You probably have seen it. Eyes glaze over.

They may roll in the back of their heads. They may even fall asleep. And they may just say, no, you're trying to control me. And so we don't want that. But there is a way to talk about how governance is so critical to most of most every use case that we're delivering and we need to tell that story.

And by the way it doesn't mean that we have to set up this big upfront governance practice. It doesn't mean that we you know catalog everything before we can start day one. It doesn't mean that we enlist data stewards before we even you know think about analytics. It has to be a much more agile process.

The way I view it and the way I think we have to sell it is governance as a team sport. It is essentially the enterprise trust initiative. So I almost don't use governance with business people.

I say, you like to be trusted? You want to be trusted externally? You want to have an excellent reputation? Then you need to have governance.

And that's how we need to sell. It is the enterprise trust initiative. Everyone wants to be trusted.

So I'm going to just focus on this one, given we just have a little bit of time. So this one happens to be an AI one. It's a great example, again, AI governance. They are, in this case, Axon is a global leader in connected public services. They've assembled an independent AI board, which I think is really interesting.

The AI Ethics Board really helps them in the development of their AI-powered devices. Their development team has access to them. Their ombudsman has access to them.

And this board sort of over-represents civil liberties. privacy voices and concerns. It shares its feedback with Axon but also publicly. So Axon credits this board with really helping them with their AI strategy.

They feel like it's a crucial part of their sort of AI first development strategy. So as a reminder we covered you know business and business build a value centric data and analytics and I native strategy. Prioritize talent, skills, change management as core competencies. Again, you can't deliver value if you don't have the right skills, if people don't know how to use the analytics that you're making available to them. Build, of course, drive innovation by leveraging emerging trends.

One of the aspects of an excellent strategy is that you incorporate trends and drivers in that strategy. Many strategy documents that we review don't have that. Set audacious goals.

Think about fast innovation cycle times. Build products, not projects. Build a scalable, composable, agile data analytics and AI ecosystem and foundation. Focus on AI-ready data. Focus on processes that allow you to scale, not just do proof of concepts.

And then finally, treat governance and risk management as a key value driver as it is. So I'll just leave you with one question. Given all of this, maybe you can think about what type of leader will you be?

Hopefully, you will be both a business driver and an enabler to help your organization shift from being data-centric to being value-centric. So thank you for listening. I'm really excited about you all being here with us this week. Hopefully, you have a safe trip home and enjoy the rest of the conference if you're going to stay for another session.

Thank you.

Transcript for:Essential Habits for Data-Focused Leaders

Transcript for:
Essential Habits for Data-Focused Leaders