Toggle Menu
[Webcast] How to Raise Your Data Driven Game in 7 Steps

[Webcast] How to Raise Your Data Driven Game in 7 Steps

The surprising effectiveness of Enterprise IT in rallying to meet the challenges caused by the coronavirus has shown that aiming higher and moving faster was within reach all along. In this recorded webcast, we outline 7 data-driven steps to raise your data-driven game.

  • How has adapting to the implications of the pandemic impacted productivity and the pace of innovation?
  • Where is the “low hanging fruit” to raise your organization’s game on creating data products?
  • What equips organizations to excel at creating new revenue streams from data?

Introduction

Bryan Kirschner (00:00):
My name's Bryan Kirschner and I work on strategy at DataStax. In our last webinar, we looked at patterns and practices associated with achieving competitive advantage as a data-driven enterprise. In this webinar, we're going to do something a little different. In our last survey that we just completed in October, out of more than 500 organizations in the US, zero, precisely none said data strategy was not a priority.

Bryan Kirschner (00:31):
So we went and gathered seven specific actions you can take no matter where you are in your journey to help you make some progress on it. We assembled some data to use as a friendly provocation, not just for you the viewer today, but maybe for people and teams in your organization. So I decided to go there. I've been holding off on retweeting or using the who led the digital transformation of your company, the CEO, CTO, or COVID 19 meme, but since the fine folks at MIT Sloan Management Review who do really good work have tweeted it, I decided, okay, it's fair game.

Bryan Kirschner (01:08):
But there's some truth behind it, because, for all the suffering that this pandemic has caused, it has also unlocked more potential in enterprises and in IT than we thought was possible before the crisis. So for example, in this time of disruption, many companies stepped up, their employees noticed and it showed up in data. This is a longitudinal study of culture ratings on Glassdoor that one firm does. And for some companies, culture ratings spiked because of the constructive way they helped their employees through the initial disruption.

Bryan Kirschner (01:49):
For IT in the Harvey Nash, KPMG CIO Survey, almost two-thirds think the response to the pandemic has permanently increased the influence of the technology leader. Now, if you think back about the history of enterprise IT, by some measures, the failure rate of large projects was 70%. Yet, in the face of a disruption that occurred with barely any ability to plan, under tremendous levels of stress, IT stepped up to make most companies successful at implementing remote work faster than anybody believed possible.

Bryan Kirschner (02:28):
So at the same time, you're getting this over 80% success shifting to remote work, but almost exactly the same number of technology leaders are concerned with the mental health of their team. So let's take the glass half full and the glass half empty. How can we build on the potential that's been unlocked by breaking free of inertia and discovering we could do more than we thought we could with the stress? And not just the stress of the current crisis, but the stress of adapting to faster change, the stress of working in new ways and figuring out how to come back with all the good and not the bad.

Bryan Kirschner (03:07):
So onto the seven specific activities, we suggest. Carpe diem First, seize the day. So as companies have adapted over the months since the initial shock, the percentage of companies saying that the adaptation to the crisis has increased productivity and also increased the pace of innovation has risen. So it's Two-thirds say on balance, productivity has increased and the same percentage say the pace of innovation has increased, and a third of companies strongly agree. So it's a big notable change.

Bryan Kirschner (03:44):
We should seize this opportunity to bank what's been good and also figure out how we come back and build on it. So today, if we're being extra productive working remotely, what are the great things about working together, for example, in person that we can build back in, so we're even more productive? This to me was a pretty striking provocation. So I think all of us, my organization included, yours as leaders in technology, your business, our HR partners, facilities partners, I think we all probably thought we were doing our best to make employees happy and productive prior to the crisis. Yet, in the survey from 451 taken just recently, almost three quarters, over three quarters of employees say they have welcomed the changes made in the ways of working as a result of the crisis.

Bryan Kirschner (04:37):
So again, we got a shock, a massive shock to the system, a massive negative shock to the system. And people said, oh, things got better in terms of being happy and productive, at least from a way of working standpoint. So the provocation, the data-driven provocation is if we could do that without planning facing economic headwinds, as we go back and get beyond the crisis, six months into whenever we get to that point, will 80% of your employees in your organization welcome whatever the new changes that are that you've introduced to raise the game beyond where we are today?

Bryan Kirschner (05:15):
And that may be physical and technical ways of working in person and remote tools and processes. It may also be how you work together. So progress toward a data-driven enterprise is dependent on shared accountability between IT and business domain specialists. So maybe it's about how we take the productivity gains or the opportunities to innovate and apply that intentionally to getting your teams new ways of working that raise their games.

Sprint

Bryan Kirschner (05:47):
The second is, again, let's take some of this opportunity, this productivity dividend and plow it into a sprint. And I mean a specific type of sprint. So in our latest data, we let the data speak for itself. We took several dozen attributes of a data-driven enterprise and found that they were merged into five clusters. So if you organize those clusters from least data-driven to most data-driven, including a whole bunch of patterns and practices both business and technical, one, you see their belief that data is a competitive advantage tracking this cluster analysis.

Bryan Kirschner (06:28):
It also predicts how much revenue companies are deriving from data and analytics, which is not part of the definition. So we feel really good about using this as a framework to detect best practices. So using net scores to understand that spread between high confidence data is a competitive advantage, and essentially almost no one feeling confident data is a competitive advantage. And we look at why would this be? So here's a super pronounced pattern.

Bryan Kirschner (06:56):
The net score for a belief that the organization has clear data product owners or not. Almost universal at the high end are the most data-driven companies who are deriving the most revenue from data and analytics today. And, data governance by domain, where the work happens versus somewhere else. So using data to deliver value back to customers, partners, employees is very domain-specific. It doesn't just have to be somebody's job. It has to be the job of someone, whether that's an embedded technologist or whether that's someone in the business side or a cooperative effort, someone who understands what among that data is the most potentially valuable. What among that data plays into key processes, all data is not created equal. It's going to vary by domain context.

Bryan Kirschner (07:48):
So to build that skill of being able to ask out of all the data, how do you [inaudible 00:07:53] out the value of the potential value? How do you find the fit for a process or a customer experience to make things better? Do a sprint on what, one of the things we call the right questions to ask. The right questions to ask are grounded in that value discovery process. So, one for example, do we know what data about interactions with us our customers will really value? What among our processes are the most error-prone that we can bring data to bear on improving or fixing? What's stopping us from shipping data products as fast as apps?

Bryan Kirschner (08:31):
Any one of these questions is a great collaborative effort to focus the energy in your organization on understanding how to turn data into value. Building that skill can be a snowball effect where one successful sprint, one successful team, one successful collaboration brings in others. It demonstrates it can be done versus in some cases maybe having silos that are so vast that no one has the answer about what data interactions with customers might ask for, because it hasn't been a priority. It hasn't been something practiced in the past.

Bryan Kirschner (09:08):
So by the start of next year, can you complete one sprint and have a hands-on partner? If you're on the IT side, someone from the business side, if you're on the business side, someone from the IT side who can actually read the case study for here's how we work together. Here's how we find a great valuable use case for some data that we already have. Read. Share. Write, then Test.

Bryan Kirschner (09:27):
Also, don't just look inside, look outside. So read, share, write, and then test. What do I mean by that? These are a couple books that everybody should probably read. The one on the right's a little older, Platform Revolution about two-sided platforms and platform economics and the impacts for business. The other is a little more recent, Competing In The age of AI, looking ahead to how we all become AI factories and make the most of it.

Bryan Kirschner (09:57):
It's not that these things are necessarily right. It's not that these things will have some cookie cutter strategy that your organization should pursue, but they encompass concepts, you, and I mean, broadly you, all of us who help run companies and our boards and advisors and stakeholders, these concepts and tools we should be aware of. So we can't actually assess the strategic horizon for what we do without being aware of how two-sided platform economics work and how AI changes, where the value lies in the value chain.

Bryan Kirschner (10:34):
And again, it's not an assertion these are 100% correct, but they're all written by folks who teach at some of the best business schools. So, the competitive context, the context in which we operate are folks who are coming out of school to run companies like ours are going to be versed in these concepts too. So we at least need that common vocabulary. And there's a great anecdote, Jes Staley from Barclay's actually did a fireside chat with Geoffrey Parker, one of the coauthors of Platform Revolution and said, your book completely changed the direction that I was thinking about and how we grow the bank and got me realizing that one of my executives was communicating a message that I just didn't get.

Bryan Kirschner (11:24):
And so, his point was, they have the largest consumer banking customer base in the UK and his original view was okay, we're saturated, so we will go replicate that and build a massive consumer base in another country. That's not a bad strategy necessarily, but it turns its back on, oh, if we have this large customer base, how could we use data to create more value among the customer base? How could we use that customer base to match offers not just from us, but potentially from other providers and build our business that way. It gave him that frame to think about a different way to grow.

Bryan Kirschner (12:06):
Likewise, if we think about not just the raw data, but the implications of how the data may change our business, I love this comment from folks at the National Geospatial Intelligence Agency. So they look at massive amounts of data in part to help protect national security here in the United States. They did an analysis a few years ago and concluded that there was so much more data, so many more sensors, if they didn't actually apply AI and machine learning, they would eventually wind up needing eight million more imagery analysts than they have today to deal with that volume of data. So of course they needed to start to think about automation and artificial intelligence. Otherwise, they'd simply have to let potential data that could be valuable, go by the wayside and be ignored.

Bryan Kirschner (12:54):
So challenge yourselves, understand these concepts and implications and what some of these folks believe about the future of different industries, and then create a point of view, write a plan. And the plan could say, you don't think we'll need eight million more analysts. The plan could say, we think we're fine, but you'll at least have a plan that you can come back to test against new information and get everybody on the same page about their beliefs.

Engage the BoD

Bryan Kirschner (13:22):
Fourth is engaging the board of directors. For some of you, the board of directors may already be engaged in data strategy. For some of you, this may be something that takes multiple quarters to build support for. But, unambiguously, if you look at progress toward being a data-driven enterprise, again, it becomes almost ubiquitous among those who made the most progress. And it's essentially a linear connection between having a board-level discussion and data strategy and making progress.

Bryan Kirschner (13:54):
At the end of the day, this makes sense, right? So among the most data-driven, a good number of them are now attributing more than 20% of their revenue to data and analytics. That doesn't happen by accident. And it's also a direction, a strategic choice that you actually want the board to weigh in on and have a point of view on. So, if it's not already, before 2021 ends, will you get a date on the calendar to do that first board discussion of data-driven strategy?

Get a grip on your data

Bryan Kirschner (14:27):
Point five, getting a grip on your data. One of the big challenges of getting to be a data-driven enterprise may actually just be the fundamentals of deciding it's somebody's job that we're going to invest in, that we have a plan and a strategy for handling our data. So I think I've used these quotes before as polar opposites almost in some of our webinars. These are interviews from this year of technology leaders here in the United States. And on the left is one whose company's organization is doing digital business on a massive scale. He's generating a terabyte of data a day, but it is nobody's job in the organization to do anything with that data.

Bryan Kirschner (15:10):
So they have data policies to control risk, to control cost. And so he does a rational thing given those policies, he just deletes it. The organization hasn't turned the corner to say, okay, let's start to try and understand, in that terabyte of data, what might our customers value? How might it fix our most error-prone process and so forth? Or, the most data-driven side is this pretty deep into the history of the data-driven concept. For two or three years on the right, data has been designated as a strategic asset and we've been working on it.

Bryan Kirschner (15:45):
But even among folks further down the journey, most companies are not using all of their internal data effectively yet. So we broke out not five buckets, but three to keep it simple. Wherever you are, can you move the needle to get into the upper 50% of like companies? And by that I mean, if you look at the bucket of folks who've made the least progress to data-driven enterprise, or maybe barely started, all these organizations scored themselves on a five-point scale. So very few of the least data-driven companies today, only 3% strongly agree that we make good use of internal data.

Bryan Kirschner (16:31):
You cross into that 50% threshold at a three. So this is a very simple metric you can use with your teams, stakeholders across the company to say, how good are we at using internal data? Are we great? Are we not? Are we not even on the boards yet like the 4% is strongly disagree in that category. But you can make progress. You can get into that top 50% within the number of companies in this stage of the journey.

Bryan Kirschner (17:00):
Likewise, at the middle cluster, the folks who've been at it awhile, you cross into that 50% threshold at a four. So pretty good. And finally, the leaders setting the pace, more than two thirds say five, it's strongly agree and that dovetails with the conviction that it's a competitive advantage. So wherever you are, find out where you're at in your journey and make some concrete progress. You don't have to go from zero to the top immediately, but commit to it, track it, actually use some simple survey metrics or interviews to take a pulse and assign a number to it and see if you can make progress.

Look beyond data “Wild West”

Bryan Kirschner (17:47):
Now, we get into the technical side of things to the extent we'll talk about technology. Look beyond the data Wild West. So since this is a US-centric phrase, I want to make sure everybody is familiar with it. The dictionary definition is, it's the period in the Western United States characterized by roughness and lawlessness. So anything goes in the Wild West. And we actually did, because we fielded this in the US, we use that rubric and we asked folks, do you believe, do you agree that your data layer in your organization is the Wild West? And we looked at the net of agree versus disagree.

Bryan Kirschner (18:30):
And this is super right interesting because you see a peak in folks in the middle, and this makes intuitive sense. These are folks who've made progress towards using internal data, toward understanding how to productize data. And they're probably running up against the barriers, the limits of their legacy approach to data and working that out. But you also see it pop back up among the most data-driven. In our experience, conversations with people at the cutting edge, they also have made a lot of progress, but they're thinking about what's next? How do we take this to the next level of data velocity and so on. I'll come back to that in a minute.

Bryan Kirschner (19:10):
But the big risk, the big problem is the folks at the back end, the least data-driven folks who only marginally tilt toward believing their data environment is the Wild West. And in some cases in the second cluster, they actually disagree on the balance that their data environment is the Wild West. Well, here's the thing, if you're like most organizations, the default for your data layer is "the Wild West."

Bryan Kirschner (19:35):
So we've interviewed dozens and dozens of people, both from the executive level to the manager level and the hands-on practitioner level about data and databases. And they will be honest when they're going to be attributed anonymously. It should be a consideration of the best database, but it's just not. We use Cassandra in the case of this CIO because the architect knew it and fought for it. Usually I just say, whatever guys, just make the requirements work. Our developers don't think about operations. Our databases are hodgepodge because people in a particular project are making the best choice they think for that project and whatever context they have, not because they're bad people, but those are the tools they have to make a choice. So if there isn't a clear strategy, the odds are you have a hodgepodge, you have a lot of local optima at best.

Bryan Kirschner (20:30):
And we see this cascading through awareness and action. And so, the leaders have invested a lot in optimizing for data velocity. So we've probably, most of us have seen the evolution of DevOps and application development velocity and DevOps practices and tools make a big difference in how fast you can ship better quality code. We're going through, I think the same process of data ops, what people call data ops emerging as the practices and patterns and tools that help drive data velocity into organization to be able to deliver data products or data features as fast as apps and code.

Bryan Kirschner (21:14):
So the framework that I found really compelling is the idea of a data mesh. And what we hear from the leaders is they're looking ahead to how that data infrastructure and architecture becomes something like a data mesh, which has a lot in common with the idea of a service mesh, but goes beyond it and is well worth forming a point of view on. Something you can read, share among your teams, talk about forming a point of view wherever you are, it looks like it's the shape of what's in the future, whether that's near term, if your organization is like our friend on the right, who's already been at DD for three years, or maybe it's a little further out like our friend on the left whose organization still hasn't turned the corner on data strategy. This is Zhamak Dehghani who writes at martinfowler.com. In this case, I highly, highly recommend just diving in investigating her point of view in data mesh in general.

Lean into skilling up

Bryan Kirschner (22:10):
Finally, speaking of learning, lean into skilling up. This is an opportunity because at a very concrete data-driven level, part of that increase in productivity companies are seeing is because working from home has cut out a lot of commuting time in fact. There's been a pretty big economic study, about just how much time has been carved out of commuting. And about a third of it, people are devoting back into the primary job and they're devoting other things to childcare and so on and so forth, but about a third, they're putting into their primary job.

Bryan Kirschner (22:44):
So don't let it happen randomly. This is not particularly about channeling that time, but I was really struck by a CIO in my own home state here in Washington, where if you're a non-tech company, at least to the extent geographic placement matters today, you've got to compete with Amazon, Microsoft, the big and growing Google engineering center, Tableau, others like ... If you want to hire some of the most in-demand talent, you have to have a strategy.

Bryan Kirschner (23:12):
And so, this is a non-tech company and he said, you know what they did. They went out, they had to hire two very expensive senior data science people they would maybe not otherwise have hired, but it was a strategy because they're the nucleus that gives them the credibility and the ability to now do college hiring for those skills because they've gotten mentorship, they've got a career path and so on. So, that's a strategy. That may not be your strategy, but make a strategy for building the skills you need.

Bryan Kirschner (23:39):
In particular, near and dear to our heart, embrace the modern, what we call the modern data stack. So conceptually, if you think about the experiences that data-driven enterprises deliver, real-time, personalization and recommendation, real matching that with real-time inventory tracking and management or logistics, and so on, these are cloud-native microservices. They're reactive to changes actions by the client and changes on the backend. They're streaming data-friendly, sending lots of data back, taking lots of data in. And they're often conceptually if not literally graph-oriented, right?

Bryan Kirschner (24:19):
What have you put in your cart? What interactions have you had with the brand across marketing and across support and sales and email and chat? So that's the shape of these modern data-friendly experiences and this stack is great for doing that. This is what we see the leaders using to deliver those experiences, Apache Cassandra, Kubernetes, Kafka, and so on.

Bryan Kirschner (24:39):
So, these things are probably in your future if they're not in your present. Probably more of these things are in your future if they're not in your present, as you progress down that road to being a data-driven enterprise. So take advantage of the current conditions where potentially you have slack time, or we have a productivity dividend to get people to skill up on these things.

Bryan Kirschner (25:02):
And of course, there's been such a massive shift to online education and training that again, the burden of getting people someplace cost-wise, lifestyle-wise, productivity-wise has been reduced. So come up with a plan to make the most of those conditions. Strategically, what's the talent you need? How much of it? How do you get there with the people who have or the people you hire?

Bryan Kirschner (25:29):
So, as we come up on the half-hour mark, to recap, seven data-driven steps to raising your data-driven game, all presented with the data behind them, the observations behind them, again, intended to be a friendly provocation, a stimulus you can think through yourself, but also potentially share across your organization to pick what makes sense and actually go get something done, get people to roll up their sleeves, build some of these skills, build some talent, and then come back, hopefully given the favorable indications around the vaccine, come back into a new, better way of working.

Q/A

Bryan Kirschner (26:09):
So, thanks for joining us today. We are now able to take questions if we have any questions. There's one here. What does it mean more concretely to make data strategy a board-level discussion? So I have two things I'll say about that. I think we've seen, or if you've lived through the last 10 years, maybe 10, 12, we've seen cyber, cybersecurity rise to a board level discussion. Then we saw more recently digital transformation rise to a board-level discussion. So I think, if history rhymes, future rhymes with history, I think we will see data-driven enterprise rise to a board-level discussion.

Bryan Kirschner (27:10):
So it's more about getting one's arms around the fact that this is coming and being proactive. Substantively, what does that mean? How do you do it? I would say, it'll look different for every company, but I think there are two things that stick out to me, right? One is, many companies are attributing 20% or more of their revenue and an even larger number, 10% or more of their revenue to data and analytics today. So it's possible. And if your company is aiming for that or not aiming for that, that's a choice. And probably when you talk about at least 20% is probably a board level engagement and choice to discuss. So there's scenario planning or strategically planning, does that lie in your future?

Bryan Kirschner (27:59):
And then, there's also just a question about what digital transformation and data opens up from a feature business standpoint. It may not be as extreme as NGA saying that the growth in imagery data is so vast that we need eight million analysts. But, if you think about how some CPG consumer packaged goods companies are responding to the current disruption, some of them are moving directly to consumers where they didn't plan that before. And you can see a pattern or a paradigm for doing that. Part of Nike's digital transformation is becoming direct to customers.

Bryan Kirschner (28:38):
So, does data open up a different business model discussion that would be a board level scenario planning to think about two or three different things? Again, will look different for every company, but it's probably worth getting a handle around how big do you think the changes are? And if we don't think they're very big, probably talking to the board about our data strategy is based on change not being that big.

Bryan Kirschner (29:21):
Something better than the Wild West of data sounds good but getting developers to, for sure think more carefully and being more thoughtful about databases seems like a heavy lift. Exactly, it is. And if you dig into data mesh and the concepts of building a data platform, it's all about abstracting away that choice and providing affordances that give them the type of data storage and access and service that they need, but more from more of a platform perspective so that their choices are constrained, but they're constrained to things that become an obvious default as opposed to something that prevents them from getting things done.

Bryan Kirschner (30:20):
And okay, one more, we're just after the half-hour, but I'll do this one quickly. What about if you're not one of the companies that have had a productivity increase, or I'll add in lost ground? I think the numbers keep shifting toward productivity increases. So I would say that data suggests, don't accept that circumstance. Obviously, if you're in a vertical that's been really hit hard by the current disruption, it can be challenging. As one executive said last month in retail sector who are, there's a glass half full and digital is a glass half empty, and everything else that's happened, we're trying to be very mindful and it's a little brutal, but they are wrestling with how to move ahead.

Bryan Kirschner (31:14):
As not having productivity and innovation gains become a smaller, smaller slice of the pie, a more minority position in the market, the potential pain of letting it happen versus suffering some pain to get your arms around and then turn things around, that starts to shift. So I would say don't accept that status quo, that's how it is today.

Conclusion

Bryan Kirschner (31:46):
All right, so I don't see any other questions. I think I will say thank you for coming, and we'll wrap up this webinar. Thanks, have a great day.

Open-Source, Scale-Out, Cloud-Native NoSQL Database

Astra DB is scale-out NoSQL built on Apache Cassandra.™ Handle any workload with zero downtime and zero lock-in at global scale.

Astra Product Screen