Today I’m chatting with Vin Vashishta, Founder of V Squared. Vin believes that with methodical strategic planning, companies can prepare for continuous transformation by removing the silos that exist between leadership, data, AI, and product teams. How can these barriers be overcome, and what is the impact of doing so? Vin answers those questions and more, explaining why process disruption is necessary for long-term success and gives real-world examples of companies who are adopting these strategies.
Highlights/ Skip to:
- What the AI ‘Last Mile’ Problem is (03:09)
- Why Vin sees so many businesses are reevaluating their offerings and realigning with their core business model (09:01)
- Why every company today is struggling to figure out how to bridge the gap between data, product, and business value (14:25)
- How the skillsets needed for success are evolving for data, product, and business leaders (14:40)
- Vin’s process when he’s helping a team with a data strategy, and what the end result looks like (21:53)
- Why digital transformation is dead, and how to reframe what business transformation means in today’s day and age (25:03)
- How Airbnb used data to inform their overall strategy to survive during a time of massive industry disruption, and how those strategies can be used by others as a preventative measure (29:03)
- Unpacking how a data strategy leader can work backward from a high-level business strategy to determining actionable steps and use cases for ML and analytics (32:52)
- Who (what roles) are ultimately responsible in an ideal strategy planning session? (34:41)
- How the C-Suite can bridge business & data strategy and the impact the world’s largest companies are seeing as a result (36:01)
Quotes from Today’s Episode
- “And when you have that [core business & technology strategy] disconnect, technology goes in one direction, what the business needs and what customers need sort of lives outside of the silo.” – Vin Vashishta (06:06)
- “Why are we doing data and not just traditional software development? Why are we doing data science and not analytics? There has to be a justification because each one of these is more expensive than the last, each one is, you know, less certain.” – Vin Vashishta (10:36)
- “[The right people to train] are smart about the technology, but have also lived with the users, have some domain expertise, and the interest in making a bigger impact. Let’s put them in strategy roles.” – Vin Vashishta (18:58)
- “You know, this is never going to end. Transformation is continuous. I don’t call it digital transformation anymore because that’s making you think that this thing is somehow a once-in-a-generation change. It’s not. It’s once every five years now.” – Vin Vashishta (25:03)
- “When do you want to have those [business] opportunities done by? When do you want to have those objectives completed by? Well, then that tells you how fast you have to transform if you want to use each one of these different technologies.” – Vin Vashishta (25:37)
- “You’ve got to disrupt the process. Strategy planning is not the same anymore. Look at how Amazon does it. ... They are destroying their competitors because their strategy planning process is both expert and data model-driven.” – Vin Vashishta (33:44)
- “And one of the critical things for CDOs to do is tell stories with data to the board. When they sit in and talk to the board. They need to tell those stories about how one data point hit this one use case and the company made $4 million.” – Vin Vashishta (39:33)
- V Squared
- LinkedIn: https://www.linkedin.com/in/vineetvashishta/
- Twitter: https://twitter.com/v_vashishta
- YouTube channel: https://www.youtube.com/c/TheHighROIDataScientist
- Substack: https://vinvashishta.substack.com/
Brian: We did it. Today is the 100th episode of Experiencing Data, and I just wanted to reflect a little on this before we hop into the next interview. This show has been a place for me to learn and experiment with different questions across guests, spanning the spaces of data science, analytics, design, UX, and product management. I don’t share questions in advance with the guests and I hope that means that you’re getting a biweekly dose of unfiltered perspectives on what other leaders like you are experiencing in their work to build useful, usable, and indispensable data products, your Apple Podcast reviews, emails, LinkedIn messages, all that stuff, it’s really the best feedback I get about the show since, ironically, the analytics on podcasts aren’t particularly helpful. There’s really no recipe for making a podcast successful, and so please keep that feedback coming whether positive or negative. You can always reach me over at designingforanalytics.com/podcast.
And finally, I just wanted to send out some thank yous. First to the bandmates in my ensemble, Mr. Ho’s Orchestratica. They’re the ones that are performing my music that you sometimes hear playing in the background.
And secondly, I have to thank the folks over at Humble Pod, it’s humblepod.com, and particularly Chris Hill. He runs a really sharp team, and they handle all of the audio and production elements of Experiencing Data. I could not have published this show and the episodes so consistently without their support. And highly recommend them if you’re considering starting a podcast yourself. Their work really allows me to focus on the content of the show and the guests, and not all of the operations and logistics that goes into to publishing. So, thank you so much, Chris and team for your help, and Ashley, and all the other staff there. And of course you, without your ears, none of this matters. So, thank you for listening, and please keep those comments coming. And now, here’s episode 100.
Vin: This is Vin Vashishta, founder of V Squared, and you’re listening to Experiencing Data with Brian T. O’Neill.
Announcer: You’re now Experiencing Data with Brian O’Neill. Experiencing Data explores how product managers, analytics leaders, data scientists, and executives are looking at design and user experience as a way to make their custom enterprise data products and analytics applications more useful, usable, and valuable. And now here’s your host, the founder and principal of Designing for Analytics, Brian O’Neill.
Brian: Welcome back to Experiencing Data. This is Brian T. O’Neill. Today, I have Vin Vashishta on the line from V Squared. How are you, Vin?
Vin: I’m good. How are you doing today?
Brian: I’m doing great. And I wanted to talk to about data products and data strategy. These words are thrown all over the place, and to me, they don’t mean the same thing depending on who you talk to, so I just like hearing different people’s opinions about what these things are in hopes that maybe the audience can, like, paste them all together into some soup that tastes good and make their own thing. So, that’s what I wanted to jump in to talk to you about today. And the first question I want to ask you about was this—we talked about this AI last-mile problem, what is that?
Vin: Really, dragging projects over the finish line. That’s literally the AI last-mile problem. The best way to think about it is you’ve got this clunker—and especially when it’s the beginning, early stages of maturity—you got to clunker that you got to drag over the finish line because it’s been cobbled together, mistakes were made along the way. And so, we have two sides to this problem. The first side is how do we get this thing that we’ve built to deliver the business value that we need it to?
The other side is, how do we get it into the hands of users and customers? And so, those are the two sides of the AI last-mile problem. And what’s incredible is, there are companies who have gotten to very mature stages, they’re able to—they’re not delivering clunkers anymore; it’s not something that needs to be dragged across the finish line, they have mature processes in place, but monetization is a struggle, prioritizing, getting the right initiatives done, and then integration. How do you get it into the hands of customers and users? That is probably the first piece that matures, and once that happens, they start delivering to customers and customers say, “Well, yeah, that’s not interesting.” Users pick it up and they go, “Oh, this is awesome.” Then you talk to them six months later. “What was that again? Did you—was that—no, I don’t remember that. When did that happen?” And that’s where the second half of the last mile problem comes in, is that now we’re delivering product, but we’re not delivering value.
Brian: Yeah. So, dragging it over the finish line suggests either it’s giant and heavy and we can’t even lift it anymore, or nobody cares about it. Like, I’m very curious about this. When I talk about last mile, I’m usually talking to people that have technical backgrounds, and so they tend to think of the deployment and putting it out into the wild as being something that’s at the end. I actually think that’s the beginning of the game. It’s the start of the game. [laugh]. It’s like kickoff for the real game.
Everything else before that was rehearsal or practice, and now it’s just starting; it’s not actually the end. But how does it happen that there’s a question about whether someone would even want it or would consider using it? What is happening that we’re waiting until the end to find out that it’s not what anybody wants or needs?
Vin: Well, technical strategy is absent for most businesses. They have a business model, they have an operating model, and then they have technology kind of sitting off to the side and it’s not integrated. And so, when we do strategy planning, we’ve got all these great opportunities. Here’s what the business is going to do, here is core business strategy, and then technology comes in afterward. And it says, “Hey, we can do all of this other stuff,” and now the business is letting technology lead strategy.
And that’s the root cause is that piece right there isn’t integrated. And so, part of the opportunity identification process is disconnected from the way technology can help with each one of those opportunities. And when you have that disconnect, technology goes in one direction, what the business needs and what customers need sort of lives outside of the silo. And that’s where the AI last-mile problem—you know, when I talk about dragging it over the finish line, that’s where you get, it sounds amazing when we pitch it in a pitch deck and we say, “Oh, it’s going to do all these amazing things and it’s going to be rockin’ and you’re going to love it.” You know, slide one, ‘AI,’ slide two, ‘question mark,’ slide three, ‘profit.’
Vin: And that sells it and everyone loves it, and then they get it, or it gets close to being delivered and users actually start looking at it, and customers actually start giving feedback on it, and just the level of excitement plummets. And that’s why I say, “Drag it across the finish line.” And you kind of nailed it. You said, from a technology standpoint, the AI last-mile problem is that infrastructure, how you get it into production, how do you integrate it, how do you do all of those pieces? And so, that gets solved first, but what doesn’t [laugh] go back and get solved is what should we be delivering? How do we decide what opportunities are our best ones?
And so, there’s the other side of the AI last-mile problem that really should be the first one to be solved. It’s not as important to, you know, build all the infrastructure and everything else to deliver product if the product you’re delivering no one wants.
Vin: And that’s the dredging across the finish line.
Brian: You know, coming out of the design world—I mean, my roots are in design—and the way I frame this and talk about it to my audience and when I’m training people is that the first hurdle to business value is adoption. So, if the humans in a loop won’t use it or can’t use it or don’t care, then you’re never going to get to step two, which is business value. And most people, the data people, are always talking about business value with data, and they’re rarely talking about the adoption hurdle, which means there has to be utility, there has to be usability, there has to be perceived value in the eyes of the beholder, whatever that means. It doesn’t matter what we theoretically can say the value of it is because that’s all subjective thought. And users have their own mental model of the world and what’s helpful for them and what’s in it for them.
So, we attack that by understanding the problem space much earlier in the process, and continuously throughout the development of whatever the data product is going to be so that we minimize the risk of having to drag something over the line that no one may want at the end. We should not have a big reveal or a surprise by the time it goes in production. And to me, production means real human beings are touching and using it. It doesn’t mean we’re not in a QA environment; we’re in a production environment. Production to me sounds like it’s [laugh]—there’s value, there’s people using it, like, you can see things going back and forth. Like, it’s actually real. It’s not just sitting on a different server. Is that—tell me how you frame it or how do you think about it, addressing these issues earlier?
Vin: Airbnb CEO said something really interesting yesterday because he talked about going through the pandemic and surviving it and how they had to get real small. They had to shrink down their teams, they had to focus on core business. Part of it was figuring out what their core business was. And what they went through two years ago is what most companies now, if they’re confronting the AI last-mile problem, that’s what most companies are doing right now is they’re looking at each one of their initiatives and reevaluating it and saying, “Why are we doing so much stuff? Why are we so far away from our customers? Why are we so far away from our core business?” R at least that’s what they should be asking.
But I think that’s what we’re confronting. And so, when you say, you know, “How do we solve this sooner? How do we get to the front of this process?” It should happen at that opportunity evaluation stage. Someone needs to come in and say, “There is a ridiculous goldmine sitting in front of you. You have the talent, you have the infrastructure, you have data, you have access to run experiments and build some really, really accurate, reliable complex, internal, external operations products, you have opportunities.”
But we have to figure out what opportunities the business should pursue. And then from there, say, “Okay. Now, what can we do with all of this potential? What can we do with data? What can we do with analytics? What can we do with AI to help the business execute on those opportunities better?”
And I slam home this point: it has to be better. You can’t just throw data at it. Why? Why are we doing data and not just traditional software development? Why are we doing data science and not analytics? There has to be a justification because each one of these is more expensive than the last, each one is, you know, less certain.
So, solving the problem is going back a little bit and making sure that the connection is there from the beginning, that all we care about is customers, is value, is whatever the business has decided that’s what they want to go for, that’s what we care about. That’s what we’re going to do. And it doesn’t matter, yeah, we could have, like, 50,000 things down the road that could be done that could be amazing, but who cares? Let’s do today, and plan for tomorrow. Let’s set us up.
And I think that’s, you know, when you talk about where can we go to fix it, that’s where we go. We get very, very far up into the strategy planning process, we embed ourselves in there, we become partners. And not just, you know, at a product management standpoint. Having data scientists sit in on strategy planning and start talking about, “Okay, so you want to do that? Here’s what we can do for you. Here’s what we can offer you. Let’s think about that.”
Brian: Typically, I mean, the roles and activities—the activities you’re talking about to me, that is typically the domain of product management. And I agree with you that my model for this has kind of gone from a three-legged stool to four. In mature digital spaces, right, you have product engineering—product management, kind of at the top running the team, you’ve got product engineering, and you have product design. And this trio worked together. I think with data products, data science or analytics or someone with a data strategy background, some there’s a different expertise there that’s distinct from software engineering, so now it’s a four-legged stool.
One of the challenges that I see is I don’t think a lot of very skilled people that understand the technical side of data science really well are particularly suited at doing the product management work without training, guidance, coaching; it’s a different skill set. And you talked about, which problems should we solve? The should we solve, and is it viable, and will it be valuable all these kinds of squishy questions is the domain of product to me. So, I don’t see a lot of companies yet hiring data product managers to do this work yet—and I’m now I’m kind of shifting to the enterprise—the large internal data teams servicing internal customers. They don’t have this skill, yet they continue to have the problem associated with it.
I’m kind of like, “Why do you keep doing this?” And you keep talking about how mature the digital natives are and they’re kicking everybody’s butt, but you’re not copying the model they have. And that model has product management, product design, technical expertise, and probably data science if it has anything to do with machine learning or AI—or even analytics—there’s going to be someone with data there. All you have is the data leg of the chair. I’m not seeing—there’s no design, there’s no user experience, and there’s no product management happening. Those skills aren’t there.
Some people are trying to get to that through training, some people are doing some hiring. I’m starting to hear changes, like, people are saying we want to do this product approach to data stuff. What is the gap there? Do you think the gap is going to be closed through training here? Are teams going to start hiring product people? Or are they just going to keep reverting back to what they [laugh] were doing yesterday because the status quo is—I don’t know. It just shocks me because they’re saying, “I have this problem.”
And it’s like, well, there’s demonstrated solutions that work. This is not new stuff. If you’re coming out of the digital world, this is old table-stakes ways of doing things. You need the discrete roles involved with these products. You don’t have them. Like, I don’t know, do you—[laugh] it’s just kind of like, I don’t get it. Talk to me about that. Like, do you see this too or is this just my little solar system, and I’m not aware of what’s going on outside that? I don’t know. I find it interesting.
Vin: This is everyone, literally—
Vin: —everyone. Literally.
Vin: I mean, on the three continents that I operated, it’s everyone.
Vin: On the startup SME Fortune 100, it’s everyone.
Vin: Face—I mean, Meta. Meta is dealing with this. Google is dealing with this right now, where they say they want to be—you know, Google said, “We want to be 20% more efficient.” But really what they’re saying, “Is we have teams working on stuff that we don’t understand, that isn’t producing any value, and we don’t know how we got here.”
Vin: So, if you listen to—and Meta is saying the same thing. “We have this amazing wealth of talent, but we’re not entirely sure what they’re all working on.” And so, what they’re confronting is what everyone else is confronting, too. So, what you’re seeing, completely common. And the solutions are—and you kind of hit it with that four-legged stool—where data science is different, and you know, most people say digital product, or you know, they kind of lump everything into digital.
And digital is the last wave. Digital is software, it’s web, its, you know, cloud, you could probably throw cloud into digital, although it almost morphs into its own thing. And so, you have a digital strategy, you need a cloud strategy. Data is different, so you need a data strategy. Analytics are different; that needs a strategy. AI, different; needs a strategy.
So, we’re in this place now where we have to have a level of separation, but we have to figure out a way to take all of these technical strategies and put them under one business umbrella so that C-level executives can look at one thing, not 15. Because this doesn’t stop. And so, when you’re talking about, you know, what are the extra skills that need to be brought in, definitely you need product managers that specialize in data science, in putting models into production and monetizing models, where people pay for inference. We need people who can understand the differences between that product and a traditional digital product and say, “You know, it’d be better if we did digital. It really would. We should just do this the old-fashioned way because it’ll get the job done, and it’s way cheaper. And we’re better at it, so let’s do it that way.”
And then you have other people who have a different skill set, who can say, “Wait, wait, wait, if we use machine learning for this, here is the benefit.” And that’s where we need this partnership, like you’re talking about, this four-legged stool, where it’s a conversation about, here’s the opportunity—you know, not here’s all of the opportunities we could be going after, but at the C-level, we’ve decided these are our opportunities; this is core business; this is what we’re doing. Don’t—you know, don’t try to drag me into something I’m not, as a business, ready to do. Here’s our opportunities. And then you have this four-legged stool comes together with different domain capabilities and different domain expertise, and they start saying, “Okay, given these opportunities, what’s the best technology? Is there a benefit here? Does that work?”
And that’s the conversation in the partnership that needs to happen in order for C-level strategy to really hone in on how and why it uses technology to create value, to deliver value, to improve productivity, all of those different core pieces. And you have a product management side of that, you have a technical strategy side of that, you have just this real, real need to get close to your customers, close to users. Drag data people in front of those [laugh] people who are going to use their products and say, “Look, no. This is what they do. I don’t care what you think they should do. Here’s where your customer is. Here’s where your user is. Meet them there. Don’t meet them on the moon because that’s not working.”
And it’s that, you know, there’s so much value in all of these traditional roles, but also in adding pieces, adding sort of these hybrids, where you have technical strategists, where you have—and how do you get those? You can’t hire them; they don’t exist; you have to train them. But you can hopefully pull somebody out of Amazon who does data science and machine learning product management—they have them—but that’s really expensive. And you want to look at a very large amount of money that you’re going to have to put into one person, that would be hiring out of Amazon or hiring out of Microsoft. So, it’s far more effective to do the training route.
Let’s get people who are smart about the technology, but have also lived with the users, have some domain expertise, and the interest in making a bigger impact. Let’s put them in strategy roles. Let’s put them in product management roles. Let’s look at people who are leading and we’re becoming strategic leaders, who are beginning to own pieces of the strategy process for those technology strategies that I talked about. Let’s make them, let’s put them into the executive leadership level, and let’s get them to own their piece of the strategy puzzle.
And let’s begin to develop these people give them training—definitely they need training, they need mentors, they need exposure to other organizations and other leaders—and let’s train them up. Let’s give them a career path. Let’s give them a learning path. Those are all critical pieces.
I mean, I built classes for this because this is the need that I see. I see that there’s just this huge gap. You know, everyone at the technology team is almost abandoned when it comes to leadership training, when it comes to strategy and business acumen. It’s almost like they are just left on an island and told, “Yeah, you need to have business acumen.” “Well, what is that?” “You know, it’s business acumen.”
Because the rest of the business just knows. It’s almost like the technology team talking about technology: they just know. So, when they say, “Yeah, we’re going to use cloud, probably Kubernetes. We’re going to—” you know, business users are like, “What?”
Vin: And it’s the same way.
Brian: Trying to [crosstalk 00:20:21] a restaurant?
Vin: You need business acumen. And [laugh] technology goes, “What?” And the business is like, “How do you not know?”
Brian: Yeah, as soon as I hear—when I see that in a strategy, I’m generally cringing because it’s like, it. It’s implementation, masking itself as strategy. And I think especially as you go up the C-level, these things should be almost more simple. They should be more simplified for a more generic audience. If I read your strategy, I should basically be—anyone should be able to understand what we’re talking about without any business domain knowledge.
I don’t even think it probably should have AI or machine learning or any of that because that suggests an implementation path which may or may not be relevant. Or it may or may not be the right first step because it’s like, well, if we can get this thing that we want at 100, well, if we can get it to 20, in the next two weeks instead of waiting four years to get it to 40, maybe we’ll do the thing that gets us to 20 because any improvement is demonstrable value. And so, someone like you might say, “We don’t need to do AI yet. Let’s get the two-week thing out and get to 20 and get the win, then go on to this other thing.” That’s all implementation stuff.
That shouldn’t be in the strategy discussion. The strategy should be a clear understanding of, like, current state, future state. Like, I should be able to see the change. The change is of this: there’s an increase in x of this, or a reduction of this over this time period.
What does it look like when you help a team with a data strategy? And I tend to think of it—there’s a business strategy, which really has a data and digital component to it; it still should be in service to an overall business strategy, which itself is in service to a customer. But what is the data strategy? What does the deliverable look like? What would I see?
If you were to come in and help me with my company’s data strategy, what would the end of the project be? Like, tell me a little bit about what that looks like. You can use an example or maybe just swap out some names, but like, I’m trying to visualize it for the audience. What does it look like? What kinds of things do I see? Is it a PowerPoint deck? Like, what—how high in the sky is it? How low? Like, give me an idea what it looks like.
Vin: Ugh, if it’s a PowerPoint deck, fire me—
Vin: —[laugh] literally that minute. If I ever deliver you a deck and say, “Here you go. Good luck.” [laugh]. You know, that might be the last time I ever work for anyone again.
So, we’ve got a business model, right?
Vin: Statement of monetization. Got an operating model, statement of why you—you know, how you create value, why you create it that way. You got to technology model that supports both. Why do we use technology? That’s a strategy. That’s it.
Don’t make it any harder than that. When you talk about high-level strategy planning, you’re not making it any more difficult than, why did we pursue that opportunity? Why are we going to use marketing? To pursue that opportunity. Now, you got a marketing strategy.
How are we going to leverage sales and our advantages with our particular sales organization to achieve that outcome, to achieve that opportunity? That’s your sales strategy. How can we leverage pricing? Do we have pricing power? Do we have a way to achieve pricing power? Do we need too? That’s your pricing strategy.
Technology, same thing. Create a technology model, and it explains why the business uses technology to achieve its strategic objectives, to go after those opportunities. And like I said, you have different technologies. So, underneath the technology model, underneath your technical strategy, which is that statement of value creation and how technology supports your business model, how does it support your operating model? How does it incorporate into products to make more money than it would have without technology there?
I mean, why do we use data? What’s the point? Why do we use analytics? What’s the point? Why not stop? Literally, answer me that question and that should be part of strategy planning. If I stopped doing everything technology-wise that we do right now, would there be a loss?
And you begin to really dice down your projects that way. It really just filters down almost immediately. It’s a prioritization. You start talking about core business. And so, that’s the artifact, is you got to technology model that explains how technology supports your business model, your operating model, you have a statement of value creation. You have a statement of monetization.
And you’re explaining why the business uses each one of the technologies that it does, and those are all the sub-strategies. What do you use cloud? Why do you use digital? Why do you use data? Why do you use analytics? AI? Soon, quantum? Why are you using a platform-based business model or operating model?
You know, this is never going to end. Transformation is continuous. I don’t call it digital transformation anymore because that’s making you think that this thing is somehow a once-in-a-generation change. It’s not. It’s once every five years now.
Vin: And when you begin to look at it that way, that’s what the engagement looks like. That’s what I’m leaving the business is planning for continuous transformation. So, what’s that roadmap? What’s the timeline? How do you build the timeline? How do you know when you need to have each one of these pieces in place?
Well, you have these core opportunities. It goes back to that. When do you want to have those opportunities done by? When do you want to have those objectives completed by? Well, then that tells you how fast you have to transform if you want to use each one of these different technologies.
And then how does your data strategy amplify your analytics strategy? How does your data strategy amplify your AI strategy? And you should be looking at, year over year, increasingly high annual recurring revenue from that investment. So, if I invest in data, year one, it should give me a certain amount of money, but when I move to analytics, when I move to AI, now that investment should be giving me that money that I got out of my data strategy and I should be making more on analytics because of the decisions I made in the last transformation. I should be making more with AI because of the decision I made two generations back. And that’s the process of continuous transformation.
And those are the pieces that when I work with clients, that’s what I’m putting in place is that top-level strategy so that, doesn’t matter, you know, 30 years from now, it doesn’t—you know, whatever crazy technology ends up coming, your core concept of having a technology model still works. You’re still talking about it in terms of value creation not technology creation. And that’s what it ends up looking like is tons of these roadmaps that align the entire business. Your transformation roadmap that aligns because there’s a difference between the technology is ready and the business is ready. There’s a difference between the technology is ready and the customers are ready. If product management doesn’t know what to prepare customers for, your [laugh]—that AI last-mile problem is more like a last… would you call it—
Brian: Marathon mile.
Brian: Marathon mile [laugh]. Or something.
Vin: [crosstalk 00:27:33] fifty mile problem. [laugh].
Brian: No, I understand. Is it the CDO’s job to own this data strategy and help create it? Because I guess one of the things I think about, too, is that if I’m carrying a machine-learning hammer, we don’t want to be just hitting every problem with it because I have a PhD in that, just like, you know, the answer can’t be always use this tool that I have. To me, there’s a framing of the business problem and then this is why product management as a role is important because product management shouldn’t be dictating the implementation method. It should be, “We need to see this increase. This is the users pain—” whether it’s an internal user or an actual paying customer—“This is what’s tough about this, so the use cases, the stories, are this.”
And it’s okay—in fact, I think it’s better when there is technical representation participating in that, but then they should be in service to those use cases, right? It shouldn’t be, like, “We’re going to try to use machine learning to do X.” There’s a space for experimentation; that’s not my department, and I understand the need to maybe run small experiments to learn how to do this stuff, but when we’re talking about a core business strategy, I find there’s still this disconnect where the data teams are getting these asks to do machine learning, to do AI, and they’re like, they don’t know what they want. So, whose job is it to surface the unarticulated needs? Like they’re telling us they want this, then we give it to them, then they don’t use it, and we drag it over the finish line.
And I just keep seeing this data—it’s the data tennis game. It’s just this ball just keeps getting rallied back and forth, [laugh] and it just keeps going back and forth. Can you unpack a little bit about how you get from the high-level strategy—how does the data product person get from the overall business strategy, which should have no implementation detail in it, down to something that becomes an actionable work increment for a data product team, where there’s data science work, there’s engineering work, there’s design work? How does that get broken down? Like, what are the skills?
What are the steps to do that? So, let’s say it’s like Airb—I’m just going to pick Air—you keep mentioning Airbnb. So, it’s Airbnb, and they’re like, “We want to increase top-line revenue by 20% and we think that the best way to do that is to be into new markets.” Like, “We need to be in new markets.” “Well, where? What markets?” Right?
So, maybe the start of it’s like, “Well, where should we go?” Right? Where should we get more people to open Airbnbs? What’s the opportunity there? You can probably imagine some data would be helpful to make a decision like that.
Unpack how the data product manager or whoever the role—or the CDO or whoever it is—takes it from that high level—let’s 20% increase in the next 12 months—and the hypothesis from the CEO or the leadership team is new markets is the best way to do that. Go. Like, what how would you unpack that conversation? And I’m just—or using your own hypothetical, but I’m trying to help people picture what those questions are, and how do we get it down to, “Go build this thing.” Which I think everyone’s ready to do. They got the hammer for that, and they’re so itching and ready to do that [laugh].
Vin: Yeah, I mean, Airbnb strategy was—like I said, this is indicative of where most companies are going because they had to go and answer the question at a higher level. It’s like, what do we do? How do we survive? You know, their CEO talks about staring into the abyss because that’s where they were.
Vin: You know, when everything was locked down, they were—it was over. You know? That was a moment of this could be it for this giant company and they had to decide, now what? It wasn’t, you know, we’re going to go into new markets; it wasn’t anything like that. It was, “Oh, my God. What now?”
And so, data played a role there, and it does in strategy planning. And this is one of the most important pieces that data analytics, and eventually models, it’s the highest value use cases; you bring them into the strategy planning process. Because you have senior leaders, especially now, who are facing uncertainty. So, they need to be more prescriptive, they need to be more forward-looking and understand the implications of their decisions long-term so that they can choose the best opportunities. Senior leaders, especially at the C-level, need to understand the problems that they are facing and what’s causing them. They need to have that diagnostic side of data science and machine learning.
So, we need to be part of the strategy planning process because we are providing decision support tools. And that’s where it begins. When you talk about where does this whole thing begin, you think about the highest value use case for analytics is strategy planning. If strategy is more certain, more forward-looking, if the opportunities that are being identified, if it’s a wider range and they have a deeper connection to likelihood for success for things like growth. And what Airbnb has figured out, that they’ve identified, is their source of growth is really going from these very small apartment rentals that they’re doing, to now families are taking trips. And so, you heard how you kind of prescribed a solution. You said, “Okay, we want to grow by 20%”—
Brian: Well, let me—
Vin: That is a—
Brian: —let me pause you just for a second. Let me pause you for just a second. I don’t think I framed my question well. I wasn’t—and am okay to talk about the historical reality of Airbnb if you think that will make the point. I was using Airbnb just in a total hypothetical situation about, imagine we were the leadership team of Airbnb now, with this current fictitious goal of 20% revenue increase hypothesis is that new markets, we should be in small towns across America or wherever, and that’s how we’ll get to this increase.
How did how do you get from that hypothetical current new goal down to something like a data strategy, or something that a team can begin to build the right thing? Or maybe their job is to validate that that’s actually the right strategy. It’s like, well, maybe let’s start with that. Should you go—
Brian: Into small towns—
Brian: Across America? And then how would you go and do it? Or something like that.
Vin: That’s it.
Brian: Tell me how you get that—I’m just trying to picture, for my audience, how would you break that down? How do we get from that high-level bullet point level down to build something, or provide evidence, or decision support? Like, how do you do that?
Vin: You’ve got to disrupt the process. Strategy planning is not the same anymore. Look at how Amazon does it. Look at how some of these companies that are data-driven—look at Disney. They are destroying their competitors because their strategy planning process is both expert and data model-driven. They have the best of both worlds implemented. They disrupted the way that they make decisions. They completely rewrote that playbook.
And so, when you say, “Where do you start?” That’s it. That is your starting point is to showcase at the C-level, this is what’s changed. And that tangible demonstration, providing decision support systems, changing the way strategy is done. And, you know, that’s why I’m using Airbnb because they are the—they’re just a great—especially front-of-mind right now—example because their CEO was faced with that existential crisis. How do we save the business?
And that’s what makes businesses turn to data. That’s what makes businesses reevaluate the process that you’re explaining. Because the traditional process is, “We want to grow 20%.” And you know, “We think this is the best way to do it.” But no, if you throw data into that process, the C-suite is now saying, “What are our best opportunities?” And data is helping them figure that out.
You know, is 20% the right number? Can we do 50? Because in a lot of cases, the numbers too small. The opportunities are there to go with a more aggressive goal. Using data as a competitive advantage can put the business into an industry leadership position.
And then it’s an opportunity for C-level leadership to say, “Okay, is that what we want to do? Do we want to choose this wave, to become—you know, to put ourselves into that top quartile of growth? Do we want to do that?” And that’s where it begins. So, when you say, “Where does it begin?” You know, how do you get that, from, here’s our goal to here’s what we’re going to use data for, you can hear the collaboration that I’m describing.
And that’s why the technology model is so critical, and that’s why having somebody at the CDO level, who is both a strategic leader—and those two are really critical because your CDO is going to own part of the technical strategy, but they don’t define it. That’s what the business does. The business defines the technical strategy, just like the business defines innovation. The CDO owns it. The CDO owns their part of innovation. They own their part of the technical strategy, data analytics, AI.
And that’s the piece that—that’s where it begins. Now, you’ve got clear responsibilities. You have someone who is no longer trying to define technical strategy and then tell the business what it is. The business defines technical strategy, and it is then owned by the CDO. And the CDO is part of that definition process. They are in the room, they are working to provide these opportunities, talk about how data can even uncover opportunities and help the business fix problems.
Those are the earliest use cases. So, when you say, you know, “What does that look like?” That’s the project. That’s it right there. That’s the product.
You know, the first person you sell is the CEO. Once they see the value of it—and you know, if you want to see the power, listen to Bob Chapek talk about how they use data. Listen to CEOs at some of the largest companies that are successfully weathering these disruptions. Listen to way that Apple talks about their supply chain and resilience, and how certain Tim Cook is that they can manage this amount of supply chain disruption, and here’s where it’s going to start playing—you know, here’s some problems that we’re going to see in the near future; here’s our guidance, and here’s what we’re going to do. You know, we have a plan. We’re not just going to sit here and wait until this wave hits us; we’re going to do things and here’s what we’re going to do.
And then listen to other CEOs who are like, “I don’t know what’s happening. Danger, danger. We have no idea what the guidance is going to be. We have to take it back. We don’t know.” You listen to—you know, it’s a story of two different industries, where you listen to Intel’s CEO talk about uncertainty and how they’re managing what’s happening going forward, and he doesn’t sound like he has a firm grip. He’s got a great story, but you start getting into it, and you’re like, “I don’t know if he knows what he’s doing.”
Then you talk to Nvidia, they’re running into the same struggles. It’s not like they are somehow immune from the problems that Intel has. It’s not like they haven’t faced problems like crypto. You know, that crash impacted their business hard and they didn’t see it coming. That happens. No one’s omnipotent; data isn’t perfect.
But you hear them talking about how they’re launching new product and how they’re making—you know, they’ve survived the supply chain impacts and they’re still launching on time. Their products are still hitting market on time. They got hacked and they had an IP theft. You know, again, they’re not omnipotent, but they have a plan to recover from that. And so, they’re not delivering late because of things they could fix; they’re delivering late because something you really couldn’t have anticipated, even with data, even with models.
They got hit by a black swan. They’re going to learn from this and it’s not going to happen again. And you want to listen to these CEOs because that’s your first initiative. That’s your first project. Because as soon as the CEO hears—as soon as investors hear it, it goes from just being a story about using AI and analytics and being data-driven to it is. Like, this is real. This is execute.
This is something that I can see in every communication that we have. And one of the critical things for CDOs to do is tell stories with data to the board. When they sit in and talk to the board. They need to tell those stories about how one data point hit this one use case and the company made $4 million. The process really is for CDOs to start saying things like, “Here was the challenge. Here was the data we have, the impact we knew we could have. We had some struggles. We had to overcome some problems in the process of delivering. And we did. And this one data point had this outcome.”
And when you start telling the board that story, that’s your use case. You know, it’s so small: we delivered one data point. Who cares? No. That was huge. That’s a massive story because it had an impact. And you begin to hear this, this is an initiative. I am explaining an initiative at a strategic level.
And CDOs get dragged into how did I get the data point? Where did we develop it? What was the technology behind it? No.
Vin: The business just cares that here’s the thing that we did—
Brian: Here’s the outcome.
Vin: The old way. Here’s what we did. And look at this. Look at this. That’s your first use cases. You know, get to the C-level; give them the power. Once they see it, well that checkbook opens up. And that is the absolute power of it, is that’s where strategy gets written. That’s where they begin to understand that there’s a difference. Data is different. You can do things now that you could not do before. There’s the power.
Brian: Vin, it’s been great talking to you. I know you have some training that’s related to the topics we’ve been talking about. Tell me a little bit about that, and where people can get in touch with you.
Vin: If you go to datascience.vin—thank you, whoever created the.vin domain—
Brian: [laugh]. Just for you? [laugh].
Vin: —because that was absolutely perfect. Thank you.
Brian: You selfish [laugh].
Brian: Who do you know at ICANN? [laugh].
Vin: Yeah. It’s as simple as possible. And you head there, you’re going to find all my classes. I’ve got two new certifications; this week, I’m coming out with a third one. Get certified to build data and AI strategies, you can get certified as a value-centric data professional, somebody who’s capable of taking their technology skills, taking their day job into a value-centric direction. And that’s also critical.
And I’ve got one coming out this week, which is your strategic leadership certification. So, teaching that executive level and C-level data leadership role, the organizational level data leadership role. That person that I talked about who’s going to own strategy. And so, I think between those three, I’m trying to meet a need that I’ve seen for—I’ve been doing strategy for seven years—this is the need I’ve seen for seven straight years is we need someone who is a technical strategist, who can live with the C-level strategy planning process, and provide that hybrid skill set. Here’s the technical input that you need, but I’m going to give it to you in strategic terminology to help with the definition part.
CDO plays a role in that as well, but their role is strategic leadership. They’re going to lead strategy, but they also lead the organization strategically. And so, they are making the organization more value-centric. And that’s what I teach in that side of the certification course.
Brian: datascience.vin is the—
Brian: URL? What’s the format of these courses?
Vin: Oh, they’re all self-paced online courses. I give you—and I had to redesign these courses because strategies and intelligent process, there’s no memorization that works. There’s no—you know, I can’t give you a set of facts and now you’re a strategist. I had give systems, models, you know, the process goes away and its frameworks. And so, that’s the very beginning is you go from theory to implementation.
There are tons of use cases, so you understand how this has worked at other companies—like we’ve talked about Airbnb—talk about very similarly with companies like Apple, and Meta, and Walmart, and Nordstroms, and several other different types of companies. And then there’s implementation cases where, here’s how you do it. Here’s what it looks like when you take it from soup to nuts. And all of those combined to give students when they leave, they can do it. They know not only the core concepts and the theory, not just the implementation, but they know how to synthesize it, they know how to adapt it to the decisions I can’t even guess that you’re going to come across.
Because that’s what—you look at what product managers are really good at, they’re good at synthesizing their knowledge to the stuff nobody could have told them was coming and adapting quickly to deliver value. And so, that’s what I try to teach.
Brian: Mm-hm. Cool. Cool. That sounds great. Is that also the URL to get in touch with you? Or is there a better place, social media or LinkedIn? Like, what’s the best place?
Vin: I’m so far all over the place. You can get in touch with me any place. You can definitely—there’s contact on datascience.vin. You can direct message me on LinkedIn, find me on Twitter. I’ve got a YouTube channel. I got a Substack. I feel like I’m running several-ring circus, so it’s pretty easy to get in touch with me.
Brian: All right.
Vin: Any one of those channels, you can reach out and find me.
Brian: Great. Well, it’s been great to talk to you, and thanks for coming on Experiencing Data and sharing your experience.
Vin: Oh, thanks for having me. Thanks for the great questions. This was a great conversation.
Announcer: We hope you enjoyed this episode of Experiencing Data with Brian O’Neill. If you did enjoy it please consider sharing it with the hashtag #experiencingdata. To get future podcast updates or to subscribe to Brian’s mailing list where he shares his insights on designing valuable enterprise data products and applications, visit designingforanalytics.com/podcast.