So this started in the green room of the Texas conference for women.

It sure did.

Yeah, when we met and I had heard of your work. I came across your work through another podcast guest who I'm assuming you all know, Dr. Linda Hill.

Oh, yes. A dear colleague.

Yes, I had heard of your work. And then when I met you in person, the green rooms and all, I was like, wait a minute. Y'all need to be on the podcast.

It was wonderful. It was actually wonderful to meet you there and to even see you in action, Brene. And how you move and touch people. It was extraordinary and I'm so glad we met and so glad we're here today.

Well, I have to say that everyone in our organization is like, can we listen in in real time? Because we all have so many questions. So I'm going to jump in. Y'all have written a book, the digital mindset, what it really takes to thrive in the age of data, algorithms and AI, which makes y'all the most relevant, disliked, desperately needed and feared people in my experience going into organizations. Does that resonate at all?

I think a little bit. Pong, go ahead. Yeah, I was going to say many of those adjectives resonate. Feared is probably the biggest one that I think we hear a lot. And that was a big impetus for why we wrote the book because we actually don't think it's that scary to develop the skills and the mindset you need to really be successful in the digital age. And hopefully that comes across in the things that we wrote and the way we tried to approach these topics. But I can see how it's scary at the outset, for sure. At the same time, it's very clear to managers, executives and individual contributors alike that the world has changed and will continue to change in this digital area. And the question is, how do we equip ourselves? How do we equip our organization? And how do we do it quickly?

OK, I'm going to start with some basic questions. I'm so excited. It's

like really

coaching. OK, define digital transformation for us. I mean, it sounds really scary, I think. And I think everybody's in it and doesn't even know that they're in it. So what is it exactly? How would you define it?

So a digital transformation is changing the organization. How it functions to use data. Technology in the form of algorithms, computing power and reimagining your business models in order to serve your stakeholders very differently. So digital is about data. It's about technology and it's about organization design. And it's a new way of approaching work and working as a result of it. It's a new way of approaching customers and service as a result of it. So it's a wholesale radical transformation of a company's DNA. Yeah, and I just add to that. Yeah. Digital tools are technologies. And as tools, they allow us to do new things in new ways. They give us opportunities to act. And what a digital tool does primarily in our organizations is it provides us access with data. And so at the heart of any real digital transformation initiative is this idea that we need to learn how to cope with data, make sense of data and use data for our strategic advantage and to help our employees and to touch our customers in more profound ways. So if we can think of how tools become an avenue for us to get access to make use of all these new forms of data, then we're on the right track to thinking about what digital transformation is.

OK, I want to stop for a second. They always joke. They call this the pause cast because not only am I, I think, are in a pauser, I have no need to fill up a lot of space with bullshit. So let me think about this for a second. I have to tell you, this is the first time I'm hearing some of what you're saying. And I spend 90% of my time in organizations freaking out going through digital transformation. So it's a definition that's really different. So I want to think about it for a second. I don't know that people always think about digital transformation as data. The collection of data, the analysis of data, the implementation of new strategy and processes based on data. I'm not sure that that's how people in general think about digital transformation. It does that resonate with y'all when I say that's not been my experience?

I think you're very right. But the reason why we're in this digital space today is actually, if you can imagine three circles, one is access to data in ways that we've never had before. We're talking about million data points or what's called metadata, which is data about data. So the presence of data in extraordinary ways, if you think about another circle, is computing power. OK, computing power has allowed us to crunch data to have so much data being processed in ways that was not possible 20 years ago.

Can I raise my hand and ask a question real quick? Sure. Are these like Vin diagram, like Olympic rings? Are you making a target?

Think of them as Olympic rings.

OK, got it. OK, three of them. OK.

So data.

Computing power.

OK, extraordinary. I mean, at this point, we're even talking about quantum computing at Harvard. And the third one is models, algorithms, statistics. So the presence of these three forces is why in that center, we're seeing digital transformation. And an example of this, an example that I think people can relate to is Netflix. So how does Netflix figure out the types of things that they should recommend to you? It's all algorithms, right? It's collecting historical data, being able to predict what are the things that an individual or a family or group consumes and doing some matching and being able to make recommendations, recommendation engines. That's all because of data, computing and algorithms. That's what digital transformation is.

OK, I've drawn it. And if you're listening right now, you know, that I love the listeners, I'm going to make sure we give that to you on bernaybrown.com under the podcast that a graphic will be there so you can look at it. I'm clear on access to data and metadata. I'm clear on computing power. Paul, tell me about the model circle.

Yeah, so models really are kind of strings of how would I put this in kind of an easy to describe in an audio way? Just tell us the hard way and tell us the easy way. OK, well, you can think of sort of strings of equations that put together lots of inputs, permute that data in some way. And then there's another set of equations that will output that data in some format that you want to read. So what a model basically does is says, OK, we've got all these data points. We need to do something with those data in order to make sense out of them. And then we need to calculate those data in a way that provides lots of actionable suggestions for people in our organizations. And so models, I would say, are sort of commodities in many ways these days. Like if you're trying to run some kind of machine learning algorithm, which we hear a lot about, Microsoft and Google and Amazon, AWS all have these models that you could download and use and play around with. What's really important is understanding what are the kinds of data that we need to have? What do those data represent? How are they going to speak to particular business issues that we care about? And then have the wherewithal of what to do with some of the predictions that those models are making based on those data. And those two ends of that spectrum, the knowing what data we need and then knowing what to do with the predictions, I think is really at the heart of what digital transformation is.

Interesting.

That's what we really try to get at in the book is that all the stuff that's happening in the middle, you should know how that operates and you need to get to a certain level of fluency about that to just be a good consumer. But the smarts, the real thinking that we have to do are on those two ends.

So under my model circle and my Venn diagram, is that where algorithms belong? Absolutely. OK, great. So access to data, the computing power and then the model or the algorithm we use, we collected the information, we crunch it. And then what model do we pick to make sure what we're crunching and how we're crunching it leads to actionable things we can do to make our businesses better? Is that yes, a clunky, but fair?

No, you got it. The model is really trying to turn data into some kind of insight. And one of the important things to remember is that the insight that that model generates is usually based on some kind of statistical model. And so really what it's doing is it's making a prediction. It's saying given this that we know today, this is what we predict is going to happen tomorrow based on the data sources that you fed into me as the model. And this is an issue that we see a lot in many organizations is that people have a really difficult time understanding prediction and that prediction is not the future written out for us, right? But there are predictions about things that might happen if we made certain choices. And so the real sort of management and leadership role is to figure out what are the choices we should make. And in order to do that, you need to understand a little bit about how those statistics operate for sure so you can have confidence about whether they're pointing us in the right direction. And perhaps more importantly, we need to understand what those data look like because if we don't have the wrong data or we classify those data in ways that don't really make sense, then those models are ultimately useless for us.

And dangerous, right? Yeah,

potentially. I mean, lots of stories about how bias creeps into these models. And we discussed a fair amount about that in the book as well.

You do. Yeah. It's really scary to me sometimes when I go into organizations and I love to do focus groups with kind of the least powerful people and organizations first. And then I hear, oh, no, there can't be bias in hiring because we use an algorithm. And as a researcher, you're like, wait, you know, shit in, shit out. Exactly. I don't understand what's happening here. Who developed the algorithm? Exactly.

And what data are you putting in? Can I tell a fun quick little story about that? You know, I love a story. It's even more basic than that. So I was doing some work a number of years ago at this large research lab. And they had implemented this new computerized system to try to track all the work that technicians were doing across the lab and who was sort of the best at doing different technical tasks. And the manager who had recently got his MBA and was running this part of the group decided that the data that were in this tool would just be the perfect data to determine everybody's performance evaluations for the year. Because they showed who was doing what jobs, how quickly were they completing them, how good quality were those jobs. And so he was using this to try to make sense out of who should get promoted and who got raises and so on and so forth. And there was one woman at the organization who had worked there for about 25 years and she was just consistently getting really low evaluations. And she was so frustrated that she decided to quit. And that manager was so happy because he's like, well, obviously, she was the lowest performing person. And over the next two quarters, the customer service ratings, right, just went down and down and down. And what he didn't realize was that he was capturing one kind of data in that tool and it was data about solving technical problems. But he wasn't capturing all other kinds of data that were really important for keeping the organization working. So this woman knew, for example, what solutions users were likely to like. She knew who maybe we should prioritize across the organization to keep the business running the way that we wanted it to. So it was all this sort of social and cultural knowledge that was not captured and stored as data in the tool. It was overlooked and it led to really poor decision making. Wow.

See, I think such a cautionary tale because I do also think operationalizing those contributions is really hard. It is. And so I think when you start building models, you're like, uh, this seems kind of fuzzy and soft around the edges. This is not binary. I'm going to leave this out. And then what an important tale. Yeah.

And I think it points to one of the dangers. And so I see this happening in so many companies that when things are easily quantifiable, they take on a permanence or they take on a seeming objectivity. And the things that aren't easily quantifiable, like knowing what jobs we should be doing first or knowing what jobs we should be doing second, don't seem to take on as much authority as data that are presented in numbers. And I think that's where a lot of leaders really go wrong.

It's really funny. I had when I was in my PhD program, she was actually my MSW program is where I met her, Karen Stout. She since passed away, but she studied femicide, the killing of women by intimate partners. And there was a lot of pushback for me in my PhD program because I had the first qualitative dissertation in my program and, oh, my people were pissed. Really? Oh, yes. To the point where Barney Glaser who developed grounded theory methodology. Yeah. He was my methodologist on my dissertation committee. You're like, I need some legitimacy here. I need some legitimacy, but they were so hostile about the dissertation proposal that Barney legally removed himself from my dissertation committee until the proposal process was over, then added himself back because they required a lit review in it. And he's like, if you already know what literature to review in grounded theory, then this is a waste of time. And so as it turned out, I reviewed all of the, what you would expect and none of it was irrelevant. But there was

a going

with the story with Barney.

The quality of the legitimacy of quality of the cognitive versus deductive.