Welcome to our newest season of HumAIn podcast in 2021. HumAIn is your first look at the startups and industry titans that are leading and disrupting ML and AI, data science, developer tools, and technical education. I am your host, David Yakobovitch, and this is HumAIn. If you like this episode, remember to subscribe and leave a review. Now on to our show.
Welcome back, listeners, to today’s episode of HumAIn.
Here we have our guest speaker, Victor Bilgen, who is the partner and head of McChrystal Analytics. I’m really excited to speak with Victor today because both of us work in the data field, and what everyone knows is, it’s more than data science and machine learning, if you don’t have analytics right, you have to start somewhere.
So, Victor, looking forward to this dynamic conversation with you today. Thanks for joining us on the show.
Thank you, David. And thanks for having me on the show, really looking forward to the discussion today.
I know this year and last year has been all about, can we use machine learning to predict everything? but it doesn’t always work that way.
Yes, you’re absolutely correct. Being a partner at the McChrystal Group, which is a management consulting firm and a leadership advisory firm, our job is to be able to use our data, to be able to help leaders understand how to run their companies, how to be better leaders and, ultimately, how to prepare for a constantly changing environment.
So, a lot of times we’re brought in to answer some data questions. When, really, you have to be able to take in the entire context of what’s going on, get some qualitative information and also have the expertise to know what to do with that information. So, it’s always been a tricky, but very interesting environment we work in.
It’s so fascinating because my work, which has been both with startups and scale-ups a lot of times, I come in as an advisor and investor for startups and they say, David, we’re bringing you in.
And I say, okay, I’m so excited. Why are you bringing me in? We want to know the secret. Tell us about the AI, tell us about the data. This is why we’re bringing you in. And it’s going to help us raise our valuation.
We’re going to grow, scale the business. We are the next unicorn. And I say, not so fast. Let me hear about your business. Let me hear about what you’re doing, what you’re building, how are you going to scale your operations? No, we don’t even have an API yet. And I said, okay, you got some work to do.
Let’s start from the ground up. And so today, my role as well, I’m with SingleStore, we’re a massive startup scaling in a distributed SQL analytics space. A lot of the space that you play in and customers also come up to us. How can we speed up machine learning and AI?
And that’s part of it. But companies are often rushing into data maturity, too quickly. So when we’re thinking about 2021, new budgets, thinking about strategy analytics, everyone’s thinking about what they should invest in? What should they build?
What have you started seeing and what are some recommendations you might have for the industry?
That’s a great question. I’ve been doing what I do now for the better part of a decade. Prior to coming into the management consulting space, I spent a good amount of time in market research and getting my graduate degree. And I can tell you that the methods to understanding data can start to look consistent across those different avenues.
But as we got into the leadership space and we got into understanding management and organizational structures and functional team dynamics and geographical expansion, there became a need to really start to hone how you understand yourself using some pretty interesting new methods.
I’d say the thing that first popped up for us many years ago, but has really caught on some steam here is the organizational network analysis, which allows us to take data, whether you actively collect it through surveys or you’re passively collecting it through collaboration information, emails, chat, calendar. And you can actually understand how people are connecting.
So we work within organizations. We can map out where people are actually connecting with each other and what that means as far as how you run your business, because every organization is going to have their structure, their strategy, and they’re going to design things very cleanly.
But the way that organisms and organizations operate, usually doesn’t look like that. So the network analysis piece has been one of the critical components for us to be able to hone our understanding of connectivity, information flow, decision-making. And it’s really one of our primary lenses for people to understand themselves.
For the company that I’m involved with today, SingleStore, we are scaling massively. And organizational network analysis is relevant for all companies, whether you’re a startup, a scaleup, a Fortune 500.
For ourselves, as an example, we are a startup that’s becoming a scale-up, we’re a few hundred people. But our culture is so different among the different organizational elements.
We have the classic Silicon Valley engineer, world-class engineering, amazing talent. And then, our business team and sales processes and the cultures couldn’t be more apart, but more unified in their workflow, especially in a remote culture.
I imagine that not every company is like SingleStore. I imagine there are companies where unification takes a lot of effort. So thinking of organizational network analysis, what’s the lens that’s relevant that you think companies should be looking towards?
That’s the question that I probably get asked the most. When we are delivering our results to clients, we tend to, first and foremost, get asked the question, what does right look like?
And that, as I get to the answer of what you’re asking, makes me think about the fact that most of the time we’re dealing with our organizations and our partners.
The way that you structure your organization is you’ve got the thinkers and designers that are establishing how you run your business, your operations, what your strategies are, and you have the core components in the back office to understand your people.
Generally speaking, that falls within IT and HR. Because those two elements tend to be places where you’ve got siloed thinking, I know where we want to go. I’ve got the financial understanding of where we’re going to take our company. And the other side is, here’s what your people are saying about their jobs. Those two areas never meet. And that’s why we exist today, because a lot of times that means of those strategies, one is failing.
So as you ask what are the different ways, you can start to understand yourself the idea of what is right look like for organizational network analysis is always going to be a moving target.
The key is that you tie your understanding of how your network is connecting to your ultimate goals. Start to marry those two sides of a business. So you brought up a good example of your organization. I like to think about, for example, an organization that we were working with that traditionally was driven by geographic markets.
They had leaders at every single market. So one in New York, one in San Francisco, one in Los Angeles, Atlanta, and they worked great as that sort of geographical silo. And so, when we ran our network analysis, we saw exactly how you’d expect it to look. Clusters of people that said, I get my information and guidance from my leader in my geography.
The problem was that their market was shifting. It was shifting to something where their customers are no longer just geographically centric. They were dealing with global customers, ones that were larger and spanning across multiple geographies. So when we came in and ran our analysis, we said, this is how you look, you look exactly like how you designed yourself, but this is not going to work for your new problem to help them understand what the network should look like as it starts to connect more geographically.
And I can tell you that even today, they’re looking a lot more how they want it to end up operating, but they’re probably going to have a next moving target as they continue to shift their strategies.
And everyone’s shifted, as we know, in 2020. They went online only, not to say by choice. It’s because of the virus, but so it’s really like who is right and who’s wrong. There’s really no wrong. In 2020, everyone was figuring this out together.
And I know that organizations are moving back into the hybrid model. Now some work from office, maybe all work from office. I imagine it’s been a fundamental shift in perspective and productivity, and then partnerships, by seeing how companies went remote only.
What did you see throughout the entire pandemic? Were there theories that you knew about organizational network analysis that were proven true, proven false or any displacements there that you found fascinating, Victor?
It’s definitely been the dominant top of mind sort of thinking there, the first place that I would start would be the McChrystal Group itself.
The mentality that we put out, especially with our clients and our partners and thought leadership that we have, it all stems off of the founder of McChrystal Group, General Stanley McChrystal, and his line of thinking around how he changed the Joint Special Operations Command under his leadership to be able to be more connective, agile and fast acting.
So the way that he did that was really to establish behaviors and processes that allowed leaders to connect flattened communication, and be able to share information that is relevant at a faster clip, so that people work with autonomy as they’re all over the world. And so we took that and we applied it to the types of analytics that we run.
We try to understand where companies are at with that. So when the coronavirus pandemic first hit, we were presented with a massive global challenge. All of our clients, all of a sudden realized that the standard way of operating, the standard way of having people in the office was going to change.
We didn’t know how long that uncertainty would become a very large point of conversation and topics. I’ll tell you that the original point of view on working in the office was that as a leader, you have more perspective on what your people are doing. You have an expectation that when people are in the office, they’re going to collaborate better together. And you have a situation where you can maximize your team’s innovation by having them in collaborative spaces.
The fear of moving to a remote environment was that you’re going to lose that and productivity would fall. What we’ve found is that, instead, the individuals are a lot more nuanced and diverse in their experience. There are some people who benefit from working from home that can be heads down and focused.
And there’s others that probably struggle a lot. Because if you remove the serendipitous interactions that individuals might have in an office in order to make up for that, people will start adding more formal meetings to try to keep those connections. So what does that mean for us from an analytical perspective?
What we ended up finding was that it was really important for leaders to understand how collaboration has changed in these new environments. And it was very important for people to understand the climate and organizational health of their employees. So what we ended up doing was, we worked a lot with Microsoft Workplace Analytics.
And we developed our own set of tools to help organizations understand the changes to their environment. So what that looked like was we could, actually, using again, things nor-organizational network analysis, employee engagement surveys. We could get the leaders a very specific lens to understand how their teams have been impacted, whether now there are new silos, because teams are not connecting as well. Whether there are groups of employees that are more stressed by the uncertain and remote environment, and whether there are individuals that are critical nodes within your network, that were necessarily going to be impacted by the lack of seeing people in face-to-face situations.
So, David, I’d say that using data and analytics became a primary concern for leaders that knew that they had to impact their teams. They had to understand something they can no longer see with their own eyes.
Data is everything. And when we’re thinking about leaders and we’re thinking about budgets in 2021 and scaling organizations, the missing piece for organizations has been everyone’s remote only, and people are being hired remote, people are scaling remote and you miss the context.
You miss the ability to understand every moving piece of your data.
Typically, when we think of products, there are trace logs. We’re understanding every movement, every action for software and the programming language. But does that always translate to our people as well? And what I’m hearing in our conversations so far is that may not be the case, it’s possible for that to be the case.
And there are things that leaders can do. So that becomes true. So, what’s your take on strategy to influence leaders, as we’re building organizations for the future?
The idea that leaders need to understand, the data that’s out on hand, as we talked about, is a critical conversation to have. I can tell you that I’ve experienced this both from a practitioner, as well as an executive partner perspective.
And in my time, there are really important parts of the conversation that need to happen. But the first piece that is usually skipped, but shouldn’t, is if you talk to a leader and say, this data is going to show you X, they’ll likely say, sure, that is valuable to me, but how do you convince them that they actually need it?
And so there is a change management aspect to these discussions, and for any data practitioner that is trying to influence leaders and thinking about leveraging data like this, you have to be able to make them understand how there’s value for them. They need to understand the problem at hand, and they need to understand that there is a time-sensitive need for change.
So what I found more specifically, as we looked at this environment, you talked a little bit about product tracing. There’s obviously a concern for some, on things like survey fatigue, and there’s a concern for how we’re going to get this critical information when our environment is changing so often.
And so, the way we’ve been thinking about this is we have a lower tendency to send out surveys. We have a survey that we use for context. Because qualitative context is key to all of this. We’ve found that it’s far less effective to say here’s a bunch of charts. These charts are higher, low.
We need to be able to put these things into the leader’s vernacular. What do they care about? What are they trying to achieve? But then, being able to track change over time is an age old question, but is one that I believe there are answers out there now.
As mentioned, we have a partnership with Microsoft Workplace analytics.As an example, their tool set is what I believe is going to be one of those things that really shapes the market moving forward and changes things.
Their tool essentially allows for a structured query of Microsoft Office 365 data. Things like email, chat, calendar, and you get massive data dumps because even in a hundred person company like McChrystal Group, you’re emailing, chatting, and meeting all day, every day.
So there’s always these edges between people that show how they connect. And using our qualitative approach upfront to be able to contextualize why people are connecting and, what their roles are and how people are feeling. And then overlaying that on this passive data creates a really rich story on how people are connecting, how your business is operating and some core KPIs that I just believe people are not tracking right now that they should.
As in examples, how many one-on-one meetings are you having with your employees? How many collaboration hours do you have between sales and marketing? How many out-of-normal work hours meetings is your R&D team having in order to accomplish their objectives? These are critical KPIs that can proactively predict the success or failure of a lot of those strategies.
As long as they’re contextualized with an understanding, a management system framework, you can look at this data and understand which numbers actually matter. And so, as we look forward, those types of collaboration metrics are going to be the ones that really help shape how people are leading, especially as we’re in this hybrid work environment.
You have to contextualize your metrics. The metrics have to be unique for your organization, and depending on how much data that is, that could be for 10 people, 100 people, 100,0000 people.
There’s so many elements of the data. And viewing that as analytics and dashboards, whether in Microsoft Analytics or integrating into other tools where you use technology, like the SingleStore and the snowflakes of the world, you realize that data at scale also takes a lot of computing.
It takes a lot of processing power to be real time. And as an executive, that’s one of the big things. Time is money, and everything you wanted in real time. And so change management is making these decisions with as much data as you can have available, because if your data is incomplete, then you’re making an incomplete decision.
And of course, you never have 100%, but the more you have and the more insights and the analytics you have, you’re steering the ship as good as Tom Hanks steered the Greyhound across the Atlantic.
So, we’re steering, the world, we have vaccines distributing. We have organizations ready to move into new hybrid modalities of in-person online models. And you have certain people, certain very interesting people in a way organizations where they may not be the executives. But they may be a special person that’s connected to a lot of other people.
We learned about this in business school and data scientists learned about this in social network analysis, that’s there’s that specific individual or group that somehow connects so many people together. What do you call them and how are they important for organizations?
I love that part of this discussion because social network analysis really is the foundation to organizational network analysis. The main differentiator, when you’re thinking about it from a data science perspective is, with organizational network analysis, you also have to have an integration and understanding of organizational design and structure.
But the idea is very similar. You’re trying to understand pathways. Whether it’s between entities like teams or between people. You want to understand what they’re connecting on and how often and where, and that allows you to compare the design that’s based off your strategy and the way that people are actually operating.
So within our experience, we have leveraged techniques that were first started and really grew into the larger organizational network analysis fields by people like Rob Cross, who’s currently at Babson, Dr. Nick Christakis at Yale. There’s a few other experienced experts that we’ve talked to be able to grow this practice. And I can tell you that what we’ve found is, the first thing that’s really important is understanding those key network influencers.
And we have about four or five that we tend to look for. The first is our central network influencers. Central network influencers are the ones that people are probably pretty aware of. They do tend to go more senior. They are geographically centered to the network map. They are named a lot of times or interacted with many times with high frequency.
They tend to have important cross-functional roles. Generally speaking, those central influencers are the ones that leaders should be able to understand who they are and make sure they are on the same page when it comes to where you’re trying to take your network.
Actually, my Chief Of Staff came up with that term. As we found them as this unique entity within networks, they’re usually not named that much. They’re usually not that high impact to the rest of the network. But they are information sources to your central network. That’s the general councils, for example, individuals who the CEO might go to, but no one else would name.
The third is what we call the regional hubs. Those are individuals that act like central influencers, but are disconnected from the central network. So generally speaking, those are individuals who might be very important to a geography or a specific team that’s not connected. And understanding those individuals is critical to making sure that you’re pulling your network together, that you’re getting the high impact individuals on the same page.
And finally, there’s brokers. Brokers might end up being individuals who are five or six layers below the senior leadership, but they uniquely connect otherwise disconnected geographies on otherwise disconnected teams. And by identifying those individuals, you can start to understand how you can bring together and bridge groups and teams by understanding, is it because of their functional role? Is it because of their hierarchy, their tenure? Do they rotate between teams or are they just really good at connecting with people?
So when you understand the key network influencers in your network, it gives you an ability to create change at scale, because you can’t make 10,000 people change in a day. And especially if it’s within an organization. But you can change the top 10% of network influencers, which can then have a network effect, a scalable effect of change for your leadership in your organization.
With how we’ve been virtual only, it must be fascinating to see how influencers have been at work in the organizations.
As you mentioned, very central, very focused about in-office and with stakeholders. What I’ve heard a lot, at least in the education perspective, and then, building product perspective, is being online-only has actually flattened communication. It’s made it so that anyone can communicate. You can talk to the CEO by just picking up a Microsoft Teams meeting, a Slack meeting, where before you may have not been able to get that time.
And so, a lot of influencers are stacking many meetings and sessions in a day that normally would take a lot longer. So perhaps there’s been some productivity improvements in the organization, but that’s not everything. There’s so much to do.
What are some of the ways that you found the fact that you have to help leverage these different types of influencers?
The first step is, obviously, identifying them. And that’s where the organizational network analysis comes in. But once you’ve identified those key individuals within your networks, we’ve found that there’s a few really important ways to have them impact your network the way you want them to, the first you’ve mentioned, flatten communications.
There are a couple ways of thinking about this, but one of the ways that we actually found was really important is most organizations have regular meetings.
When we moved into this remote and hybrid environment, those regular meetings increased by a large amount. Again, because people felt as though they needed to introduce meetings as a way to fix the lack of serendipitous communication and collaboration. They created a formal structure to fix and a non-formal event that they believe went away.
You can make that more efficient by putting those influencers into critical meetings. You can start to impact the number of meetings they have to be part of. Create a formal structure where you do get that communication vertically, including your network influencers. And all of a sudden you are creating scalable communication.
The next thing that we have seen as really valuable is the idea of cohort building. So we’ve had situations like the previous example that I provided, when people were trying to create cross-geography connectivity what they did was, they found network influencers at the middle of their company from different geographies and had them come together to solve problems.
Like an integrated planning team or like, again, being able to come up with agendas for that big meeting or something like that. Being able to have cohorts of people that are all network influencers allows you to start bridging those cross-functional divides that exist.
And then, finally, one other that we’ve seen before would be rotational programs. Now, those rotational programs actually come into two different forms. There’s the standard rotational program, which people are aware of, either it’s an entry-level and you’re going to pop around into different positions, or it’s an executive level, and they need you to understand the different teams you’re running. So they’ll have you work with those teams.
We also have another type of program we call a liaison network where there’s probably a better term for this, but it’s like a prisoner exchange. You take your best person on your team and the best person from the other team. And they switch jobs.
In that situation, again, what you’re doing is creating an environment where you can drive awareness and connectivity between teams that otherwise wouldn’t be connected, and you can have your network influencers have a deeper understanding of the other side.
So in general, leveraging your network influencers is something that you can do systematically to reduce the burden that this hybrid environment has likely already put on them.
I actually love this analysis. You just said a prisoner exchange. It reminds me of the Netflix show 60 days in. They bring the non-prisoner humans into the prisoners and see, how do you change actually? How do you change organizational network analysis? Do complete strangers impact the dynamics, the hierarchies of different teams and their interaction.
Of course, it’s a Netflix drama. So it wouldn’t be without all the interesting things that we enjoy about the viral shows. It brings up a fair point that the cross-culture, the cross-technology is critical to pollinate that, both virtually and in person.
And the teams that we’ve seen be very successful in the last couple of years have done that, whether they’re remote or in person, because they’ve always been willing to build a culture that’s thinking about constant improvement, being agile, being open to change, being willing to understand that they don’t know everything, that anyone can bring up a good solution, a good feature, a good request
At our company SingleStore, we recently ran our moonshot project. In fact, our Chief Product Officers from Google. And at Google, they always focus on moonshot projects. What’s the next big project that’s going to get you your next million users, your next billion dollars? And part of it is, anyone can have the ideas. It doesn’t matter whether you’re the chief scientist or a manager who’s going through a rotational program.
It’s so fascinating to see that everyone can be leveraged. And in fact, anyone can be an influencer. When we take everything that we’ve been discussing here today about strategy and change management, and we bring it back to data, it’s important to see, moving into 2021, with organizations that can do the responsible use of the passive data it’s being collected, it’s being stored in databases and data warehouses and tables, but that’s not always being mined.
What can we start doing? How can we start responsibly using that data for our organization?
That’s the number one question that we get into at the front end of these data collection discussions. Because for the data that we’re using as an example, we are talking about employee data. We’re talking about employee perceptions. We’re talking about employee behavior.
One of the most critical aspects that we first need to make sure people understand is our intent for that data. So what we tend to do is prior to doing this work, we make sure people see examples of how this data is used. Because otherwise, you’re going to have a massive trust issue.
We’ve had conversations in the past. I was talking to the Chief Data Officer for a large healthcare organization, and he had mentioned that he tried to turn on some sort of passive data collection tool in the past. And he saw a 20% reduction in work email during the time that he turned on this tool. That is a surprising number.
And we asked about how he went about doing this. And essentially, there was an email that went out saying, we’re going to turn this on. Now, don’t worry. Trust us. This was essentially the point of the email.
And that obviously did not meet people’s standards for the context needed and why you would need this information to know how you’re going to use it. Who’s going to get access to it. All of the critical information necessary.
I live in Los Angeles, so obviously, California has some pretty strong data privacy and protection. We do work with countries in Europe and GDPR is a massive set of regulatory procedures that we also have to adhere to.
So we’re very aware of that. There’s the sort of protection end of it that we need to do from a regulatory environment, but we also fully respect and understand the point of the respondent. I don’t like taking surveys unless I know that there’s a value. And so, the key thing for us, first and foremost, is to say here’s what this looks like.
My job is to make your life better. That’s what I’m trying to do. And with people seeing how this is actually being used, that can be the first step for leaders to responsibly use this information. Outside of that, the organizations that have built tools like you’re talking about, like Workplace Analytics, and other companies that collect passive data or companies like ours, are very rigorous in making sure that that data is protected and is very difficult to access unless you’re a practitioner.
Those two things, being smarter than regulations and making sure people understand what you’re doing with the data are two key critical areas.
A lot of it is about security. And you mentioned GDPR, CCPA. We’re looking at European and California standards.
We have a Chief Security Officer at SingleStore, and we’re talking every day. How do we get more secure with our product? How do we get compliant with all these standards and working towards them every single day is essential and it takes everyone in the organization.
On the data science side, I’m known to have my data science standards where I talk a lot about design thinking. And particularly one of the exercises I run through with students often is from Carnegie Mellon University. They have a data science and AI ethical standards. If you’re building any data, AI processes, it’s basically a checklist to see, am I involving every human stakeholder possible? Am I asking questions about, am I doing no harm? Am I improving the product?
We’re looking at all these things and it’s so important, because collecting data for the sake of collecting data, that’s good. But if you don’t have all the right security processes, then we start seeing we can just turn to the news and see what we’ve seen over time.
And fortunately, there’s a lot of new security tools in the market, and there’s a lot of new encryptions coming out. So I’m quite hopeful that 2021 is going to be a year where security is going to take back control. We’re going to be leading there first in data systems. We’ll be thinking when we build data systems, how to build secure systems from the start. And as we’re continuing to move there, security is everything, it’s on the top of my mind every day. It sounds like it’s on the top of your mind, as well.
A hundred percent. The state, local and world governments are getting smarter on this. And putting the protection of people’s information and privacy above all is a really important step.
Companies that want to be successful in this new data environment need to not only accept that as the reality, but embrace it, because with a system where people are more trusting of the data environment, there can be a lot more value created for the people involved in it.
Well, I’m looking forward to learning more. Victor, I’ve had a lovely conversation here today.
I know for all our listeners, you can find out more at mcchrystalgroup.com. You can also follow Victor on the interwebs. Check out LinkedIn, learn more about his publications and discussions.
This has been a conversation about how analytics around people can inform your strategy, not only today, but into 2021 and beyond.
Victor Bilgen, Partner and Head of McChrystal Analytics.
Thank you, David.
Thank you for listening to this episode of the HumAIn podcast.
Did the episode measure up to your thoughts on ML and AI, data science, developer tools and technical education?
Share your thoughts with me at humainpodcast.com/contact. Remember to share this episode with a friend, subscribe and leave a review. And listen for more episodes of HumAIn.