Live discussion: Building Engaging Internal Data Communities

This is where I would normally say, hey, welcome back to the Data Science Hangout. But what I am welcoming you to is a special panel discussion with three industry leaders where we are talking about building engaging internal data communities.

The Data Science Hangout has been going on for five years. Rachel started it five years ago, believe it or not. I have the great privilege to host it. Rachel and I have learned a ton about building communities inside and outside of organizations over these years.

And we are joined by three industry leaders in data science, so specifically data science and data analytics, who have a variety of experiences with building community, with seeing how community can affect things like innovation and culture growth, adoption of new tools and technologies, all kinds of stuff.

So we are here together to talk about not only building community, but also how community is not one thing. It's going to look different for everybody. We have multiple industries and fields of work covered here today.

So what I will do is I will go ahead and start with asking our guests to introduce themselves and tell a little bit about what they do and what community kind of looks like at their organization.

And of course, we have Rachel and I here today, so I would love to introduce myself too. My name is Libby Herron. I do community stuff here at Posit. I run the Data Science Hangout. I also run the Data Science Lab. Those are two places where our data science community comes together every week on Zoom to talk about data science stuff, technical topics, career topics, industry topics. They're beautiful spaces, and I also run our Discord server.

So Rachel will share a link in the chat to join the Discord if you would like to. The invite will be open during our call today. And that's a place where if you are not already there, you can see data science community happening.

Also, Posit.com registration is open in person in Houston, Texas, virtually online with myself and my co-pilot, Isabelle Velasquez, if you know her. And we would love to have you there, speaking of, like, you know, wonderful data science community.

Rachel, would you like to introduce yourself? Sure. I was just going over to type, I just got my flight to Houston for Posit.com, so seeing who else is joining.

But I am the head of customer marketing here at Posit, so aside from all these great community events that we host, I also work on our customer spotlights and love getting to share all the amazing work that y'all are doing. I'm located in the Boston area, so if you're ever in town, let me know.

Guest introductions

I am so excited to introduce Chris Englehart. He's the vice president and head of AI services at GenRe. Chris, would you like to introduce yourself? Tell us a little bit about your role and about what community has looked like for you.

So as Libby mentioned, I'm Chris Englehart. I'm the head of AI services at GenRe. I'm located a little east of Cleveland, Ohio, in a town called Shaker Heights. GenRe, for those unfamiliar, is a reinsurance company where we provide insurance for other primary insurers. I'm also a social and personality psychologist by training, so I have a background in academia and I very much bring that background and training and education to my work every single day.

And so I'm responsible for a variety of different areas, including bringing people along on our AI journey. That's kind of my chief remit for the team, where we work with an eye toward helping people build trust around AI tools and technologies and fluency with those tools. We also build and deploy custom AI applications and manage an AI portfolio projects. But more pertinent to this conversation, we also enable enterprise community building and AI adoption.

I've been with GenRe about four years now, and I was brought on initially as a data science engineering manager. One core component of that role was to lead an internal R meetup, which we called Regular. It's a really clever portmanteau, obviously, because it speaks to both what the focus was in R and the frequency with which we have it, which is bimonthly. And we help educate people on best practices as they relate to R and Posit products. We also invited other data scientists, actuaries, and developers to that meetup and community to speak on projects or ideas that they were working on.

We were obviously very intentional about the culture that we had in that community, which I think will be maybe one common thread that you hear about today. So it was an R users meetup for everybody, a welcoming space where we had discussion, discovery, new users, and education. All of those topics were welcome and invited. And we framed the community as your community. So we weren't necessarily the owners of it, but the ownership was broadly distributed across all members who decided to join, that we wanted to hear from everybody, and that they could reach out to any of the organizers if they wanted to be part of it.

I would also love to introduce Dudie Roy, Executive Director, Head of Innovation, Clinical Data Sciences and Methodology at Boehringer Ingelheim. And Dudie, if I messed up that very long title, you let me know which of it I did.

I do, but I think I'm here primarily today in my capacity of Head of Innovation, Clinical Data Science and Methods in Boehringer Ingelheim, for those who may not know, is a German pharmaceutical company. And of course, we have a global presence, meaning that there are roughly 60,000 people who work for this company, and particularly stats and programming, the team that I'm talking about. We sit in US, we sit in Europe, we sit in China, and also in Japan. So that's a little bit of scope of the community that I'm looking at.

Maybe just going back a little bit, my background, trained as a statistician, and joined the company as an individual contributor, really building and designing clinical trials, right, in different disease areas. And after a while, there was a mandate for me to start looking into the transition from SAS to open source R, as some of you may be familiar in pharmaceutical industry, SAS has always been the language of choice for very many years. And at some point, it hit that ceiling to say, okay, we really want to do things that at the moment is not possible, or easily possible using SAS, so what other languages are there?

And in the meanwhile, the stat methods community has been churning out innovative approaches and methods in R for almost 15 years. So it was kind of like a jump to say, okay, can we really start leveraging R company wide? And of course, it's easier said than done. I had that mandate, I was like, where do I even start?

And that was also roughly the same time I came in touch with Rachel, right, and then one of the conversations was really around community and the value of community, right. So when we are thinking about bringing something new to our people, oftentimes, we kind of also must think about community. But I think one thing that I was able to bring differently in the organization was to really start thinking from community from day one. So we set up like a dedicated community focus as we were talking about this transition and how this would look, right. So this was one of the many pillars of that.

And just a few months back, beginning of this year, I had another rule change and now my mandate is to lead innovation overall in clinical data science and methods, meaning SAS, methods, programming, etc. So now community to me has a different but a broader meaning, right. So now I'm looking at community building in several different domains of expertise. You can think about a statistical community, community focused around programming, data science, some method focused communities.

So it has obviously broadened, but the principles that I learned and used and saw value from the same principles are usable. So I'm very excited to be here today and also like talk a little bit more into details, share some of my stories and learn more about, you know, from all of you.

And we also have James Laird-Smith, Senior Data Scientist at Bank of England. James, I would love it if you would likewise introduce yourself and tell us a little bit about what community looks like at Bank of England.

So, yeah, like Libby said, I'm a data scientist at the Bank of England. It's always worth noting or just worth saying whenever I introduce that to the American audience that we are not actually a bank. The Bank of England is the UK's central bank. So it's much like the Federal Reserve for the people who come from the United States, that we have a lot of responsibilities, but holding customer deposits is not one of them. So we don't act like a normal bank, but the bank has lots of responsibilities. So we are the financial regulator. So we regulate the financial system, which is other banks, insurance companies. We also control monetary policy. So we're responsible for, among other things, like setting the bank rate.

Unlike maybe the others on the, well, I have never had to like build community from nothing. So the bank has a very big analytics community. So there are many hundreds of people who work for the bank, who data is their job. And so they interact with data in a variety of ways. So we have many, many hundreds of our users, many hundreds of Python users. And so I definitely do participate in community building, but I've never had to build a community from scratch. I was sort of, in some sense, gifted this community of many, many analysts.

I happen to have a semi-formal role in the organization where I am the R business owner, which means that I take on sort of some of the leadership for the direction of our tooling. So we have many hundreds of our users, and I am sort of nominally in charge of the direction of that sort of tooling.

Being a big organization is we also have sort of people whose full-time job it is to administer those tools. So we have full-time members of staff who are responsible for our Python installation, for our R installation, for Posit Connect, and things like that. I am not one of those people. So I am a sort of like a citizen sort of data scientist, but I do have some elevated responsibilities.

Much like the others on the call, we also have sort of less formal avenues of community. So we likewise have semi-frequent user forums where we divide it up among the tools. So we have like an R user forum, we have a Python user forum, and several other user forums for different tools.

I will say that just because we have a big community doesn't mean it's always active. So people have busy jobs, and everybody will sort of recognize and sort of like the participation that we get, for example, at the user forums, but also elsewhere, does vary considerably depending on how busy the organization is or how busy the particular sections of the organization are at any given time.

It's also the case that although we do have a very good user base, one has to be honest and like to say that the top sort of data and analysis tool at the bank is still Excel. So we have many hundreds of users of the data science programming languages, but there are still people who use Excel and Power BI and sort of similar tools like that. So there is still room to grow, as it were, and that we still look for avenues to get people on board with other data science tools sort of beyond Excel, which most people of the organization are probably at least a bit familiar.

Scale of user communities

I think it would be great if I could really quickly just go around and ask for a little bit of scale. I know, James, you said hundreds of users of certain things and maybe a large percentage of people who use data working in Excel. Do you have an idea of a number of users for like R or Python?

Yeah, so we do have some telemetry on this. So we do collect data from sort of user logins. And so I think right now R is sitting at about 700 users over the course of a year. It's a bit lower than previously had it. These things do sort of vary. Python at about 400 or 500. We don't count Excel because we don't just count like the entire organization, I suppose. Everyone's got an Excel license, right?

Chris and Doody, same question. If you have that number, I know that James happens to sort of like know that already, but I don't know if either of you do.

So there are R users, for example, beyond my own organization, there are also adjoining organizations who are using R. I would say also roughly around 600 to 700 folks globally, where I would see people regularly use R for their daily business.

On the GenRe side, I would say roughly 275. And there's predominantly R users compared with Python users.

These are big numbers. Good reminder that communities can be two people. Communities can be three or four people. They can start small. And this goes back to what James was talking about, right? Just because you have these many number of users doesn't mean that all these people are super active in the community, right? We would love to, right? That's definitely the goal, but this is not oftentimes a reality. You do have some power users in every community who are really, really active, keep it going. And then you have some people who are kind of engaging more than the other ones. And then you have some silent people. This is generally how I think most communities look like.

That's the same for us, Dudi, too. So we have roughly 250 who would be users, but that doesn't necessarily mean they join every community event. We have roughly 70 to 90 who join the bi-monthly meetups that we have. So we have a pretty good turnout based on the total population of R and Python users.

I think that we see a similar thing with external community events, right? You might get 100 people that sign up for something and really only 50% or less might attend, kind of depending on how broad the signup group is, right? And I think that the makeup of the total population of possible people makes a big difference. Where are they? Are they used to talking with each other already? Do they already know each other? Do they already have mechanisms for fluidly communicating with each other? Because if they don't, I think that that can be a pretty big barrier.

When everybody registered for this event, we asked everyone, do you have a community already? And there were some people who said, no, I don't have a community. And then someone else from their same organization said, yes, I do have a community, which drives home the point that you might have a community at your organization and not realize it. Or you may have people who are getting together regularly and talking about things, which is a community, but they just don't call themselves one.

What it takes to build a community

I wanted to ask you what you feel like it takes resource-wise to actually build a community at Boehringer.

And of course, you know, when I was sitting with the task in front of me, I was quite new, like I had no idea how to make this really work. And I got some really good advice from several people, including Rachel here.

So like I mentioned earlier during my introduction, right, so we kind of had this focus that we would like to put community building as one of our central pillars of this transition. And that meant basically twofold things, having one person dedicatedly assigned to that task, looking into responsible for starting the community forums, thinking about what topics are critical for the community. But also, like Libby, you mentioned that Boehringer is present in different places, right? And it is very difficult for one individual, no matter how dedicated they are, to somehow reach and connect with every different type of audience spread over different time zones. So we also took this approach of regional leads.

So it was basically like four people, right? One of them was the global lead. And then there were other people talking together, planning together, but essentially executing in their own region. That's what like organizationally it initially looked like.

But we would like to kind of draw a difference between what it takes to start a community and what it takes to keep it going and really build that up. And I think that really takes a village. Three people working is not going to make your community behave and act like a community. And that really comes from your rest of the community people who are not the leads, connecting, taking an interest, bringing topics up, solving each other's problems. And that took almost a year and a half to really start seeing a difference, right? From having these couple of people who were obviously taking the lead because they were assigned that task versus, oh, this person has nothing to do with the lead role, but they're just interested to contribute. And that takes quite a while.

Three people working is not going to make your community behave and act like a community. And that really comes from your rest of the community people who are not the leads, connecting, taking an interest, bringing topics up, solving each other's problems.

So my question for you would be, what does it look like for you as far as a model of distribution? Is it more centralized? Is it like here are these people in this central organization who are in charge of community or is it more distributed where there's sort of a responsibility enmeshed in a lot of different places in the organization?

It is for us mostly centralized. So for us, we have a central team and sort of people who are appointed organization wide. There is some, well, there is a big informal network, but it is just that. And there is some attempt to recognize sort of so-called champions for some of the tools, but with sort of mixed success.

I should say that there's not many like formal requirements for my job. The one is that we have an annual tool upgrade cycle. So we upgrade our tools sort of once a year and there's a strategy for doing that. And the business owner, that is their one formal job, which they sort of, if they sign up, they sort of have to do that and take responsibility for. Everything else is basically bonus.

And in these circumstances, it really does, at least for us, matter a great deal if that person is good and if that person is very invested and if that person puts in a great deal of effort. My particular way of doing that is through shared tooling. So I'm less good at hosting user forums, but I'm much better at developing shared tooling for the organization. And that is my contribution to getting people on board.

My experience is that the best way to get people into the community is to somehow solve their problems. Now solving problems can mean a lot of things. And it's just one of those things where you have to try and keep your ear to the ground for what people are, with what people are struggling, and then build sort of something that you hope will solve that problem and document it really well and broadcast it quite widely. And I find that the momentum that you can get from that does pay off.

So we certainly have people who are involved in our community who wouldn't be if not for a particular package that either I built that they use or someone else built and they use and kind of has made their job much easier. And they sort of, not just from a self-interested perspective that their job is now easier, but they want to engage with the people who made that possible, but also be exposed to other things which either are available right now or could be made available in the future, which would further benefit them or their department.

I just want to support that statement so much that James made, because ultimately, you know, we are having a job. We have things that we need to get done. And if you cannot establish to your people that there is a benefit, tangible one at that, right, to do or join something, I think just relying on, oh, yeah, it's good to good to have or, yeah, it's nice to hang out, this will just not be sustainable. So just wanted to give a cheer to that statement that if you really want to create communities that sustain themselves and keep it going, better know what the needs are and then try to somehow address one or some of those needs.

Culture, psychological safety, and change management

And I know that I've talked to James about this, too, that there's a very collegiate culture at Bank of England where everyone's really aware that everybody else is busy and has things to do and that not everyone will drop everything right now and fix something or that sometimes people are less busy and that they can do that and everybody respects each other's time.

If the culture is not conducive to a community existing, I think that can be a tough battle to sort of work upstream against.

And so there's a conversation in the chat as well around individually talking to people and how Rachel established that as a culture very, very early on at the Hangout, the Data Science Hangout, and how important that was.

I saw the Data Science Hangout randomly on LinkedIn. I think I must have followed Rachel five or six years ago because I was a user of Posit Products at that time, RStudio products, and I had been for years, saw that something was happening, decided to show up because I'm the type of person who's interested in that. And then immediately after, I got a little message from Rachel on LinkedIn that said, hey, thank you so much for coming. What did you think? Do you have any feedback? Well, I always have feedback because I'm that type of person.

So I gave her a bunch of feedback and I attended the next one and she had taken my feedback and immediately blended that into what she was saying, how she was saying things, which made me immediately trust her. I think that that one-on-one interaction sealed the deal for me. I was in, I was bought in, and I then started building the community and helping Rachel do that external to her so that she didn't have to do that.

You are building a community with every single interaction that you have.

All righty. I would love to talk to Chris for a minute, too, because this is all about psychological safety for me. And Chris, with your background in psych education, I would love to hear what drew you to community building as this mechanism for successful change management, because I think that's how a lot of people benefit from it, is we see change in an organization through community as a sort of conduit.

Yeah, yeah, Libby, I think part of the appeal for me is the challenge that change management presents, right? Because it's the biggest problem. And how do we attempt to change people's behavior? And in general, we're just not very good at this. We spend millions of dollars trying to change people's voting behavior. And I think we consider it a success if we can barely move the needle on that.

And I think there's at least two ways to go about trying to get people to change. There's clearly more. But one ineffective way that I've experienced, and perhaps many of you have as well, is to simply tell people that they need to change or take some authoritative stance on what people's behavior is going to look like or should look like. Again, I've never found that to be quite so effective. And that's backed up by a host of literature, because in that scenario, you can have people cling more tightly to their original beliefs because they have a reactionary effect to trying to instantiate this change that may be unwanted or unwarranted.

In my experience, however, I think another more effective way to bring about change is to use community building, you know, more to the topic at hand here as a vehicle for change, as people are more likely to change if, you know, they see a peer showing something that has worked for them and now they're vouching it or vouching for it in front of others. I think that gives people kind of a greater sense of autonomy or control, you know, over that process. And that way they can more easily build a mental model that if they can see a peer doing it, that they themselves can do it as well.

I know you touched on psychological safety as well, Libby, and that is kind of a key piece of fabric that, you know, we instill into the communities that we're building here at GenRe. And the idea of psychological safety goes back to at least the 90s, where the general idea is that the community perception is that the community is safe for interpersonal risk taking. Where, for example, it's OK to admit mistakes, it's OK to say that you don't know, it's OK to present a half-baked idea, it's OK to be you. And it's OK to bring like everything that you are, your lived experiences to the table because those are all worthy of value.

And, you know, that is a deliberate, intentional kind of mindset that we attempt to instill in the community because we know that that's predictive of things like team learning and team performance. And if we can have those types of environments within the community, you know, I think it's a great way to motivate change where maybe people might not be so willing to explore those things initially. And, you know, having that safe space for people to come and share ideas and be themselves and, you know, be free to admit fault and share learnings. And all of that, I think, is really setting the stage for instantiating change that you might want to see in your organization.

The idea of psychological safety goes back to at least the 90s, where the general idea is that the community perception is that the community is safe for interpersonal risk taking. Where, for example, it's OK to admit mistakes, it's OK to say that you don't know, it's OK to present a half-baked idea, it's OK to be you.

I fully agree. I think that building a space where, again, people see how you respond to each other, and people see how other people respond to each other, just watching interactions is super powerful, especially if you have leader-level, staff-level, higher-level people in your organization who are willing to be vulnerable in those shared forums, like on Teams or on Slack. They're willing to ask questions, or they're really, really nice when they answer questions. That can go such a long way in making other people go, oh, you know what? I have a question, and I've always seen people be nice to each other in that space. Then they're much more willing to ask. But if you have an organization where people get shut down a lot, where it kind of feels like Stack Overflow, and everyone's being rude to each other, then you're going to naturally, I think, see communities start to dampen because the psychological safety is ebbing away.

So likewise, with the one-on-one thank yous for showing up and asking for feedback, there could be one-on-one conversations with people who are not responding very well, and you could let them know, hey, it would be so great if you could respond differently or keep in mind that these people are new or think back to when you first started. I have had some conversations that are tough like that with people throughout different communities, and they have actually really helped, even though they're tough conversations.

Training, tool adoption, and keeping up with changing technologies

Well, I wanted to share a question that I saw in the Slido. We don't have a super, super dedicated way to ask Q&A questions today, though we got many, and we'd love to talk about some of this stuff. But one of the questions was around how members of different communities keep up with changing technologies, and are they organized programs to help people? This is a great topic because this is actually about training and development. And tool adoption. So I wanted to ask all of you if you have structured programs, training programs that incorporate community in any way. James, you have a really, really long established user base of technology. Do you do any standardized training or regular training at this point?

We do, yes. So we have various levels of training. So we have an R for beginners training, we have Python for beginners training, and an intermediate training too. People have asked me whether we should have an advanced R training. And I have to be honest that I don't know what that is. It's just because even after a beginner level, the skill set diverges quite a lot. I even struggle to describe what it means to be an intermediate R user. So that could mean you're really good at Shiny, for example. Or that could mean that you're really good at functional programming. And I think we have a reasonably good hold of what it means to be a beginner, and what skills you need to get started.

We do try and make sure all the links are there. So we do try, if you go to R for beginners, we do try and make sure that you're then on our internet data community. So just to make sure that we cover our bases. We do try and make sure you're on the, we don't have Slack, but the Microsoft Teams channels, just to try and be disciplined in getting people on the ladder.

We also do the things which one can take for granted, but maybe shouldn't. We ask for feedback for all of our training. So this is not me, but our training team are very good at taking on feedback and acting on that feedback. So yeah, just a continuous sort of growth mindset, I think, underpins quite a lot of what we do.

I think that's so important. It's sort of a training as community building. And then the community building helps the training because people who are new together can help each other together. And people who have more experience can also be in those same groups, helping those people up the ladder. Like you said, this is something I talk with Rachel about frequently is people need a reason to be there. Training is often a great reason. And for some people, just a conversation, just showing up to be with people is enough, but often it is not. And that is understandable, right? That's not going to be everybody's motivation for being there. So learning something, keeping up with a new tool, learning a new skill is often the motivation that gets people in the room together.

And I wanted to talk a little bit more about the beginner thing, because this is something I've talked with Judy about before, is being sensitive and naming things. How important is it to name things in a friendly way, especially putting beginner anywhere in there?

Yes. And I think Chris brought this up, right? With this regular meeting, right? And I also saw the same inclination to kind of incite that feeling of, like, it's a regular thing, like, come and engage. Within Boehringer, we also had this discussion around naming. And of course, we were well aware. So, you know, in the pharma industry, it almost feels like this is a generational situation, because there is this 20, 30 years of experience in this industry where people have been very comfortable, used to using a certain language to deliver their work. And now here comes these new languages that are kind of, like, more familiar on the newer colleagues who have joined, right? And they are more familiar with this language. They are fluent. But then there is this generational gap that we see oftentimes. So, it was very clear to us that we need to be careful about how we name things.

And we basically have something called Begin Our Office Hours, which basically also is trying to get to, yeah, it's a beginner. So, please come. And that was definitely helpful to have that build up their psychological safety, at least an effort to say, it's okay. No question is stupid. Just come in and have a conversation.

But I also wanted to, Libby, maybe add a couple of points from my side on the question that you originally asked, right? Which was essentially, what's this learning component or training component and community, right? Because in my new role, as I also mentioned, that my scope has kind of broadened, right? So, we have situations like in the past, for example, if you're looking into a particular methodology or statistical topic, we would have these, like, massive trainings where you have two full days or three full days where people did nothing rather than just listen to someone speak. And that's the way the training was delivered, right? But then if you reflected, maybe, like, 5% of those people were actually using what they learned. The remaining 95% spent the time. But when it really, like, six months later, maybe they come across the problem, they have by that time forgotten what they had learned. So, that wasn't really creating that value that we wanted, right?

So, we had kind of, for technical topics, this is and that includes, like, for example, learning programming languages or learning to do something specific tool with a programming language or learning statistics, for example. We have separated this out and said, well, our concept nowadays is more towards self-learning. So, you have online resources that you have available. And then you also have dedicated hours with the trainer, for example, but not that the trainer is monologuing. The dedicated hours is there just so that you are already reviewed the content and you are there to clarify, you're there to ask questions, you're already building that deeper sense of knowledge associated with the topic. So, this is something we just recently started experimenting with, right?

I'll be curious to report back in a year or so how this is. So, that's one thing that we have this dedicated question and answer hour with the trainer, but then we also have this community of practices where the intent is there's no trainer available, but then you have, like, each other to ask questions and exchange experiences and exchange knowledge and what worked, what didn't.

So, that's a little bit on the technical front. And, of course, there is this other part around human training, right? Human skill training, which means non-technical, right? So, we also look into training, for example, in statistical leadership. And there, the model is that you come engage in face-to-face trainings where, again, the focus is not monologuing. You're mostly learning by acting and improvising on the spot and then keeping that topic alive via community of practices where maybe this topic kind of gets revisited every quarter. So, people are engaging on how does this statistical leadership look like for you this month and what were the challenges that you faced, what tools did you apply, did it work, did it not, and kind of, like, encourage learning via community. So, for us, I think now training and learning and community are even more closely coupled.

Measuring success and leadership buy-in

Anytime you can get human beings to talk to each other, you are building community, because then they are more likely to talk to each other again, even if they're not in the same context. They will learn things about each other. And not every group of people is going to re-engage, right? You might introduce 50 people to each other and only two of them end up getting together later and chatting about something. That is still a win.

And there was a question in the Slido about, like, well, what does success look like? How do you measure this? And if you've gotten buy-in from leadership, for example, to take the time to do this, what does success look like? I think that training can also help with that. It's, like, number of people who got through training can be a number that is success. But I will also share something that Regis James said from Regeneron when he joined us at the Hangout, which was years ago. His metric was a little bit of unstuckness, right? Like, if you can help people get unstuck, and I know that's a hard thing to measure, although Regis measures everything much better than I think I do.

That is sort of a nebulous thing. But getting people engaged with each other, even if it's only a couple of people here and there, can really help people get unstuck. Because then they sit with their problems by themselves less frequently. They're more willing to reach out and ask for help to a group or to that one person who felt safe. It's not really as unmeasurable as you think, because we have a very similar approach these days.

I can share a little bit briefly. We have this group of people who are leading this transition, for example, and they look at governance processes, tooling, upskilling, community building, all of these as a whole thing. And one of the ways we keep track is we encourage our community of users to submit tickets when you have a problem. It could be about anything, and you just tag the right group, right? And that gets tracked. So if over time you can show that after a training or after some event or after something has been rolled out, the number of tickets corresponding to that topic went down, there's your metric, right? So I think this is also somehow reflective of how effective your work that you have been doing have been.

And I think from my senior leadership, this always has been an ask, right? Number of people who took the training cannot be a metric, because that was the only way to the means. What have you done with the training is a question, and then you get into some murky waters, because then you have to really hold people accountable. And I just want to highlight there's a difference between if there is leadership buy-in versus not. Because if there is leadership buy-in, then you can also have that accountability part pretty clear, right? Because the ask is directly coming, as opposed to, well, nobody really cares if this is working or not. So I think there is also a big difference in that.

Yeah, it can be hard to measure things that have to do with people. And you might have to rely on some proxy metrics, metrics that are good enough. Or you might have to rely on qualitative metrics, which are, well, we talked to all of these people who are managers, and they are seeing better things. They are seeing more collaboration, they're seeing less people being stuck, things like that.

Expanding community to AI adoption

I like that you mentioned that you were making a change. And I know that, Chris, you have sort of moved into a new challenge. You started with, let's do R adoption. And that's where that begin R and regular comes in, right? But now you are shifting your focus to AI adoption. And I think a lot of people at a lot of different organizations are dealing with this as well, which means that your scope is now much more broad. You have users who are not exclusively R or Python or SAS people. You have people from across an entire organization. And I'm wondering what shifts or challenges you are expecting to have to make in order to make sure that you can still have a community for those people that are very broad and diverse.

Right, and I think you've hit on exactly the challenges, right? So, you know, the data science communities that we had, the regular communities, by the way, Judy, like the beginner labeling as well, I think that's really clever. So, we have, you know, that, and that's been going on for several years. Again, that ranges between 70 and 90 folks, depending on who turns up at one of those meetings. And the remit now is around kind of enabling AI adoption at an enterprise-wide level. And so, to me, you know, there's two sources of challenges there.

The first is that, you know, we'll have multiple skill sets joining there in terms of their background. So, we'll have, for example, actuaries. We'll have, you know, people from HR, legal representation from our various business units spanning property and casualty and life and health. So, we have a variety of different backgrounds who will be participating in this community. The second challenge that I see there is that we'll have a wide variety of skill sets. So, we'll have people who maybe have tried a copilot once, hated the output, never looked at it again. And then, on the other end of the spectrum, we have people that are developing AI-native applications and deploying that in the cloud for external clients. So, we have kind of a wide variety of skill sets there. And how do we engage across that spectrum effectively?

So, I do view that as a significant challenge. But, you know, that's kind of the gaps that we're facing right now. I would be curious to hear how others, if they've, you know, taken on this challenge before, because I'm just going into this for the first time. But I think we have to kind of strike the right balance, you know, between those two challenges, right? Like, how do we accommodate kind of the variability and backgrounds? And then, how do we also accommodate folks who may have different, you know, parts of the journey that they're in? You know, how do we meet people where they're at effectively? And so, I think we'll do that through some combination of general topics, plus, you know, maybe some offshoot communities that would be more area-focused. So, it's not necessarily just one community, but talking about community broadly, where, you know, there could be maybe more niche, you know, kind of areas that kind of spin off from the primary network community that may be more enterprise-wide.

Practical advice for getting started

Well, I would love to use our last 10 minutes here to chat about practical advice, knowing that everybody comes from different levels of experience and different sizes of organizations. What could you do right now if you wanted to start, you know, organizing something or making community happen?

My first piece of advice would be actually to ask around and see if people are already meeting and you just don't know about it. See if people are already chatting somewhere. And to James's point, it could be from training. So, I would look out for training places, right? Like, is their training already going on? And are the people who are taking the training already talking to each other? That's probably the beginning of a community.

But the most important thing is to put something on the calendar. And Rachel and I put together a whole blog post about this with a checklist at the bottom that'll help you get started. And you'll see that number one is put something on the calendar, weekly or monthly. It doesn't really matter as long as you get started. If you sit and, oh, is this the right thing to do? When should I do it? When can everybody get there? You'll just drive yourself kind of in circles and it will never ever happen because there is no perfect time, right? Like, I can't imagine that at an organization like the Bank of England or at Boehringer Ingelheim that there's one single time for everybody to get together.

Never, right? Like, that would never happen. But if your time exists, then that is a time that people can show up. So, put something on the calendar in a recurring slot and start inviting people. And really, it doesn't have to start with 100 people. It could start with one person, two people. Just invite people. Talk to them about what you want to do.

Reach out to your colleagues of diverse roles and levels of experience and let them know that you would like to start doing this. And you will find people, the more you reach out, that are interested in it who just don't have the time or the inclination to do the setup stuff. It is a lot of work. And so, someone has to take the responsibility to do it. I think that self-sustaining communities sound really good, but for a lot of organizations that are smaller, that is not a reality. It's not self-sustaining. Someone has to put in the work. And if that person leaves, the community can fall apart. So, recruiting more people and getting them on board is important.

I would also suggest finding some way to offer anonymous engagement. We use Slido. There's a free tier of Slido. I do believe that Posit pays for a paid tier of Slido, but there is a free tier. So, you can provide some way for people to ask questions anonymously. That works well if your group is maybe 10 or more. It doesn't work so well if your group is teeny tiny because then it's pretty apparent who's asking what question.

And then, recognize people one-on-one. That's the one-on-one part. Thank them for coming. Ask them for feedback. You'll find that a lot of people have feedback but don't actually want to take any action on it. So, take all feedback with a grain of salt, right? Do what is possible for you to take action on. If someone says, I'd really love to participate, but I just don't know anybody, maybe you need to start introducing people one-on-one. Send them some one-on-one messages or emails and connect them to each other. Start lubricating people socially in those ways.

Rachel likes to say, think about it like you're walking into, they're walking into your house for a party and they don't know anybody. You could introduce them to someone, get them a drink, get them a snack, show them where the snack table is. For me, with community building online, that is telling them how they can participate, what they should expect. Somebody mentioned in the chat, oh, well Libby does this thing at the beginning of each thing where she tells us how we can ask questions and she lets people know that this is a safe space. Rachel started that at the beginning of the Hangout and that's really, really valuable. If you can take two or three minutes at the top of the hour to meet, to say, here's why we're here, here's what we're going to be doing generally, and here's how you can participate and what is expected of everybody, that can be really, really a relief for people.

Yeah, I think one thing I wanted to add to that part, I put it in the chat, but I think it's so important to remember that every time, it might be someone's first time joining the community event. So even if you've been running it for five years, like the Hangouts, we have new people joining all the time. So I always like at the start of an event, I try to remind myself like, okay, someone might be here for the first time, what might they need to know to feel comfortable too?

I think that if you're anything like me or if anyone is anything like me and they don't know somebody before going to a party, they just won't show up. If the only person I know at a party is the host and I know the host is going to be busy, I'm probably not going to go. But if I had a buddy to go with me and I knew that they knew other people, I'd be so much more willing to show up. So just introducing people one-on-one can go so far in helping other people show up because once they do show up and then they see the same names and faces pretty regularly, things get easier. I've seen that at the Hangout as well.

And then Rachel, there's also the follow-up aspect, not just thanking people one-on-one, but also if you have that Teams space or that Slack space, pulling out some little tidbits of things that we learned that were fun. That goes back to what James and Judy and Chris have all said about making sure that people know why the joining is beneficial, why it's important to be there. Create a little bit of FOMO. Why they should have gone and what they missed by being there.

Create a little bit of FOMO. Why they should have gone and what they missed by being there.

Just to add on top, I also think that most organizations at the leadership, they also recognize the value of community. So what is also another on-top advice would be to talk to a leader. Ask them what they think about community building. Do they already know some activities that are going on? So this speaks to that lack of visibility some people have that I don't really know what's going on in the other corner, but maybe a leader does. So it's also good to kind of start having that conversation with your leaders that brings them closer to you. They can see that you are a proactive individual. They can see you're excited to do this. That can also open doors later on. So it's also worthwhile kind of connecting that dot.

Yes. And there was a question that Paul had asked in the Slido, which I wish we had a lot more time, but I will leave it up to anybody who wants to answer it. It says, do you get engagement in your communities from the IT side as well? What a great question. So this is the facilitator side, admin side, not just the end user side. I wonder, I feel like James, you are more connected to both of those things as far as having an administrator view of things.

Do you have any experience with that? Yes. So again, at the bank, there are people whose full-time job it is from the sysadmin type side. We don't get a lot of engagement from technology beyond those people, but it's good to have them. And it's also a very frequent or very active line of communications as I, as the business owner, talk to what we call the technology owner, maybe not daily, but certainly every week. And so it's really good to have very frequent communication between the stuff that I know very well and the stuff that that person knows really well.

Yeah, very, very advantageous if you're lucky enough to have somebody like that. Absolutely. Make friends with people at IT, make friends with your sysadmins, give them cookies, send them presents. They're doing so much work behind the scenes that you will never, never know. Invite them to speak at your events, ask them how their job works so that your end users can build empathy around what it takes to