Building a vibrant community with Scott Hirleman - AgileData #45

Welcome to the Agile Data Podcast, where we talk about the merging of agile and data ways of working in a simply magical way. Welcome to the Agile Data Podcast. I'm Shane Gibson.

And I'm Scott Hurleman.

Hey, Scott, thanks for coming on the show today. Just before we start, can you hear that? Can you hear that? No, you can't hear anything. Do you know what that is? That's 1,000 Slack communities where nobody's posting. And one of the reasons I've got you on the show today is you have run one of the most successful slash communities I've ever been a part of. We've talked off and on, and you keep saying to me, there's no rocket science, there's nothing magical, there's just a bunch of simple things to run a community that is as successful as the Data Mesh Learning Community that you've been running. And I don't understand those patterns. So today, I've got you on so you can explain to us in the data world how we can build communities with our stakeholders, with the people we work with, and potentially with external audiences. But before we do that, why don't we just start off with you giving us a few minutes of your background and how you got into this wonderful thing of building communities?

Yeah, we'll get into specifically data mesh and what I was doing there. When I went to school, I was really obsessed with kind of the stock market. I was watching, I was that kid that would come home from school and watch CNBC to see like what had happened with it. This was back in the heydays of the late '90s and early 2000s. I still remember where I was when the NASDAQ peaked. In, in 2000. I was on a skiing trip with my brother and stuff, the only one I've ever been on in my life. It was just very random. I went to school and focused on that. I was a stock market or equity research analyst covering tech and things like that. I was in Milwaukee, Wisconsin, which is not the hub of tech, so I decided to move to San Francisco. They let me stay on for a while while I was doing that, looked around, got hired on to do some venture capital stuff, and I was supposed to be focusing on semiconductors, which was what I had focused on as an equity research analyst. And then the company was like, you know what, about 2 or 3 months after hiring me, they said, you're going to focus on semiconductors. And then I got in there and they're like, you know what, semiconductor investment right now, because it was like 2010, it's way too capital intensive. It's ridiculous. So find a different thing to focus on. So I started to focus on big data because this was the big wave of Hadoop and all that stuff. So, started to dig into big data, met with something like 150 companies, and we made only 1, 1.5 investments. There was one that was in the BI space, but we made an investment in DataStax, who's the company that's the main push behind the Apache Cassandra project. Was working there for a bit and realized, hey, I wanna go over to DataStax. I think it's gonna be a rocket ship if we could do the right things. So, I went over and they hired me originally to do sales operations and financial planning and analysis, or FP&A. So I was doing a whole bunch of things. Originally it was just supposed to be sales operations, but then I did sales enablement, FP&A, I started to do some community marketing stuff. And then they brought in a new CFO and he wanted to bring his whole team in wholesale. And I was doing the financial planning and analysis stuff. So they said, find another role internally. I'd been talking to people on Twitter, trying to get understanding, especially for helping salespeople to understand what's actually going out there, like what's happening out there. I'd been talking to them. And so the community team saw my ability to go out and find people that were using Cassandra, have an engaging conversation where it was an actual conversation instead of a sales conversation. And so they had me come on there. I did that for a couple of years, found this really amazing way of sourcing leads. So I spun out to go do that and source leads for a whole bunch of companies, but it was based on a LinkedIn thing and they shut it off the day before I was supposed to start. So I bet this whole thing, and we can go into that at some point, but yeah, that, that was crazy. But I've just always been good at finding people and then finding what they care about and just talking what they care about. Did some sales and marketing consulting, had somebody recruited me to move up to the Seattle area to do tech recruiting. So I did that for a bit, but it was at an agency and you're really just doing the same thing over and over. You didn't get to talk with the hiring managers, so you didn't really get good feedback even on the resumes you were sending over to people. Why did they reject them? Why did they interview them? All that. Didn't like that. Decided to just bet on myself, quit and moved down to the Bay Area, did a little bit of business operations for a company, but it was a little bit of not the right fit. And then I had an old contact from the Cassandra community reach out and say, hey, we're struggling with our AWS costs. Can you come in and manage those? So I managed those. The company was just launching their SaaS, so their costs were skyrocketing. They were up 30 or 40% month over month for 4 straight months. And then about 2 months in, they stabilized. And I think over the next 18 months, they only increased 25%. 5%, despite our number of users increasing 400%. Pretty good things, but I was focused on that AWS stuff, but that really got me back into the really deep communication with the tech people because cost management isn't about keeping costs as low as possible. It's about being as effective with costs as possible. Had an old acquaintance again from DataStax recruit me into a startup. It really wasn't the right fit. They were just trying to figure out what they were doing, but I was doing business operations stuff and then I just decided to I gave them notice in late January of 2020, and they asked me to stay on into April of 2020 to help them close out their quarter. And I was like, yeah. And then it gets to be like mid to late March of 2020. And they're like, do you still want to leave? And I'm like, yeah, you know what? I do. So I spent a couple of months just faffing around, I think is the word that you use down in the ANZ area. And then people from DataStax reached back out, asked me to come and help with Cassandra community stuff. Now I was focused on that for about 8 or 9 months, but really my issue is when you ask me to go and intersect with community people, once I get going, I can bring in 30 or 40 people in a month, not just like to be in the community, that want to interact, that want to do specific thing, that want to create content, that want to be involved. And they just didn't have the capacity to take that much people. So then we crafted a role for me where I was supposed to jump into a bunch of these different communities, and that was the friendly research spider, was my title, and to jump into these communities and make these connections and bring them back and make connections to people inside DataStax. Again, what I was finding was I would find all of these people and then we just didn't have the best capacity to actually deal with those people. I had to wrap them up in a bow. I had to go, okay, here is this person, here is exactly how you should interact with them. I had to do all of this stuff instead of letting somebody create a relationship and explore what, like, what should be doing, what should be happening. But the first community that I started to, or the second community that I started to help was data mesh. I had reached out to Zhamak. I actually asked internally somebody and said, Do we have any information on this data mesh thing this person is looking for? It'd be nice to help them. And the person misread my comment and said, yeah, learn as much as you can about data mesh. I'm ADHD, ASD, so I went and I learned as much as I could about data mesh. And at that point, there were only like 80 articles total, but I started talking with Zhamak and everybody was trying to ask her, how do we do data mesh? How do we do data mesh? And she was the only outlet. So I was like, hey, should we create a community around this? Should we create just a Slack or whatever, where people can go and talk to each other? And being me, I did this big prep work and went, and the community went from 0 people to 500 in 4 days, 0 to 1,000 in 14. And when I was focused on growing it, I think it got up to 4,000 in about 6.5, 7 months. And that was only in the Slack, but then there was the LinkedIn, which had 2,000, 3,000 followers. The meetup was, when meetup still had that thing that you could look at, it was considered the fastest growing meetup group in the entire world around the word data, but there was a ton of interest, right? It's not as if nobody was interested in data mesh and I generated all this stuff. I tapped into something there, but I managed that community for a little over 2 years before handing it over to an independent company that kind of runs it like a foundation. And so they're doing some stuff there, but along the way, I started to focus on different things. So at first I was focusing on bringing people into top of funnel, and then I was focused on trying to generate these one-on-one conversations, and then What I was trying to do that I failed at, and we can talk about this as well, like the pattern as to how I think it would be successful, is trying to bring other voices up and elevate those voices and keep the momentum behind them instead of a flash in the pan versus a somebody that's actually gonna take on a lot of load and do that type of thing. And then I just saw everybody was still writing, what is data mesh articles back 10, 12 months into where the community was, cuz we'd launched in early February. And so this was in December of 2021, where I was just like, there's just nothing for practitioners. So then I launched a podcast around practitioners, like actually talking to people, how do we do this? What do we have to learn? So you and I had an episode, and you're not really super involved with data mesh, but you and I had an episode about how do we do agile and do it appropriately in data? I handed over the community. To an independent group to run the actual community itself. And then I'm now focused on consulting with, as an advisor to people that are implementing data mesh and connecting them to each other, because I find that's the most helpful. You can come and ask me all the questions in the world, but I want you to talk to 4 or 5 different implementers.

So, let's unpick some of them. Where do I start? So, let's talk about this idea of bringing on 30 to 40 people that interact versus bringing on, thousands of what I would call observers, that whole thing of, okay, we're gonna start a community. Oh, we've got massive reach. Oh, look, we've got 3,000 people in 3 days. And then a month later, there's 3 of us left talking. 2 months later, none of us are talking. I don't know if you call them zombie viewers, the people that come because it's new, it's shiny, they don't engage, and then they get the hell out of dodge. So you talked about this idea of bringing 30 to 40 people that were gonna interact. Is that one of the core patterns, is actually being able to identify people that are gonna engage, interact, drive the conversations, have something interesting to say, have something interesting to ask. Is that one of the core patterns to getting engagement within that community?

And specifically what I was talking about there was more on the company side of, hey, I've got an influencer. This person is already talking rather than for the community. I think the community, you can go out there and you can say, who's trying to influence and not having a ton of luck? A couple of people, I won't name names, but one of them I think is actually really brilliant, but he's only getting 10, 12 likes on a lot of his LinkedIn posts. But I think they're insightful and I think they're deep and I think he's very smart. And so he would be the type of person that if I were doing a community, I would go out and say, I, I want you to come into the community and not just invite them, be like, send them a message and be like, look, I want you to get involved in my community and I want to do something with you. Immediately hook them on something. We tried this early in the Data Mesh Learning Community of doing what I was calling topic swarms, which when you think about swarming in Agile, you think about swarms in, in software engineering, you basically are like, we've got an issue, we're gonna have a bunch of people come together, we're gonna have everybody kind of swarm in and work on this one issue. So we tried that and it didn't work 'cause I think we were too early. So I wanna get back to answering your question in a second, but there is a difference between between observers and shiny chasers. There are people that are too embarrassed to speak, and this comes up a lot in data. It's especially prevalent, I've found, from people in Europe, because people in the US just have a little bit more ego, hubris, and things like that. I think when it comes to Australia and New Zealand, it's a little bit 50 of one, 50 of the other, where you've got people that just don't care as much about the appearance of their reputation. But in Europe, a lot of people are like, "I'm not an expert, so I shouldn't be part of the conversation." We tried to do these things and only 6, 7 people were showing up and they were fun, but it was only like 6 or 7 people were showing up of a community of 2,000 people. It's really difficult to find 30 or 40 people that are going to constantly interact. You can find a couple of people that are raring to really react. One thing that I do is searching LinkedIn and stuff, I just do search for in quotes because of Boolean search of quotes data space mesh or capital O-R, and then data mesh as one word, because then you get the hashtag data meshes and stuff like that. And so you can just watch for the patterns of who is talking, right? You can sort by latest and you can scroll through. It's a lot easier when there's not a topic that's incredibly overloaded. If you were to do that with data engineering, there's gonna be 1,000 posts a day or something that mentioned data engineering. It's very difficult to do that, but early data mesh, there were, 100 posts a week. Now there's about 100 to 150 a day. So it's much harder to go through those and keep an eye on it, but finding those people early and finding people that are going to be your flash in the pan, but not in a bad way where you're going to have people that are going to spark other people to get in and you can start to identify people. And then it's not that you identify who's going to be the most prevalent in the community ahead of time. This is the thing that I struggle when talking with salespeople. They're like, I want to know who's going to buy ahead of time. And you don't know that until you're interacting with them. It's like dating. It's not that you go and go, oh, I like their profile, therefore I'm going to propose. The thing is you don't know until you interact with them. And so it is somewhat about getting people in and getting them mixing together and seeing who wants to talk, who wants to be part of this. The Data Mesh Learning Community, I'm rebuilding some community stuff right now, and I'm a little bit down about it because I had such insane success that first time. Going from 0 to 1,000 people in 14 days is absolutely insane. I talked to everybody and they said this is probably the fastest growing data community of all time in that space. And it went from 0 to 4,000 in 6.5, 7 months. But it's gone from that to about 8,000 in the next 2 years. So you can go for vanity metrics, you can go for top of funnel of just bringing as many people in as possible, but you are gonna have people that kind of jump in and go, I'm just gonna consume what's already been written. And those are the people that could be eventually the people that are gonna be talking, or you're gonna have people that jump in and go, this is fun, and then go and find their fun somewhere else. And both are fine and you have to be okay with them being fine, but you have to figure out what are you actually trying to accomplish? Because some of the most really successful communities start with a tight-knit group of 50 people that are all chatting a lot. And I'm seeing that with the Product Ops community. Rather than trying to bring everybody who might be interested in Product Ops in, they're talking from these core people and building out and having more and more people that are attracted to them. That happened with the MLOps community. I worry about this with the data mesh community of bringing everybody in really early, and then if they didn't get what they wanted out of it, if they didn't get all the answers, are they ever gonna come back? And so, you have to figure out what experience do you want when somebody lands and things like that.

So, let's unpick some of the patterns here. When I'm working with a team in an organization, maybe we go in greenfields and we're kicking off a new data team, or we've got a team and we're expanding and scaling. One of the things we observe is as we scale teams in an organization, we'll start to see champions. We'll start to see somebody take a leadership role outside of their core skills and outside the role they have and take leadership of growing that community. And especially if we adopt things like Spotify model where these guilds is really encouraged. But you'll typically find somebody who's passionate, who's interested, and they go and build this internal community in their organization. And they tend to just have that right personality. They just like talking to people. They like helping people be successful. They like sharing knowledge. They like organizing things. They might do a book club, lunchtime book reading, or they might do to show and tell presentations, to distribute some knowledge, but they're typically also doing it because there's something in it for them. They're enjoying that engagement, or they're growing their career and looking for the next step. And so that was one of the things you said, is that when you're bringing people in to help drive the community, to contribute, always keep in mind that there's probably going to be something in it for them because they're donating their time. It's not part of their job. And so, as long as that's a reasonable expectation of what they're getting out of it, they're not coming in to shell everybody, then that's okay. But that's what you're looking for. You're looking for that trade-off of, okay, you're going to donate a whole lot of your time and knowledge and expertise or effort, and you need to get something out of it. And so, that's what we're looking for, people who are willing to engage and are getting the right things out of it. I think the other thing that you mentioned briefly but didn't really cover it into the depth I think it deserved was this idea about time to response. And so one thing that you did really was you monitored LinkedIn and whenever somebody mentioned about it, you popped into the comments with a very simple conversation that went, hey, would you like to come on the show and talk about it? And that was it. That's all I ever saw was you just popping in politely going, would you like to come on the show and talk about it? Now, I don't know whether you followed up in the background again, but it was that time to response because people had posted and you came in with this very polite offer for them to contribute. And I think what I saw was that immediate response back to them, I'm assuming, helped bring more and more guests onto the podcast. If you had waited a week, it's dead, like out of the thread, out of that mindset. Is that true? Did you think that they actually, that ability to respond just at the right time when they're personally interested in what they just posted about helped you grow that community from a podcast point of view?

So let's talk about actually two angles. The first, I'm gonna wanna respond quickly to what you said the, for the first pattern.

Yep.

And what I tell people is ask not, offer a lot. So when you're asking people to join your community, when you're asking people to do these things, you have to think about what do they get from it? And you may not be able to know what's gonna be an attractive offer to them, and that's okay. You put something out there, but make sure that it's an offer and it's not about what you want out of them. It's what they might get out of something that is mutually beneficial to both of you and the community and your audience. The second question I wanna start back with when I was building out the community versus the podcast. So the community, I had the same kind of thing that I posted in a bunch of people's comments, and this is where you can get spammy. Um, and I had a couple of people accuse me of spam, but most were very happy with the approach that I was doing was. Hey, we've got this vendor-independent community where we're talking about data mesh. If you'd like to join or learn more, here's a link. Let me know if you've got any questions. I had about a 25-30% hit rate between somebody was posting and using about data mesh and they wanted to join the Slack community. That's how it was. I was creating something that was potentially of value, and so then they joined. But that's again, inorganic versus organic growth. Organic growth is people find out about it, people are talking about it, people are doing this stuff. When it comes to the podcast, I don't find that people have to be directly in the same mindset, because most of the time when I'm inviting people on, a lot of times it's either somebody I've seen a couple of times, so I don't mind just saying, "Hey, I wanna have you on and we can figure out the topic," or it is somebody saying, "We're doing data mesh," and I especially Also, just keep an eye out for people from underrepresented groups, women, people of color, LGBTQ, to make sure that my guests list doesn't just become white men, because it's quite easy for that to happen. But I don't know that that immediacy when it comes to a podcast episode is that necessary. I think it is when you're talking about inviting them into the community. It did actually work. I saved up all of this stuff before I launched the community. I went and I found every single post that somebody had done about data mesh, and LinkedIn, maybe it was the search capabilities or anything like that, but before, I think it was February 8th or February 9th of 2021, there were only about 350 total posts about data mesh in its entirety. I went and I responded to every single one that were way, way back. So if you've got a niche concept, I think that's fine. I think if you've got a non-niche concept and you're trying to invite somebody in to talk about MLOps. Where there's already an MLOps community that's 10,000 people, and I've seen 5 or 6 other MLOps communities, that you trying to go back and message somebody from 9, 12 months ago, they're gonna be like, what's going on? But if you're focused in a niche area, which data mesh was when I first started this, and then it very much became not, and we can talk about the frustrations there, you do wanna intersect with people at that time if you can. It's that whole thing of the best time was right when they were posting about it. The second best time is right now. If you're willing to put in the effort, you can do that. Now, I do want to, at some point, make sure we, we talk about putting in the effort because that's the other thing with community that I think a lot of people really fail on is thinking that this is a single big push or that If I build it, they will come and things like that. But did that kind of wrap up the question around, do you have to be at that exact time? No, but it is more effective. I was going on LinkedIn and I had my search and I did that every day, Monday through Friday, and sometimes Saturday and Sunday. I did that every day for the early community. When it comes to the podcast, I was doing that pretty much 3 or 4 times a week. Of going through and saying, hey, who should I invite on? And it also gave me market intelligence. Somebody's like, oh, only 10 companies are doing data mesh. And I'm like, oh, oh, my sweet summer child. I wish I had been taking better notes about them, but I've seen 250 to 500 companies and I can't tell you where it is in between those numbers. On my podcast, I've had 60+ actual end-user implementation users of data mesh. So yeah, there's more than a couple of companies doing this. I think that market intelligence aspect also gives you the ability to then keep tuning your message and keep on the pulse as to what people are actually caring about relative to your topic.

At the beginning, when you first started it, how many hours a week would you do?

So this was actually my main focus because I was supposed to jump into new communities about every 3 weeks. And so I did 1 week of prep work and then we did the launch and then by week 2, we were at over 1,000 people. So they were like, hey, don't jump to a new community. We want to see what, whether this has legs. So it was my main job. I was doing like 6 or 7 hours a day out of my 9-hour day of doing this. But what I would say is, what are you trying to do with your community? If you've got somebody who's doing the interactions, like if you can split up that role and somebody's doing interaction piece, Somebody's doing welcoming, somebody's doing questioning and helping answer or pointing people in, you know, that we can talk about that. That's a really great way to grow your community of to say, hey, this other person that's in the community probably has a good answer. Cuz then that person's, oh, I'm recognized as an expert. And that's how you engender people to be answering those questions that other people have. When you @mention people, if you want a large, thriving, fast-growing community, You need to put in a lot of effort. It's gotta be a pretty full-time job if you want that. But to build a community, it doesn't have to be the most vibrant community of all time to be successful. You should always be asking yourself, why am I doing this? What do I want out of it? Because it doesn't have to be that you create the most successful community in the world. It can be that you just wanna find additional people to hang out with and talk on this topic that you really care about.

Again, it goes back to what's that goal? Is the goal to have a community that's based on the number of people who are joining or the number of conversations or the quality of the conversations or just a group of 20 great people that you chat to on a regular basis because you enjoy chewing the fat over the subject. I think the other thing that you mentioned was around rarity of content. So I think that was one of the secret sources and one pattern, which is if it's quite focused what the conversation is, then potentially you're going to get better conversations because it's more interesting than just a data community, which you can go all over the place because there's so much noise in the data community, there's no real focus. So you don't know what you're going for. It's like Twitter, you don't know what you're going to get. Life is like a box of chocolates, but sometimes you just like dark caramel ones. So I think that's one of the key points that you raised was because you had that rarity of content, because you had that singular focus, people who are interested in that focus would come. Whereas if it was just a data community, then potentially they wouldn't. So the, one of the other things you talked about was this idea of multi-channel effectively. And I'm gonna go back at the end of this cuz I wanna see how many of the patterns I can actually relate to almost a marketing pattern cuz you're talking about funnels, but we'll go there at the end when we summarize. This idea of channel and you started off with a channel, which from memory, and tell me if I'm wrong, was Slack.

It was only Slack for a while. Then we created a LinkedIn page and we created a Meetup group. I think Meetup group was 3 months in because we saw that this had legs. So yeah, it was only Slack for 2 or 3 months.

And did you plan those channels on day one or was it a case of as you started to get growth, there was a reason to go to a second channel because there was value in changing the communication style or the medium or the reach? Or had you actually said, look, we're going to try these channels over time?

So I've been listening to a podcast that unfortunately recently ended, but it's called My Dad Wrote a Porno, and it's a very funny thing of this guy and his dad came to him and was like, I wrote an erotic novel, and it's absolutely the worst writing of all time, but it's hilarious and it's so engaging. And so I do recommend people— it's obviously not safe for work, but it's very funny. But when they created the podcast, They created something because they were just like, we want to have this type of thing. I think if you have this super, super specific plan as to this is exactly how I want to shape the community instead of being reactive and going, what are the needs? What are people telling me that they want? I think you're headed for a bad time because you're not going to live up to your expectations. It's going to be different than you expect. I didn't have them planned all out ahead of time. I thought that this would be a community that got to 100, 150 people relatively quickly and then grew 10 or 20 people a week at most. And when I was asking people to join, a bad week was when 80 people would join. It was mostly in the 100 to 120 range every week when I was constantly asking people if they wanted to join. And I think planning out ahead of time, is that very marketing approach. You wanna think about marketing tactics. You wanna think about how things work. If you're just completely oblivious to marketing, you're gonna struggle in building out a big community. But again, that doesn't have to be your success metric. If you're super focused on this is exactly what I want out of it, this is my exact output instead of focused on what are the people that are joining gonna get out of it, I think you're gonna struggle because you're going to be so focused on monetization of this. I see a lot of people that are like, we wanna create a community because we wanna sell to this community. And it's as soon as they find out that you're creating it to sell to them, they're gonna leave because very few people wanna be part of a community where they're constantly being sold to. Being sold to as part of the community thing is fine. If that's not the main goal. But I saw this working in open source, open source software, DataStax, Apache Cassandra. So many of the salespeople did not understand that open source users don't want to be sold to, but you need to do awareness marketing. You need to make sure they're aware of everything that's available. And so they need to be aware of how to get help when they need it. And when it's about time to be like, look, you're having so many troubles in this way or that way, or you need these features. Come on, pay for them because they're of value. And there are people who are open source sellouts and all that stuff. We did those topic swarms and looking back on them, they were actually what we should have been doing and we should have kept doing them and finding that core group of 20, 30 people and just continually expanding it out and continually inviting people and just saying, hey, you should join and making it okay to have people come on to a group conversation and not say anything and just be in the background and that they're not a failure for doing that. But if you have overly focused on exactly where things are going to go, I think you're in for a bad time because you're going to be overly focused on exactly what you're trying to get out of it instead of what's best for the community members, which then leads them to want to continue to interact with the community.

Again, if we take that idea of a funnel though, So we know, yeah, in marketing, if we're going to do direct email, there's a certain percentage that's the average rate of email opens. If you're saying building a community around a newsletter, you're going to go and have this many subscribers. And when you send the newsletter out, there's a reasonable open rate that you're aiming for. So yeah, you can be amazing sometimes and you can be crap, but there's this industry standard for what good looks like. I'm just wondering with a community. Is there a sense of a funnel of what good looks like where you say, if we have this many people who have joined the community, we should be able to see this many conversations being had or this many responses to a thread or this many people actively communicating in that community twice a week? Do you think there is a sense of metrics that mean a community's healthy, or do you think it's unique to every community?

I've got a couple of different thoughts there. What I recommend if people are looking into this to think, to check out orbit.love. So there is a software offering, it's paid, but they've got a lot of conceptual things around community and community health. And you start to think about how close is somebody's orbit? Are they at the core or are they way out in the thing? And you're trying to convert them to bring them in closer and closer to the core. It's a little bit like that old Judge quote about pornography of, "I can't define it, but I know it when I see it." I would say that the data mesh learning community isn't nearly as vibrant as it should be right now. I think especially for the next couple of months, because it's summer and Europe is a big market for data mesh. I think anybody who's got a community that's highly focused in Europe, especially if they're US people, they're like, Why did all of our things fall off for these 2 months? And it's literally people in the Nordics go on vacation for 4 to 6 weeks consecutively, which is absolutely astonishingly, you can't even consider it in the US because if you take 2 weeks straight in the US, people are like, what's going on? That's an inappropriate type of thing to do. But I think Orbit gives you a sense of that type of thing or other tracking metrics. If you think 'Every post should get this many responses.' That's a problem versus, 'I should be seeing this many responses in total,' or, 'I should be seeing this much interaction.' One of the things that we had in the Data Mesh Learning Community that I think was actually slightly toxic was there would be these back and forths, and there were a couple of people that would go on and on in threads. And so somebody would wake up and maybe the thread had 3 responses. And then especially me, I'd get up the West Coast US and 2 people in Europe would've had 100 responses. And I'm like, I'm not gonna read that. And so it made it so that there wasn't a good ability to do that. You don't have to think about Slack as being the only thing. You could also use Discourse, not Discord, but Discourse is a message board type of thing. And I wish we had done that. I wish there was a software that actually integrated real-time chat and message boards where you could just pin this to a message board and people could read it because Slack is so ephemeral. And I have people who've said that they join and then scroll back and try and read all this stuff historically, especially because Slack, you only get 90 days because they won't put us on a free plan that does the actual historical stuff. People miss all of the framing of that conversation. I think if you get overly focused on this is exactly how it should look. It's like when you're dating someone, it's like a relationship. It's like a friendship. It's like a company. You're building out your company. Is your company exactly what you thought it would be? Even 6 months ago, is your company exactly what you thought it would be when you founded, or are you responding to what's happening and trying to iterate? It's agile, right? That's how you do community stuff is you look at what's working and you try and pour some gasoline on that. You probably try and push some force behind that.

I think we'll just coin that, that's called community ops, cuz we're gonna put ops on the end of everything. So I'm just thinking about that definition of good. An example I use is I needed to get better chops at being a product person because I'm leading the product development for us and I've never really been a product person before. So I tried a few things and I ended up at Report, part of their cohort learning, and I loved it. I loved the cohort learning because I could go with a group of people. Their material was just fricking brilliant. And they had a Slack community and it was quite vibrant. And because I was constantly going to Slack anyway, because we use it internally for our internal comms, and then there was a data mesh one, and there were other, couple of other Slack groups that weren't zombies. You'd go into a cohort and automatically people were slacking each other going, let's create a study group. 8 or 10 of us will get together and meet once a week and figure out where we got stuck or where we didn't. And that was great for me because effectively it became homework. Didn't want to turn up to 8 of my peers not having done the work, so that forced me to to do the work for the week. And I love that. And then the last cohort I went on, they moved the conversation out of Slack into their own Reforge app. Right now, this is the app that you go to, to see the reading material and watch the videos. And it died. I posted, I love doing the study groups. It's really helpful. And it was like crickets. I met up with a really cool person out of Tokyo and she was just fricking awesome. And I got to do my homework, but that whole mover channel to where people weren't naturally viewing they weren't naturally in that channel every day, definitely would've killed their collaboration. I think if they were monitoring the number of conversations, they pretty much should've seen that it's gone from being amazingly vibrant to absolutely dead, even though it's the same people who want to have the same conversations. I think it's that balance of not having fixed metrics that we're going to have 123 responses to every thread, but we are going to see where the channel or the things we're experimenting with don't seem to be working so we can make a change and see what happens.

When you think about community, you have to look at where are the points of friction. And if there's a lot of friction in participating, are they getting enough out of it where the cost-benefit that they're actually going to be like, there is more benefit than the cost of that friction. And in most communities, the answer is nah. It's just, this is fun. This is for me to learn. 'If this is so tough, I'm going to go elsewhere.' And people also don't like being captured and, 'Oh, you're putting me in your app, you're tracking me, you're doing all of these things.' I want to trust that the communities that I'm part of have my back. And so that's the other thing of, we want to control it because this is the best way to control the space. Most people don't want to be controlled because then they're becoming a lead, they're becoming the product, they're becoming the commodity. And so, as soon as you start to do that, then it becomes suspect. Again, most people are on Slack in some form or fashion, so it's not that huge of a deal. You have to think about friction to join, and maybe you do want to inject friction in that joining process to only get those people that really want to be part of this. You can do that. That is a valid way to think about it. But yes, exactly what you're talking about of This became a thing where you were getting so much value out of it, but people weren't even testing to see if they would get value out of it because there was too much friction in the process. And I have to stay logged into this random app and is it only on my phone or, okay, maybe it's on my computer, but it's clunky versus this is part of my day-to-day. I can just jump into this and have these conversations. There isn't a right answer. What are you trying to achieve? If you're trying to build the biggest community in the world, you have to lower the friction entirely, and that's where you probably want to go with Discord because there's just no friction to dealing with Discord, but it's also not threaded well. It's not like doing all these things. It's very ephemeral. And so you get a whole bunch of people and they might be posting, but nobody's actually connecting with each other versus Slack. People connect with each other and things like that. Where is your friction in your process? You need to take your product and your marketing hat and go, is this juice worth the squeeze? Is it worth the squeeze for us? Are we filtering out too many people? And is it worth it for the people that are coming through where they're gonna tell other people to come through? Because that word of mouth marketing is the best way to bring people into your community. Otherwise, you gotta have someone like me that's literally going out and going, hey, would you like to join? Hey, would you like to join? Hey, would you like to join? And most communities, people, even companies that are trying to build big communities around their product, aren't willing to invest in that person doing that.

They should.

Because that's the best way to do that. Every community should have 2 people that it's their full-time focus if you're going to be building a community or a company around that community. Friction sucks, and if you're injecting friction into people's lives, many will opt out.

Just on that, so, you know, DBT is another Slack community that went viral in my head, but I never bothered to go find out whether it was a conscious or unconscious thing, did they actually have a couple of people building that community? It was part of their go-to-market play, or was it just lucky?

They did for sure. They had at least a couple of community managers and they were focused on this. Another one that I point to that my good friend, Ale Murray, really was the one who built that out is the Kafka community. They intentionally built a community and they elevated people, I did this with the Data Mesh Learning Community when I was really focused on growing the community, what I was seeing from the community was just more conversations and no actual forward movement, which is why I created the podcast and focused on the podcast was that, okay, everybody's talking, maybe a couple of people are getting a little bit more information, but there's not a specific thing of, I'm going to pin you down and ask you these questions. There's another one in data that that I haven't joined. I have some problems with the way that they run it, but it is a very intentional way of running it, which is called Locally Optimistic. It's been in the 3,000 to 4,000 people range for a while, but they boot people out if they haven't been logging in a certain amount of time, I think. And you have to specifically ask to join and they have very specific rules. And I think some of their rules are draconian and weird, but they've built a community that the people that are part of it love it because it is insular. I'm on the social networking app, Bluesky right now. And I've been on it for a while and it's invite only right now. And they're not trying to do it as an exclusive invite only, but they're trying to do it so that they can manage it. It prevents there from being a bunch of people in that you don't necessarily want in, but that can be very exclusionary. That can have all sorts of problematic aspects to it, but it's something where you have to go through friction to join. And so then you're more likely to at least say, was this worth it? And really check it out instead of jumping and go, hey, I'm here, poke around for 10 seconds and then go, oh, okay. I talked to the Team Topologies people when I got them to join the Data Mesh Learning Slack. They were like, oh joy, this is our, I think our 39th Slack that we're a part of. And I think that they've probably added 25 or 30 cents. And it's again, that friction can be a good thing. It doesn't have to be that's bad if that's what you're trying to create, but what are you trying to do? This is the thing. How do I create a great community? What does that look like? What does that mean? It's just not, how do I do data mesh? What? That's not a question.

Like, what do you do? How do I buy a data mesh? All right. So, definitely a feeling of a flywheel, the idea of growth begets growth, and there's some things you can do to encourage that growth. No structured plan's ever gonna get you there. So you have to inspect and adapt and that's all good. I'll see how the Bluesky thing goes. I remember, I think Superman or whatever the email one that, that used the same thing. And there was that coffee chatty podcast one that tried the whole exclusive thing for a while as well and then crashed and burned. I can't remember what that was called.

Clubhouse.

Clubhouse. That's the one. And actually today they've just opened up Threads. So the evil meta Facebook tracking company is going to combine Insta, TikTok, and Twitter and Bluesky into a single thing. And we'll see we'll see how that one goes. So exciting times in the world of not curated communities. So let's go from this idea of marketing and the flywheel. And then there was another thread that came through, a whole set of patterns around community ops, around actually running it properly, where there's a bunch of things that, you know, are gonna happen. How do you deal with them? So let me start off one that I'm really intrigued with. We start off with a community, we'll have a core group of people that are interested or knowledgeable, we start to get growth. People have learned some new things through being part of that community. So their fluency in the subject's increased. And as we get more people coming in, they're starting from ground zero again. So we get those air quotes, dumb questions, but they're not really dumb because they're the same questions we were asking 3 months ago when we were noobs. And you get this disconnect around the fluency of people. You get that horrible, I'll just go read that thread. That's not helpful. So how do you deal with that? How do you deal with this imbalance of fluency, this imbalance of time in the community, still being inviting for people who are going to come in on you, who quite rightly need to ask those basic questions again?

There's a thing that I started to do around panels for my podcast where I say, here's the FAQ, but it's actually a vibe/FAQ document. So vibe is more important. What vibe are you going for? Is this a very professional place where everybody is very professional? Is it something where people can let their hair down? I think that you think about your onboarding process to tell people how that works. And then I think you create spaces. There are people that love working with people that are new to a concept and being like, join us, join us, one of us, one of us, gooble gobble. If somebody doesn't know what that is, you should go look that up on YouTube. It's so strange, but so entertainingly weird. In the Data Mesh Learning Community, we created a channel. I think it was just beginner questions when I did it, and then the new team changed it to ask beginner questions. It's a space that literally says, here's where you can do that. There's a lot of these things inside of Slack where they get sent a DM, but that's not even a DM. It's a thing that just pops up on their screen for when they're coming in, and you can explain what things are. If anybody was joining the Data Mesh learning community and saying, what is Data Mesh? People would be a little bit frustrated. I did have one person who I was talking to them one-on-one, and I mentioned Zhamak, who's the creator of the Data Mesh concept. I know you know that, but maybe some of your listeners don't. They'd been part of this community for a while, and they said, what's a Zhamak? And I was like, it's probably my favorite Data Mesh question of all time. If somebody were just to pop in and say, what is a Data Mesh? People would be frustrated about that because that means somebody hasn't done absolutely any homework before joining. I created a channel that nobody else could post in. It was only admins that said, here's your beginner content. And I created some threads around beginner content. So somebody that's just trying to learn can come in and not have to say, what content should I start on? So making it non-intimidating for those new people, but also making it clear to those existing people that new people are valid and that they're welcome. Creating that vibe is important, but it's tough to do that unless you do that from the beginning. You can just ask 5 or 6 people that when people ask these beginner questions, point them to a little bit of content or do that, and that we're not going to be really frustrated with them because it can be difficult to learn on this stuff. And they're coming to a space that they're hoping can help them elevate their career or learn on.

Excellent. So what other community ops things do you recommend? The ones that were in there were set up the automated onboarding. So when people turn up, they're going to immediately get a response, even if you're busy, that says, here's the vibe of the things, here's the things you need to know, here's the rules. Set up spaces that match the fluency. And I'm intrigued because it means that actually over time you could harvest the great content from the community and put it into those beginner guides. Because those questions are getting answered in a way that goes, that's a bloody good answer. Let's reuse that for the next person who can't go back 90 days or more because the cost of Slack is so high. What else? What else are some of those operational things that keep a community vibrant? There's obviously the Be Shameless Vendor Slack group, which everybody needs to have. What else is important if you're building a community to put in as your belts and braces, your meat and potatoes?

So if you are looking for vibrancy, you need to have a funnel acquisition strategy. You need to have, how are people gonna get in? What do you want to do with that? When we were first starting out, I asked 10, 12 people that had a decent amount of LinkedIn or Twitter followers to tweet out or post about the community and say, this is a cool thing, or reshare the post saying, hey, everybody should come join this. There's consistency of messaging. Because of my move to Europe, I haven't been posting nearly as much about the Data Mesh Radio podcast, and we've seen consumption go down. There just aren't as many downloads because I'm just not posting about it as much, so people aren't going, "Oh, I have to do this." So, you have to have that kind of external thing if you want to be putting up signposts. And then there's the person who's facilitating the conversation. I tried to create this space, everybody was really interested in it, and nobody did any work for it. But where people could find a group of people to review their content before it went out. That's a great way to encourage people. But again, the logistics of that, I couldn't get anybody to run any logistics. And so I think the first thing you wanna be doing early in your days is try and identify people. I did do this, Andrew Padilla, I really thank him for running our newsletter. That was draining a lot of my time cuz I was always like, is this incremental to me? No. Is this incremental to our audience? Probably. And so then, what's incremental? And I'd have to write it up. And I was just always reading all of this, what is data mesh content? What is this data mesh content? I really thank him for taking that on. But it was hard for me to find people that would be consistent, that didn't have too much of a self-serving angle. There were some consultants that were saying things that I really didn't even agree with, but also, that were always posting and then saying, or you can set up a call with me around sales. And so you want to find those people and encourage them more than anything, because you need to share the load. If it's a one-person driven community, which it became that for me. And anytime I would take a break, a thing in September or August of 2021, where I was just like, I need a break from doing all this external stuff, I'm gonna do some internal focus stuff, the growth stopped and a lot of the communication stopped because I wasn't facilitating the conversation between people. And so you need to figure out if you're going to have a full-time person doing that, great. If you're going to pay for it, great. But you need to figure out what's a sustainable model and what is good relative to sustainability, because it's too easy to just focus on Again, the metrics. I think the community ops thing is find the people who are going to take on some load, give them a load that's, but that's beneficial to them. Ask not, offer a lot. Find something that's going to be mutually beneficial to them and build it out slowly and don't be in a rush. I was in too much of a rush and then declared things had failed because I was doing a fail fast, which doesn't really work with community because people don't fail fast like systems and software do. You're going to have to put in a little bit more effort and a little bit more gas to test what what works and then take that, especially take the people and keep dragging them back and be like, hey, I want to put you on a pedestal. That's the other thing is elevate and tell people that they're great and people that are doing great things, but let them know that you appreciate what they're doing.

I agree with you in terms of just reminding me what you said about Europe going on holiday because in New Zealand, they're pretty much half of December and most of January is our summer. And actually, the government shuts down, everybody shuts down, we just all disappear for 6 weeks. Pretty much if you were looking for any kind of engagement from New Zealand, unless it was holiday engagement, you got Buckley's. One of the key themes coming through though is it's a team sport. Running a community is not hiring a community manager saying good luck and expecting growth. Is there a team design? If you had to think about optimum size of team, what skills or roles, I'll let you have roles for once, what would you actually build if you were going to build a community team? How many would you have in it? And what are the roles or tasks that you have people focus on? Because there's got a whole lot that's been embedded all the way through us talking. I just want to bring it out into almost a team design or a team topology.

To be successful at this, you need two people. You need somebody that's going out there and finding the right people to make sure that they're part of the community, that they're a voice that should be heard, and inviting them in. You're not going to get 100%, you're not going to get everybody coming in, but but you should be doing that. You need to have a focus on that onboarding experience so that people actually stick once they get there. You need to have somebody focused on enabling others to become a voice in the community. You need somebody that is elevating other voices and making sure that they're heard. You need to create this flow of content and you need to create a filter for saying, what is good content? And that can frustrate people when their content isn't included in the newsletter. It can discourage people. But what I found is people want you to tell them what to read and what was the most important part of this. And sometimes people just want the bullet points. A lot of people with my podcast just read the bullet points that I put out for an episode, and maybe if it's something that's really of interest, they listen to it. I try and make it so that people don't have to listen to an episode they don't want. You again are looking for the friction in whatever is related to your community. Is it on learning? Is it on doing? You think about open source community stuff and you have evangelists and, or advocates, and they're building a driver or they're building a connector or they're building all of these things that reduce friction. So you need somebody that's focused on understanding what that friction is. And if you're building a community around it, that you are pushing that information into product, that you're pushing that information into sales and marketing, that you've got that cohesively tied in. You have that person that's engaging with people, and then you have somebody that's building out whatever it is, if that's software, if that's content or whatever, that they're focused on taking that and making sure content gets created, whether it's by them or whether it's by somewhere, someone else. And that content again can be software, it can be actual written content, it can be videos, it can be all that stuff. So I think you only need 2 people, but you need to invest in 2 full-time people and everybody tries to invest in 1 and you're just setting yourself up for failure. And then as you scale, I think when I was on DataStax community team originally, and why our community was really kick-ass was. We had, I think, 7 people on the community team in total. And we had a director level and then we had me and Ale were at the senior manager level. And then we had just 4 community managers underneath us. And then we had an evangelist team that we paired with that was very good. But the thing that really made that flow was that we had people that were specialized in— I never really managed the relationships with people because because honestly, I'm not a very people person, which is funny for a community person, but I like talking to a few people every once in a while, but there are people that love to talk to people all day long. And so we had that person that was just very charismatic and very vibrant, did that. I think you wanna, again, figure out what you wanna do, but I think you can do this with 2 people. You don't have to build out a huge thing, but you have to also be realistic about what they can do and stop trying to hire, especially open source people, stop trying to hire somebody that's gonna manage your and write a bunch of content and create a bunch of software. You're overloading them way too much. That's just a stupid idiotic approach. I'm gonna just push that there. You only need 2 people to do this. And then as that grows out, you're going to see, oh, okay, how is this flowing back into the company? I'm thinking about, there's a difference between just the communities that are created for communities, but around your company, you should be thinking about how does this flow back in and not how do I sell to these people, but how do I make it so that I don't have to sell to these people and they just buy. You can extract that information from them for free. They're out there, they want help. And as part of that, you can learn, but you need to be doing that and pushing it back in. But you can do this with 2 people relatively easily and grow a big community.

Excellent. I got one question to close out, but before we get there, anything else that we haven't covered that you're like Is here's the gold, is the patterns that I haven't thought about or mentioned.

We just talked about it. Invest. If you want a community to just become very vibrant out of nowhere, it's not magic. It is constant pressure and it is not, we're going to do a big push and we're going to back off. It's just like relationships. Very few relationships are very good if you're not in contact with people if you're, oh, we want to sell you this new thing that's coming out next month. So just invest, companies, please. It's not hard to have people that are there that are talking to people and they'll learn. You don't have to hire the most, especially a community manager role, you don't have to hire the most experienced person. You need somebody who's personable, who will want to talk with people and want to make sure they get the help they need.

I think one of the key things is just keep it real, be real. Had a guest on and I wanted to get a bit of background. So I listened to a couple of podcasts they were on and one of them was from a very large vendor and I couldn't last more than 5 minutes because it was obvious that the host was reading at the beginning. It was so obvious that you wrote that and then read it. It didn't come out as a natural human conversation. Why do that? I'm just never going to go back to their podcast again because I was just like, no, I can't live with that. That. All right, so last question. Like most startups and like most data teams, they get started a lot of times by a very passionate person, and they grow and they get vibrant. But when the founder leaves, often it dies. And we see that a lot with communities. Even though the community gets behind them, the founder of that community, the person who's been driving it, blood, sweat, and tears of 7, just 7 to 8 hours a day, especially when they're on their own, they leave their community and the community just withers. Do you see that? Do you see that happen a lot in the communities you're seeing?

Even I think in the data mesh learning community, it's still reasonably vibrant. It's not completely zombie, but once I backed off from really being involved in the day-to-day conversation and trying to let other people percolate up, it didn't happen. You have to create those additional champions. And as the founder of those, you have to anoint people. You're perpetuating all sorts of issues and all that fun stuff. Fun stuff, but you have to anoint people and say, this is somebody that is part of the people that you should listen to so that there is a vibrancy. Even when you take a vacation, do you want the whole community to shut down? If you're US and you're going on your honeymoon, then you get to take 2 weeks, maybe even 3. You have to encourage those people. You have to let them know you value them. They're not just talking. You're seeing what they're saying and you value their opinion. That's where you can get that to happen. Because otherwise, yes, so many of these communities are flashes in the pan because they're cult of personality. And then somebody tries to start to monetize instead of focus on just the community, which is a valid thing. I'm starting to try and monetize the Data Mesh Community stuff that I've put together, not in a bad way, but there's so much of this cult of personality aspect to these communities. And if that personality leaves, then the cult disbands because it wasn't ever about the cult, it was about the personality.

I look at it from a different point of view, actually, and it's only talking to you on this that I came to this realization is often when the founder leaves, they're leaving a gap of 60 to 80 hours a week of work. And it's 60, 80 hours for them because they've done it for so long. They know what they're doing and they've got the shortcuts, they got the ops. Jobs. And so they're typically replaced with a bunch of people who are part-time doing it in their spare time, who haven't done it before. And the 60-day hours become 120, and it's broken up across 20 people, when we know communication between 20 people is a nightmare anyway. And they've all got lives and jobs. That momentum of one person doing that, that hard work gets dissipated. And that's probably as much a problem as the cult I think it's a combination of the two. So as we know, when change happens and there's a cost and a consequence, and so when the founder of that community moves on, there's going to be a good change or a bad change, but there's going to be change.

And there, there is one thing I would push back on within that. So I fully agree with you, but I also don't want to discourage people out there. I've talked to a bunch of my friends and said, you should create a community around this, and it can be an hour or two a week of work. You just have to figure out what's working and what doesn't, and that you have to be patient and that it's not going to be this massive thing immediately. But you can do this stuff pretty easily. You just have to have reasonable expectations. It doesn't have to be that 40 hours a week or 60 or 80 or whatever. But yes, I had the former CDO of Jaguar Land Rover on and he gave the great quote of If you want to go fast, go alone. If you want to go far, go together. If you want there to be a cohesive thing, you have to go together, but there has to be that cohesive plan. And when it is 20 people, who's coordinating all of those? Even if you did have that 3 hours a week for 20 different people, if there's no cohesive plan as to how it goes forward, everybody's just trying to do things, then everybody's just pulling in different directions and everybody's the the 5-year-olds playing soccer or footy, and they're all just chasing the ball instead of a cohesive plan of, okay, you actually have a position. You're, we're going to move the ball around in this way. This doesn't have to be a massive amount of work. You can make it that way. That's how, if you want a very big, very successful community, but you can create a community around a topic relatively easily. You just go out there and find the people and invite them in. The thing that, that I never see community doing right is inviting the people in because they're either inviting them in to try and sell to them or they're not inviting them in at all. This is the thing where I said, this stuff is so easy. You just find where the people are talking and you just say, I have created this thing that seems like it is potentially of use to you and of interest. Would you like to take advantage of this thing that is potentially of use or of interest to you? And that's it. People are so afraid. They're like, I have to craft this perfect message. And they go too far. I have this when I'm asking people on the podcast, when I send them a private message, the ones that are 400 words or whatever, get a very low response rate versus the ones that are 30 words. Sometimes it's, I know I only have to do 30 words to get this person, but the, hey, I've got this thing. Is this of any interest to you? And if it's no, okay. If it's yes, great. This doesn't have to be rocket science.

But there are a bunch of core patterns in there, and we can use those patterns for external communities, and we can use those patterns for internal communities within our organizations to create people we wanna talk to in our organization that have a shared passion. Excellent. All right. If people wanted to get hold of you, apart from the whole data mesh place, where do people go to nowadays to find out what you're up to and see what's happening next?

LinkedIn is the best place, but if you type in data mesh and Scott, anywhere, you'll find me very quickly. There's only a few Scotts that are involved with data mesh, and I'm the one that has been talking way too much about it for way too long. So you'll find me very quickly. Scott with two Ts, but that's the easiest way to find me rather than trying to spell my last name, which is a made-up last name anyway. It's a weird story behind that, but yeah.

Story for another day. All right. Thank you for coming on the show. It's been good. I think we've got some pretty good patterns here that we could try and apply. I'll have a think about some of them a bit more, and especially when I'm working with teams, how we can get those champions out there. And helping the data teams on their communities within their organizations as well as outside. Thank you, everybody who's listening, and I hope everybody has a simply magical day.

Yeah, thanks for having me on.

That, Data Magicians, was another Agile Data Podcast. If you'd like to learn more on applying an agile way of working to your data data and analytics, head over to agiledata.io.