Flock 2025 "Build It And They Will Come" And Other Myths

content, my content, right? It should should be fine. Should be I want the speaker notes because there's some extra stuff that's not in the slides. All right. Well, let's let's uh let's get it going then, shall we? So, build it and they will come and other myths. Um I this title has a long history because I've been trying to work out how to message this content for a long time. Um and what we're going to talk about here is uh quite a lot of things actually. Um there's a lot I could talk about. Uh but this is what I've chosen to talk about with you here. Uh and thank you to Eva for the long slot uh to cover it. less. Thank you for putting it at the same time as the CL AMA, but what can you do? So, this is what we're going to talk about. There are many things we could talk about within the world of community building and community architecture. Um, and I'm I've picked these four as I things I think we should think about. But before I get to that, why should you listen to me, right? Um, well, the answer is you shouldn't. Uh, I could be anyone, right? Uh, who am I? I am a longtime Red Hat employee. I have been a community architect for about the last 10 years uh with satellite with anible with bunch of other things. Um but I just recently joined Fedora on the CL team. So I'm now back to being a CIS admin. Uh so anyone the point I want to make is that we have enough anecdote in the community building world. There are any number of books and blogs and essays that you can go and read on what worked for me and that's good. There are talks at this very conference of that nature and they are good too. That stuff matters. We should absolutely listen and learn from each other. But at some point we have to say what do we know from research? What do we actually know from having done experiments and done investigations into how communities work? And so that's what I want to talk to you about. That's what I want to to understand and to try and pass that message on that actually there are things we know. There are things we have studied. Some of my examples are older than you might think. And we can use them. We can use them to design our communities. We can be much more upfront about how we build our communities, about the choices that we make as community architects, as community leaders, whether that's at the top of the project with Jeff or at any point further down or some other project that you're involved in. There are things we can use. There are choices we can make. And we can be much more explicit about how we do that. So let's get into it then and we'll start with the title slide. Build it and they will come. No, they won't. It turns out to be much much much more complicated than that. Let me demonstrate. Someone give me a number. >> Thank you. That took about 3 seconds. Um and that was because there's not very many people in the room. You do it in a bigger room, it takes longer. Um and if I said Kevin, give me a number. >> See that was two seconds. That was quicker. Now why am I doing this? Well, you can probably guess, right? So I'm I've got the example here. If I want to get something done in my community, right, how do I go about getting it done? Well, there's been research, it turns out. Um, and I have to press the button on both laptop and on this as well. So, this is research been done 25 years ago now. Uh, but in an online chat room, who recognizes this behavior, by the way, in an online chat room, requests for help are answered 50% faster when a recipient is addressed by name. Anyone seen that behavior? Yeah. Yeah. A lot of nodding around the room. Um yes, it is absolutely true that if we name people, if we ask for help from the right people, we are going to get a better result. Now, I will just draw attention to the plot. Um that only goes up to 20 people. I don't think that's linear. I think that pretty much tops out after about 50 people. And also, the response time appears to be measured in seconds. I don't think that's necessarily correct. That might be a typo in the paper. I think it's probably more like minutes. But nonetheless, I think the result does hold. And from our own experience, I think we would agree with it. Right? So this is this is me trying to give you something that to to say, look, this is something we've experienced, but there have been studies done on this. Right? Now, this is 25 years old. Right? But the point here is that we if we make specific requests to the right people, we're going to get a better result. Right? If we just broadcast it into the void, doesn't really work. And and and what does that mean for us? If we're designing our communities, that's what I started up front. I said we can be more designled in how we build our communities. What does that mean for how we put things in newsletters? What does that mean for how we put out calls for help on social media? What does it mean when we do at room in a in a channel? Right? Actually, if we know if we have better data on who's active at the moment, who might be receptive to a specific task and we can reach out to them and say, "Hey, you're up and coming. We've got something that needs doing. Would you be interested in helping?" You're going to get a better result out of that. Now, it turns out that two 25 years is actually really new. Um, there was some research done in 1968 on this, and I loved finding this paper. Um, so it turned out that if you have someone and you're doing CPR, there's been an accident. Um, you're doing some first aid, you're doing some CPR, and you shout, "Someone call an ambulance." The bigger the group, the less likely anybody will actually call an ambulance. And this is what 50 plus years old now. this research and we can all relate to it, right? Because it's the same problem just we see it online but it turns out as humans we just we just do want to let other people do things. Douglas Adams famously called this the someone else's problem field. Right? So we we have that problem and if we and and this scales again go back to the graph on the previous slide. This scales with the size of the group that we're addressing and we have hundreds if not thousands of people in our groups. So we need to think about this right we need to be much more um aware of that problem and say well just addressing everybody is going to lead to almost everybody saying that somebody else will deal with it. It's not going to work right. Um this actually shows up in a whole bunch of other stuff. Um there's a whole model of social persuasion uh from 1981 uh which uh looks at not just the immediiacy of the number of people around you but also the influence of the people asking you to make the request. Um if you identify with them, you're going to get a better result. And there was some research in 2008 on get out the vote campaigns. Turns out that h if you send people door too and knock on doors and ask them to vote, it is even accounting for the vastly higher cost far more effective than leaflets and emails and banners because it's mass appeal, right? It's the someone else's problem field again. So we know this. So let's use it, right? Let's use this in our communities. Let's use this in our decision- making. Get it in front of the right people. Who the right people are, that is a different problem to solve, right? And that's where we have to have better data. That's where we have to have better understanding of of what we're dealing with. Okay. So, going to move on to myth number two. And I'm that one's quite a quick one, right? Because I think we all recognize it. We've all been there. The next one's another one where you've probably been there. And then the last two are slightly more interesting and perhaps counterintuitive. So, let's talk about myth number two. Sadly, I couldn't come up with quite as many pathy funny titles for the other myths, so you'll just have to live with that. Um, so people contribute according to their time, skill, and alignment. That's alignment to the project. This is not D&D. I'm not talking about them being neutral good. Um, this is this is not true. Well, it is true. Of course, it's true, but it's it's nowhere near nuanced enough, right? There's there's so much more going on here. And this this leads us to talk about one of my favorite things. What I'm what I'm taking you through here, I should have said this at the beginning, is a bunch of research that comes from sociology, comes from psychology, comes from economics, and it comes all together in online communities, right? We have to consider all those things when we build a community. And one of those concepts is a thing called social proof. Press the right button. There we go. So, people follow the lead of similar others is the is the formal definition that comes from Kildini in 2001. Um, but that's not a particularly intuitive statement. I don't I think I need to explain it a little better. Um, we tend to want to follow other people in the sense that we don't want to be caught out. We don't want to be the only one doing something. So, I gave that example earlier where I asked for a number and it takes a moment before anyone will answer because no one wants to be the odd one out. Right now, I really need to make a point here. When we talk about designing communities, we all like to think we're super individual. We all love to think we can't be told what to do. And particularly in the tech communities, we're all completely indie different to all everybody else. Um we don't follow normal rules at all and that's fine and that's true. But as soon as you put a bunch of us together, then all the usual human psychology applies. And that means as a group, we can design around likely outcomes. Sure, there are going to be people for whom that outcome doesn't work, but on mass and the bigger the group, the more it will hold, right? And social proof is a great example of this. Um, and that number example is is is a very simple one, but I'll give you another. Think about um network effects. Okay, think about not wanting to move, getting your friends off of Facebook or some random social network, right? You can't get anyone to move until someone else moves, right? That's the network effect. This is the inverse of that. I don't want to be the first one. And it works incredibly well. Um, there's some some fantastic example of this in our communities. Think about a forum, right? How many times have we posted a discussion topic and got nothing? Because no one wants to be the first one to reply because they might be the only one to reply. And that's the thing that we're risking. Now, I want to be clear, it's subconscious, right? It's not we're not actually sitting there going, "Oh, nobody else might reply." No, it's not at that level, but it is a subconscious thing. We don't like being the odd one out. We are heard animals, right? And so, we've got to think about this social proof thing because it applies to forums. It applies to discussions that we have in other places. It applies again to the calls for help. It it's it's all over the place. Um, think there's so many examples to this. Uh, think about political parties that start their membership numbers at 10,000, right? Because they don't want to seem small. Think about has anyone ever seen a teleathon that started at zero dollars raised? No. Because you're not going to put go on TV and say, "Hey, are you going to be the first person to contribute to our project?" No. You're going to say, "Hey, we got $100,000 already. who's going to double it by the end of the program, right? You've got to give people some idea that they are already part of a group, even if it's not actually true. This this goes on everywhere in our hotel. Actually, I checked and this hotel doesn't do it, which is kind of annoying. Um, pretty much all hotels will will say, "Hey, you should save the planner and reuse your towels, right? And not like if you hang them up, then we won't wash them and if you leave them on the floor, we'll change them, right?" They all do that, but some will say things like, "People who stayed in this room reuse their towels." And that's now appealing to the social proof thing because the people like you, people who chose the same level of room as you, people who chose this hotel the same as you, they were the ones reusing their towels. And it turns out it's vastly more effective if you look at the rates of people actually doing this. when they start making appeals based on people similar to you, then you get better usage of of people taking up that offer or taking up that that behavior. So, this matters amazingly well, but I want to park it for a second. I'm going to talk about how we designed that into our community, but I need to add something else to the mix first. So, let's talk about goal setting. Nobody wants to talk about goal setting, but we all all our companies do it, right? I know I have to do it, right? I have to every quarter I have to be like what what are my goals for this quarter, right? Turns out there's a reason for that. It's really really effective. So here's a nice review that I really really like. Um this was taken from a movie review community. Um now this is quite an old example. Again, it's from 2004. The internet has changed, but humans haven't. So what happened here? This is a community that reviews movies. And so what they did was an experiment where they sent out emails to different cohorts and they said, "Okay, this group here, uh, we're we're doing a drive. We're doing a hack day. We're doing call it what you like. Um, we want to review as many movies as possible. You do as many as you can. You try and do eight this week. You can do 16. You can do 32. You can do 64." Right? Turns out that even even completely unrealistic targets outperform just telling people to do their best. Um, I mean, 64 movies in a week, no chance, right? And of course, it isn't the highest, but I I find it incredible that 32 did the highest. They didn't get to 32. Of course, they didn't. That's still a heck of a lot of movies to watch in a week, but they did manage to do a lot more than all the other groups. And this has been shown again and again that picking targets and trying to get people to go for them really matters. But here's the thing, we don't do it, right? We don't do it in our communities. So if goals are more effective and you also need progress, right? You also need to understand how you're doing on your goals. It's not just enough to set a goal. You have to know if you're getting anywhere near it. And I will take issue here. I have no idea how we're doing on Fedora's contributor strategy. I don't know if we're going to double it by 2028. I don't know what the progress is. So how are we doing? It's it's important to be able to measure that stuff, right? So let me get back to what I promised then. Let's talk about the design claims here. If we're going to do a hack day, if we're going to do a test day and we we're now aware of things like social proof and goal setting, let's use it. Let's set a goal. This is how many bugs we want to get done, right? Can be unrealistic. It's fine. But also, don't start from zero. Don't lie. Don't make stuff up. But how many did we do last week? Right? That could be a good baseline to say, "This is how many we've done in a week. How many can we get done in the test day?" And give the feedback. Give the progress. Where's the progress bar? Where's the bot that says, "Hey, this many bugs have been closed so far this day." We don't have any of that. And it would make these things far more effective if we did. So that's design claim number two. Yes, people will have time and availability, but you also need to set them up for success, right? You need to be able to say this is the deadline, this is the goal, this is the progress. Okay, so I am running way ahead of time, but that's fine because uh I mean this this happens all the time. This goes back to the the fundraisers and the teleathons and things, right? You've always got that that thermometer thing going on that's raising as they get more money and so on. Um, it works. It really does work and we should be using it. We can be good hackers. It's an hour session. It's not 10 minutes. Okay. So, number three, enforcing conduct scales with behavior frequency. What do I mean by that? So this is my example of something that is counterintuitive. And I definitely got this wrong when I was running u previous communities. Um let me let me explain what I mean here. If something bad is happening, we should do more things about it if it's getting worse is basically the statement here that the more prevalent the behavior, the harder we should come down on it is is basically the the myth here. And that's counterintuitive, right? You absolutely feel like you should be harsher about things. And I'll explain why. Okay, so first of all, I need to defend something important. Codes of conduct really matter. Now, this is not as controversial statement as it once was a few years ago. I think we've largely dealt with that and Fedora certainly has has had uh has that sorted. But it is important because it turns out that if you if you look at justice theory, if you look at these studies, it matters. Research on on fairness suggests that online community members are more satisfied with moderation decisions if they know how they got there. if it wasn't just completely arbitrary, right? And it's even better if they had a chance to join in with creating that code of conduct and those rules and those processes. Right? Now, that's difficult and sometimes it takes the time of moderators or even authorities. It's not always practical, but largely if you can build that code of conduct together, you're going to get more complaints with it. And that might head off this whole discussion that we're about to have. So, that's that's important to know first. But I said the target of this myth was about bad behavior. So what do we do about bad behavior? How do we have enforce a code of conduct? So to to answer this one, I need to talk about descriptive norms. Um descriptive norms are the things we pick up from around us, right? We learn these behaviors. How many people have joined a chat room and said nothing for two days while they see how the room gets used? Yeah, loads of hands. We've all This is why I like history basically in Slack, Matrix, Discord, whatever. The fact you have history skips that whole two-day waiting period, right? who you can look. Um, that is the process of discovering the norms of the room, right? That is what you're doing subconsciously. You're figuring out how people use it so that you don't stand out. We're back to social proof again, right? You don't want to use the room wrong. Um, and so these descriptive norms desri describe behavior. However, they can also describe the behavior that's not happening. Right? So, a good example of that is certainly in the UK where I live, people speeding on the motorway. Right? The descriptive norm there is that people everybody does it and and that's a difficult thing to combat right but that is what you learn as a driver is that everybody is doing this. So what does that mean here? What it means is that if you can learn norms in either direction, if you can learn the good norms and the bad norms, if you draw attention to norms, you can accidentally reinforce the other one. So if it's super prevalent behavior and you make a big deal about it, then what you actually end up doing in some cases is drawing attention to the fact that everybody is doing this and that makes the problem worse. And we have we there has actually been some studies on this. So this one's from 1991 and what they did is super interesting. They took um two different patches of ground and they did a control in each and what they did was they looked at littering. So you have a nice area, nice park. It's clean, it's pretty, it's got nice flowers, borders, benches, and you've got a patch that's considerably less nice and covered in litter. They paid actors to do littering in front of the public, just throw stuff out. And then they looked at what happened to the public afterwards and to see what their rates of littering were once they'd been kind of their attention had been drawn to the norms essentially. Right? So in the clean area, you can see that it actually goes down. So um the model litters or the model doesn't. The the plot is terrible to be honest. You need to go read the paper. But essentially what happened is in the in the in the nice area drawing attention to the fact that there is no litter here by doing the littering and people look and go why are you doing that in such a nice area. Don't do that. Littering went down for the public and conversely in the n in the disused wasteland area it went up because you're drawing attention to the fact that everybody does it. What does that mean for us? be careful, right? If it's a new behavior and you can stop it early. This happened with Anible when I was setting up the Ansible forum. We had a very excellent community member posting memes in the new help channel that had like three posts in it. And I'm like, I don't want the norms to be set that memes are acceptable in the help channel. So, I immediately made a big deal of moving that to the social section of the forum. These days, I wouldn't because the norm is already there. You can already see that the help channel is not full of that kind of content and that's not where we do this, right? We have a place for that. But if it's prevalent, maybe don't do that, right? And there's a really good example from um from MIT actually. So this um this is a a story from uh from Greg Jackson in 1994 talking about the user um the user account policy which was basically like this. They would send a message to to a problem user saying you or somebody using your account has been seen doing behavior they shouldn't. If that was not you, please come to us and we will help you reset your account. And if it was you, please don't do it again. And what was interesting was in a large number of cases, the people that came to them were known to be the problem people and they still came and got their password reset. It gave them a way out. And what was interesting that the theory goes that what was interesting here is it gave them a way to save face because quite often if you call people on their behavior, they will double down on it. They will do it again because they think I am allowed to do this and you are making my life hard. So, I'm going to do it again just to prove that I can. All right. Now, I'm not saying that you need to always handle that kind of stuff because if they're breaking the code of conduct in a series where you have to deal with it, right? I am not advocating that we should let people off lightly. But if they're in that gray area that is often the case with misdemeanors and small small breakages, maybe give them an out. Give them a way to reset their their expectations. but make it clear to them that they are breaking those norms, that they are not allowed to do that, but you can do it privately and that works. So, okay. So, that one's the counterintuitive one that I definitely got wrong because I I did in previous communities come down pretty hard on some behaviors and it probably didn't work as well as I would have liked it to. So, one to think about. The last one is an example of something where you simply don't get everything. You have to choose. So, I I I said at the start we should be intentional about creating our communities. This is a good example of where you can have X or you can have Y, but you can't have both. And it's about new users. New users are annoying. They don't know anything. Uh they have no idea what they're doing. They make terrible, terrible change requests. Uh and they are unfortunately critical to our projects. So, we have to do something with them, right? But in a lot of projects, we have initiation rituals and we have things like this. And to be clear, I have heard in person, not just online, people say thing all of these things at various times in my history. I disagree with all of them, but turns out I don't entirely get to be right about it, which is annoying. So, I've definitely heard all of these. I I the IRC one annoys me the most, but anyway, um these things are barriers to entry and we create them. Why do we create them? Turns out there's a whole set of theories around this around cognitive dissonance. Um the idea basically here is that we think we're rational beings. We're really not. Um, but we think we are. And so if we went through a hard initiation process, if we had to deal with a whole bunch of stuff, think about your your fraternity initiation rituals and things like this, we think that that means because we're rational beings, we went through something horrible. That means it must have value. Otherwise, we wouldn't have done that, right? And that's it's again it's subconscious, but that's the process. Basically, it's the it's the realignment between the pain we've suffered for the thing we got. Therefore, it must have more value. And again we've got studies on this. Um so so this is uh this is the statement here. People people come to like the things they suffered for because they want to be rational people right. Um now it turns out there were studies done on this. There's uh community where they had to do tagging in order to get their rights. So um some people had to do huge amount of tagging, some people were asked to do a small amount. Some people asked to do no tagging at all. Um, and it turns out the people who did more tagging um, did do more work thereafter. They valued that thing more and so they came back to the site more often after their initiation was done. So in other words, they are more engaged with the thing having done the initiation. So those barriers to entry aren't entirely for nothing, it turns out. Um, now that has consequences, right? If you put barriers in the way of people, there are going to be consequences. You are going to have less people in your community. It's just the way it is. That's a choice you have to make. You cannot have everything. You cannot do both. Now, I personally think that with Fedora, we want a wide contributor base because if you've got a good community team that can take that and we go back right back to the beginning about putting the right tasks in front of the right people, look at the social proof problem, give them work to do, give them things to work on, we have so much that we need to do, we have the community team there to help direct that, that's going to work. But if you've got a new project and you don't have a strong community team that's ready to start directing people, maybe a few barriers to entry will help keep your applicant cool to the point where they're going to be engaged thereafter. Maybe it's I I personally am very much on the side of let's not have barriers to entry, but it research suggests that you can get something from it, right? So, okay, I have gone really fast, but you know what? It's the end of the day. I'm not going to not going to drag this out for you. Um there is tons we can talk about in this area, right? I have not talked anything about motivation and how intrinsic versus exttrinsic motivation works. I have not talked um about uh oh god so many things. Um but I will give you some some reading material. Um if you like any of this obviously you can come talk to me some more. Um most of my content here actually comes from the first link uh building successful online communities. It's a fantastic if dense book but it's a literature review book basically. So all these papers I'm listing, yes I actually did go and read the papers. But the point here is that you can start with that and then spider off into other areas um uh and and start thinking about other things. It's a really really really good read. Um social architecture from Peter Hingens um dearly missed. He he passed away from cancer a couple years ago. Um but he was he was very good in the zeroq community. That book I would say it's really short if you're just looking for a short read. It's it's very very good. The first two chapters are pure gold. They were what I gave to the whole responsible community team to read because they needed to. Um the last two chapters little extreme but uh but worth a read and um architectural participation for the Red Haters in the room if you go and read that p that's a paper and not a book um if you go and read that you can see the history of Red Hat written in that paper without it being mentioned once because it's a it's an abstract paper about modularity versus monolithic uh work uh and the and the price of participation and uh and how community competes with with commercial entities. But you you can absolutely see um the history of Red Hat in there. Okay, I am not going to hold you here. Uh I think we should all go and like have more coffee and dinner and things. So um so let's let's wrap it up. Right, the these are the slides. If you want them, that's me. If you want me, that's my matrix.org account. Anyway, um because I forgot to change the QR code, I will be glad to discuss any more of this with you um over a beverage of some kind, but in the meantime, I will stop it there. So thank you very much. Any question? >> Well, yeah, thanks. That that's a pretty obvious question, I think. But what would you advise the federal project to do in the short, medium, or long term? >> Um, >> more precisely, >> that's true. We do still have half an hour. So, I think >> so I kind of called it out earlier, right? Um I don't think we have enough visibility into how we're doing with our with our overall strategy. So that go that calls back to the goal setting. It calls back to the progress. Um like we we have an ambitious strategy. Good. But how are we doing? I don't know. And I don't know do any of us know. I don't know that we've even decided how to count it. So um that is one thing I would absolutely look at. Uh and I would I would look at um I think the new contributor stuff. Is it possibly not what I was talking about here so much? Because we are an established project. We know what we're doing. We have a good community team. But you can always do better with newcomers. Um there is a there is a whole whole thing you can talk about with newcomers. You could do a whole talk on that because there's bless you. Um there is a there's a whole set of dynamics around newcomers because they represent a and again this is all subconscious, right? But they are a threat in a sense, right? They bring disorder to the project. they challenge old ways of doing things. Um this is to some degree present in diversity stuff as well which is important and matters but just from a psychology point of view we have to recognize that that brings in its own challenges right. Um so how you manage that how you bring people in without um completely disrupting things is an interesting problem to solve and we could look at we could always do more in that area. Um what else would I say? Motivation is an interesting one again that we didn't talk too much about, but intrinsic versus exttrinsic motivation is a super interesting topic. Um, knowing why people are in your community and what they are trying to get out of it. It's something I suspect we don't have a great handle on. The survey is open now, by the way. Go fill it in. Tell us why you're interested. But, uh, but it is a hard problem to solve, right? Uh, we know from studies that people value feedback from people like them the most. So this is for example why bots in GitHub repos don't achieve a great deal because people know that's not a human. But even like when you're at school and the teacher is praising your work, you know it's because they want more work from you, right? So that doesn't hold as much value either. So how how we how we give respect, how we give feedback, um these are all important things as well. >> New contributors Anyone else? Or should we wrap it soon? All right. Stunned you all into silence. All right, then. In that case, thanks very much. Let's go get dinner.