Dr. Scott McGrath 0:00 Okay, everybody hear me all right? 0:02 Excellent, fantastic. 0:03 My name is Scott McGrath. 0:04 I'm at UC Berkeley, and I'm just going to dive in because I got a lot of stuff I was going to cover today. 0:10 I'm going to— my background is in biomedical informatics, which is the intersection of computer science and medicine, and I want to tell you a story about our community building effort to bring informaticians over to Bluesky that ended up stalling out. 0:31 And it's not necessarily because we had a uniquely bad approach to it, but it's because I think the things that I encountered in the group that we tried to do this with will probably be similar if you tried our path too. 0:44 So it's like if you were trying to get your specific community to kind of adopt Bluesky, you might run into some of the roadblocks we did. 0:51 For some context, my professional society is the American Medical Informatics Association, or AMIA, and we had a pretty big science Twitter profile. 1:03 And so it was a hub that we would use and, you know, we had— this isn't the exact live board, but we had a live Twitter board at our conferences that you could see people engaging with. 1:14 And it was a nice central hub. 1:17 For us. 1:19 So when Musk bought Twitter, you know, the world changed. 1:23 And it didn't just trigger a migration with Amya. 1:27 We actually saw a fracturing of science Twitter at large. 1:30 There was actually a paper called "The Failed Migration of Academic Twitter" by Wang. 1:35 And they actually looked at 7,500 academics who were attempting to move to Mastodon in 2022. 1:44 And within a year, they actually saw that actual 7,500 drop down to 2,400. 1:53 And so that's only about 35% were actually still there. 1:56 So they did a little digging into it. 1:58 They did some network analysis, and they found that actually, if they'd gone to a kind of a specific domain, for example, like infosec.exchange, you actually saw higher retention, so about 57% were still active in that. 2:12 But people that kind of moved over to a general purpose server, that was more difficult for them to stick around. 2:19 And so that's— you saw the lower value there. 2:23 This kind of shows that there really is a community identity at the server or feed level, that those experiencing the kind of general feed usually did not stick around. 2:33 And so from Amy's perspective, we went from having kind of one conversation to a dozen different group chats. 2:41 And that's not a migration, that's a diaspora. 2:45 And so it was harder to see what was going on. 2:48 So what did we do about it? 2:49 Well, the signals for Blue Sky were quite strong. 2:53 There was that Nature paper that showed that 70% of the people that it polled were happy with Blue Sky, that they were over there. 3:00 There was also this Schiffman paper that surveyed 800 scientists about their habits and they had said that the tools that they had found on Twitter that were kind of helping with their career, it was actually easier to get going on Bluesky and they were enjoying it more. 3:16 So we kind of took those data points and said like, okay, there is evidence that we can get people to come over. 3:21 And we built out our migration plan. 3:23 We published a guide on our site about what Bluesky is and how to use it. 3:28 We assembled some starter packs. 3:31 We did a custom feed, and we actually engaged with Medsky as a collaborator because it lined up well with our community. 3:42 And we pushed this at our annual symposium, so our big meeting of the year. 3:45 We kind of rolled it out. 3:46 We were trying to get people to do, and it didn't quite take root like we hoped it would. 3:52 So where did it go wrong? 3:54 We can actually take a look at a preprint that came out this year that gives us some potential insight. 3:59 To that. 4:01 In this paper, they looked at 276,000 academic Twitter accounts, and they compared that to Bluesky at the time. 4:08 So it was only about 16,000 members at Bluesky at that time. 4:12 But they had found that only about 18% of those accounts had moved over to Bluesky. 4:18 And there was a different amount of adoption rates. 4:20 So like arts and humanities had around 30%, but health and medicine was only around 13%. 4:26 So that was one of the challenges that we had. 4:29 And then there was also another paper by Machado who found that there was an actual engagement gap that was occurring. 4:36 So Nature papers that were being cited on X had about, you know, 150 account mentions, but they were seeing like only about 5 on Bluesky. 4:46 And so if you got into the specific journals, it actually got worse. 4:50 And I'm not really showing— talking this— to be disparaging. 4:53 You know, I think we should be honest about kind of what— where we are in this kind of migration, then instead of saying like, this is where we'd hope it would be. 5:03 So one of the core problems that we kind of run into that we're trying to solve here is the cold start problem. 5:09 And that's not a technology issue. 5:11 This is a market structure problem. 5:13 And so the scientific social media really is a two-sided market. 5:18 We need both the creators that are the researchers who are posting the work and the consumers, those that are actually engaging with it, sharing with it, discussing it. 5:29 And to displace an incumbent platform, you kind of have to have them both over there. 5:33 And so we've seen this effect in the past. 5:36 I don't know how many of you recall back in 2010, there was the big Digg migration in which they had put out a new platform, people were really upset, and they moved over to Reddit. 5:47 And we've seen this also fail multiple times because there have been a ton of Reddit alternatives that have failed to take off, the latest being Digg itself when it tried to relaunch lately. 5:58 So it's a very difficult thing to do. 6:02 But we have some data we can look at that, you know, in Quell's paper they talked about the adoption to kind of go over to Bluesky kind of went by peers. 6:12 And so you would see one person posting and that would actually encourage you to sign up. 6:16 It would give you about a week to try and get in. 6:19 You might be interested, but once that closed, you kind of lost your interest. 6:24 And so we do know that kind of external shocks would give us a way to kind of see a surge. 6:31 Brazil's X suspension got us kind of in there, and then the US election also drove members. 6:38 We saw it kind of dip too. 6:40 These people did not always stick around. 6:43 So here's the thing that I think we really need to pay attention to is that the academics who reconnected their Twitter accounts on day 1 of Blue Sky, so they reconnected with the people they knew, stayed for about 12% longer, posted about 10% more, and followed about 25% more accounts. 7:03 And so this was a really great early network reconstruction piece. 7:08 And this kind of loops back to starter packs, which they're more than a gimmick, but they can be that tool that can get people up and running to find the people, to see the things that they're actually interested in. 7:21 So it's not entering into a room of strangers and kind of seeing like, well, what's going on here? 7:26 And I know they've kind of fallen a little bit out of fashion. 7:28 I don't see as many people kind of posting them or sharing them, but I think it is a tool that we should dust off and really engage with. 7:36 For our AMIR effort, I don't think that we really had the right tools deployed. 7:40 I mean, we had an idea, but just encouraging people to show up without kind of making it a welcoming community kind of set us back a little bit. 7:49 Another piece I want to talk about real quickly here is that there is a multi-homing problem in that— in that other paper, they found that, you know, only 19% had transitioned their accounts off of Twitter to Bluesky. 8:03 So people were actually staying on both platforms. 8:06 And we get into a problem when we kind of set it as an ultimatum, like you need to get off of X, you know, like, stop it, come over here. 8:16 Not everybody's ready to do that commitment. 8:18 And so it's okay to have the dual accounts. 8:23 Particularly because Blue Sky really lets us show off the benefits. 8:28 So let them show up, engage, and let a network effect come up. 8:32 You know, we were all talking about the amazing stuff we can build. 8:35 That will build momentum. 8:37 So we really want to see people kind of build the daily habit. 8:40 I'm going to go quickly through the things I would go through differently. 8:43 First, we want to anchor everything around a major event. 8:49 And so like we, the idea of kind of saying like, okay, there's going to be a conference, we should try and do something around that. 8:55 And it's better to have, you know, no post and kind of like things being quiet and then a surge than kind of like throwing it into the void. 9:05 Just having some activity that nobody's paying attention to. 9:09 Second, we really wanna kind of get involved in that one-day onboarding more so than kind of the zero-day marketing. 9:16 Like you need to come over here, you need to, like why you're still there. 9:20 And having the actual content there for them to engage with. 9:24 So the feeds, the starter packs, set the stage for success. 9:29 And that kind of ties into that last point about having discipline-specific identities. 9:33 We are having some amazing things like AstroSci, MedSci. 9:36 We are building the communities. 9:38 We just need to kind of invite people in to engage and keep them there, because then we can kind of show off all the bells and whistles that we have come to enjoy ourselves. 9:49 And so with the kind of last final point, I think we've built an awesome house. 9:54 We just need to learn to throw a little better party. Speaker B 10:03 Thank you, Scott. 10:05 Other questions to Scott? Dr. Scott McGrath 10:09 Um, I'm standing in the way of coffee, so I understand. 10:14 It's fine. Speaker B 10:15 I must say, I really like the analogy with the, like, having a presence on different platforms. 10:23 Yeah. Dr. Scott McGrath 10:24 Okay. Speaker C 10:28 Yeah, this is awesome. 10:29 Thank you. 10:30 I mean, we're going through this whole question now about how do you build community and, like, which platforms do you work on? 10:36 And I'm like, we can't just choose one, unfortunately. 10:38 I think we need to be on several of them because folks, like you say, there's a diaspora. 10:43 And I really like your comment about, you know, it's one thing to say to go somewhere, but you actually have to make it worthwhile for people to be there. 10:52 That's really, really valuable. 10:54 But it's also so hard because now you've got to commit to building the thing over there before people are there. 11:00 And so I guess that's my question is like, how did you convince your team, right, to commit to building this thing over there, or is that just a learning piece? 11:09 Like, how do we do that? 11:11 You know, it's, it's one thing to have a marketing campaign, and it's another thing to have a, a usable, uh, playground for people, you know what I mean? 11:18 So how do we do that? Dr. Scott McGrath 11:20 I, I think what we encountered that, um, made it difficult, it was that there is an exhaustion, um, for a lot of scientific Twitter that I spent so long building it up and it was going and then this unknown meteor came in and wiped it out. 11:39 And it wasn't that, okay, well, people left off of it. 11:42 It was something you didn't have control over. 11:44 And so even kind of pitching the idea, oh, but you can take your followers with you, that's a step too far for where a lot of people were. 11:52 They're just like, I don't know, man. 11:55 I don't think I have it in me. Speaker B 11:58 Yeah. Speaker D 12:02 Great talk. 12:02 Thank you. 12:04 I've also been thinking about the multi-homing and the cold start thing a lot. 12:08 I was wondering if you did any analysis into things like the Sky Follower Bridge and tools for moving over and then backfilling your blue sky feed with all your old tweets and this kind of stuff. 12:24 My general sense, not having looked into it too deeply, is that that tooling is super helpful, but it's just not quite there in terms of like super easy UX to onboard. 12:34 And if that tooling was just really, really slick and it was just kind of press one button and it moves all your stuff over and allows you to multi-home, then maybe we'd have a much better shot at this. Dr. Scott McGrath 12:45 Yeah, we did put that— I talked about the kind of the guide we built up and like that was in there like yeah, you can find and that's been combative, like they would try to do it and Twitter would cut back and like so it wasn't as easy as you'd hope that like okay, you just give people an easy on-ramp. 13:05 But I think that's kind of where I was pointing to the starter packs is that if you have the community and that was the balance too, we were trying to find people that were actually active and not just storefronts. 13:19 Like, oh, your favorite person's got an account. 13:22 They don't post, but maybe. 13:25 And that's where trying to get a fertile ground available for them to engage with, because that was the problem. 13:32 People would check in at different times, and you just couldn't hit that critical mass. Speaker B 13:39 OK. 13:41 One more, and then— Dr. Scott McGrath 13:42 Yeah. Speaker B 13:43 One more question and then— Speaker D 13:44 One more? Dr. Scott McGrath 13:44 Okay. Speaker E 13:45 I'll make it quick. 13:45 Thank you. 13:47 You'd mentioned using like Twitter leaderboards and stuff at conferences and things like that. 13:51 Do you think that there's been an effect of like a diminished networking effect of in-person communication to grow those follower networks since we're not doing as many in-person events maybe like post-pandemic? Dr. Scott McGrath 14:04 Yeah, I think you've got a good point there. 14:07 I think there was an atrophy in the muscle that we kind of had grown to. 14:11 And so people didn't automatically— since we'd gone several conferences without it, I think people stopped checking it. 14:19 And so that was kind of like I was hoping. 14:21 Because I was just barely— you'd see people pop in like, hey, I'm checking this out. 14:25 And that'd be like the only post they'd do for the whole time. 14:28 And it's like, you were there. 14:28 And then you kind of like, oh, the party's not hopping yet. 14:33 So I think I'm just going to slink out. Speaker B 14:38 Okay, thank you, Scott, once again.