This set of posts is great, and I'd love to hear any thoughts you all have on scaling up past the initial group of users.

I started a branch a few weeks ago on a similar topic: branch.com

What do you think some modern strategies for scaling up a "community" to millions of users would look like? Who has been successful in the past and who has failed, and what can we learn from those examples?

How do you prevent that initial group from fleeing when the rest of the internet shows up at the front door?

One way is to let people create their own communities and “silos” within your platform. If people can find like-minded friends while still also feeling they’re contributing something unique and worthwhile individually, then you’d be on the right track.

A big part, I think, is the communication and messaging of your product/platform/site. Be it Content Strategy or just plainly the nature of the copy, there are few ways to write inclusively and countless to write, well, not so inclusively. If your messaging is confusing, unhelpful or patronizing to inexperienced users, the resulting culture will reflect that, no matter how much more welcoming and awesome the early adopters are.

And of course, the general usability of the product matters.

Can you think of examples of large products that have done this well?

Twitter's follow model seems to have worked out well for them, but there's still a really high barrier for new people joining the service.

YouTube is a good example of tons of really small micro communities around each of their popular members and even around some videos, but as soon as a video gets popular the comments quality usually falls through the floor.

Path and Instagram spring to mind, allowing you to add friends from other services who are on their service as well. That considerably lowers the barrier to entry and helps smooth the signup process.

A List Apart started as a listserve, evolved into a magazine, and spawned a successful book publisher and conference, as well as a highly rated agency. It did that through good old fashioned editorial skills, plus advocacy and attention to a particular community. Boots on the (virtual) ground.

Creative Mornings started out as a little conference in NY and has since blossomed to 34 cities and just this afternoon raised $40K from a now-loyal community. kickstarter.com

These are smaller examples than your Facebooks and YouTubes. But don't underestimate the value that attention can have. Simply caring about the people that use your service, and spending lots of time with them, is the most important thing you can do.

Also, I really don't buy the "the initial group flees" claim. People didn't leave Myspace because the unwashed masses showed up. They left because there was something better.

Right and in that situation it's not a case of fleeing, it's a case of moving on. That's reasonable.

But then I see a service like Twitter, still monumentally relevant, still at a high level of quality despite any developer qualms, and while there are people leaving it for a variety of good reasons, the one that stuck out (and this is in the App.net global stream in several places) was that Twitter had become uncool because it was filled with people talking about Justin Bieber and such.

Despite perfectly adequate filtering controls on Twitter, I have seen this complaint raised numerous times and it is indicative of this elitist demographic concerned not with what works well, but what is 'cool'.

Now, perhaps I'm seeing evidence where there is none to suit my narrative, but it is a little distressing to see a community partially born out of that kind of sentiment.

Anyone who would leave Twitter because other people are using it to talk about Justin Bieber really needs to stop taking themselves so seriously.

Also, the few people reportedly leaving Twitter seem to be doing an awful lot of talking about how they are leaving Twitter *on Twitter*. I call bullshit.

Agreed. Chances are good this is just like when Facebook changes their design and people complain on Facebook, saying they will leave Facebook. Nobody ever does, really.

Just as I think Twitter will likely continue as it has up until now.

My interest basically lies in understanding the early adopter segment and what causes them to leave a service. It's likely what I've seen so far is only a vocal minority, I freely admit that.

A lot of the people on App.net are developers who are excited to build something without limitations, and that's inspiring.

That's the kind of experience a new service should have to draw people in. Something you can't wait to join because you can do amazing things with it.

That's why services like Branch, Medium and App.net are sweeping the web right now. Not really because everyone is mad at Twitter, but because these are amazing tools taking shape right before our eyes, and we get to be a part of it.

> Anyone who would leave Twitter because other people are using it to talk about Justin Bieber really needs to stop taking themselves so seriously.

Indeed. Since you control your stream, you are directly responsible for what you see.

The same could be applied to Facebook, but they make it so hard to un-friend people and instead seem to be relying on machine learning to show you more relevant status updates instead.

What if Twitter recommended people for you to un-follow, or even, suggested you replace 1 user with another? Is the problem that these services make it too easy to add people, therefore cluttering your content stream and then make it really hard to un-clutter it?

I was just browsing Hacker News, and they appear to be discussing exactly this topic...

Paul Graham posted a reply to a user's post about how bad things have gotten, and that is followed by many suggestions, both technical and sociological: Worth a peruse if you are into this kind of thing:

Cultivating communities well requires moderation and effort. Creating high-profile, critical-mass targeting startups requires growing your user base as fast as possible and getting as many people involved as possible. These two are opposites, which is why we see this phenomenon happen so frequently.

I think the key to achieving both, i.e. having your cake and eating it too, is to make the product itself ensure the quality of the content _and_ the community of users. Think for instance of a signup process that tells you the rules of the site before letting you in. Yes, it's a (small) barrier, so it may slow growth, but it could make up for that by having people be a lot more respectful and conscious about how they use the service.