After spending some time yesterday looking at the newly-launched Path (and the tech/mainstream press coverage thereof), I found myself thinking a lot about small-group sharing, and photosharing in particular. There have been a handful of products getting attention recently that are positioning themselves in this space, whether it’s explicitly articulated or not. Facebook has also been gesturing in this direction with their recent Groups product, and with Lists before that.
When we were working on Radar (Radar.net was a mobile photosharing service for ‘your favorite people and no one else,’ which is sadly now in the place where missing socks and the light from burned-out lightbulbs go), we spent a lot of time thinking about small groups — specifically about things like privacy and context and conversations and how when you manage to get the mix just right, a little bit of magic happens. People’s attitudes and expectations towards social/sharing products have been skewed in an odd way by their experiences with large mostly-public platforms — to such a degree that when users encounter something that’s been designed with them in mind it can be a truly compelling experience. Even the phrase ‘small-groups’ is ill-fitting in this way — they’re only small in comparison to groups that are artificially large.
These are just some thoughts on what I think makes this such a difficult problem to get right.
Privacy is a really funny thing. I really believe that people do care about privacy. Recently a lot of services have interpreted that belief to mean ‘people care about privacy settings’ — which they don’t. Even worse is that when users do periodically cause a fuss about privacy, it’s often interpreted as ‘the privacy settings should be impressively complex.’ This would be funnier if it weren’t such a common translation.
Privacy has something in common with myriad other complex, important, difficult-to-define aspects of life: you know it when you see it. Put another way, you know it when it’s been violated. Privacy is less about secrets and more about control (not controls). Privacy and private are not the same thing. Similarly, ‘public’ is not the same as ‘publicized.’ Privacy becomes a problem for people when a product betrays their intentions.
Dave Morin had posted a while back that “the devil is in the defaults,” which I think is a very smart way to put it. I also think it’s true.
CONTEXT IS KING. CONVERSATION IS QUEEN.
Context is crucially important to what makes all these products tick. Context is why your Mom cares about a picture of your breakfast and no one else does. Context is the ‘inside’ of inside jokes. This means that context is about who gets to see a post, but also affects what kind of content users choose to post in the first place. Make the context for sharing clear, and intuitive/easy to change (if it’s changeable). Minimize the opportunity for users to make mistakes, and make actions ‘fixable’ if people do make them (they will).
Context is clearest when the user controls it. Path has an asymmetric ‘active sharing’ model – while this approach loses some sense of group/context ‘continuity’ (e.g. the people you share with are the people who share with you), I think it’s a really smart one inasmuch as it gives the user a sense of deliberate control with respect to their audience. Making sharing deliberate is a productive friction.
Incidentally, the way that Path has chosen to limit the size of users’ audience, even if slightly gimmicky, is completely brilliant and I’m really glad they built it into the product. Limiting the number of people you can communicate with makes attention a scarce and valuable resource — just like it is in real life. This is something I’ve heard several people talk about in terms of context but never actually put their money where their mouth is for fear of hampering growth. My guess is it doesn’t.
Conversation is something that was completely constitutive to what made Radar a successful product. Comments helped to establish context but they were really the lifeblood of activity on the service in their own right. Mobile photos were a means to an end — they were an easy medium through which to share everyday moments, as they happened, with the people who cared about them, and to capture the conversations that emerged around those moments. When we finally shut Radar down, I didn’t really miss any of the photos, but I missed the conversation a great deal. Still do.
Conversation these days usually means comments (but doesn’t have to) — making conversation easy and fun and a first-class citizen of the product is critical.
SMALL GROUP COMPLETENESS
One of the specific challenges of designing a product centered around small groups (especially one that skews private) is that it’s difficult for new users to understand the value proposition in isolation. Put simply: it’s only fun when your friends are doing it. How am I supposed to know how fun/valuable it is if I’m the only one at the party? This presents a really unique timing challenge since the perceived value of the service to a new user is directly related to ‘small group completeness’ (the percentage of the group that’s actively using the service, and connected, when the new user arrives). The trick is not getting biggest fastest but getting enough small groups ‘complete’ enough in a sufficiently short amount of time. Small network completeness creates value, not critical mass.
A related dynamic that’s similarly challenging for services of this type is discovery. Platforms like Twitter are really great for discovery because they’re public, which allows new users to discover interesting content/people without having made an explicit connection with the content owner beforehand. Both Facebook and Twitter have proven to be useful tools for ‘bootstrapping’ new users by siphoning in friend graphs — but if you’re designing a product to support communication between small private groups, showing users huge lists of weak-ties on the first run isn’t a very instructive experience. Even if you’re able to quickly generate friend suggestions from another source (Path’s use of the iPhone address book is a nice fit even if a little unnerving), with a private service new users still have to wait for their friends to accept requests before they can begin to explore the product in a meaningful way. Even if this process only takes a couple hours, it often makes for a cold, lonely, confusing ‘first run’ experience for new users. You know what they say about first impressions.
LONG TIME COMING
Really smart people have been trying to design products to facilitate this kind of sharing for a long time. Path is by no means the first – and by no means complete – but I think they’ve spent a lot of time thinking about the product long-term, and built a platform that will support emergent behavior and evolution. And they way they talk about context and sharing makes me optimistic that they have a smart understanding of all the moving parts.
Even when we were working on Radar this idea was by no means a ‘new’ one. The press from a few years ago sounds eerily like some of the press I’ve been reading in the past couple days, though this should come as no surprise. The pendulum of over-sharing still hasn’t swung back this direction with quite the rushing force countless articles reported it already had. I’m not sure why it should until someone builds something better. Here’s to hoping.
(This post was originally published here by Kevin Twohy)