Making your app social
Created Slideshare for sharing your presentation slides online. Launched one year ago.
Social design of Slideshare. Idea was that presentations are hard to share, co-founder noticed that pics at conferences went to Flickr, videos on YouTube, but nowhere to put presentations. Can favourite and tag presentation.
Users drive navigation, using tags and popularity. Can do “slidecasts” where you mash-up slides with audio.
What do people share? More variety than you’d expect. Mainly expect it to be talk slides because that’s what we geeks use slides for. But within first few hours people were uploading “church 2.0” stuff, poetry, all sorts of things. Simplest way to get multimedia stuff up on the web.
Object-based social network. Social networks have been more about linking to people, but now they are more object-based, where you have the object that’s mediating the interaction, and in some sense it’s a powerful way to have a social space, because people want to share things, have an individual motive.
Lessons about Slideshare: Forget the iPod. Apple do great things but not social. People do what they want with Slideshare, cannot be a control freak. Have to give up control. Watch people make connections, and watch what emerges. Om Malik said yesterday that there’s a lot you can do when you think about online office applications, and people are going into the editing. I think that’s going to take a long time because people are specific about how they want their docs to look. Slideshare just wanted to solve one problem really well, which is to share slides.
Good beginning for a social app is one problem, really well solved.
Then take this and embed it in a social context. People don’t just keep presentations just for themselves – they are inherently social documents that you pass on. To find the social context that it fits in, which means Slideshow must be embeddable and have comments, and be linkable. Events are a powerful way for gathering slideshows.
Different choices people might want to make, e.g. use Facebook. Slideshow is a social space, but also a widget you can embed elsewhere on the web. People think of the social web as public, but actually it’s more about small social circles. Content going from public to private, it’s not a binary choice, it’s a range of choices. Google never forgets, and there should be forgetting in search engines, but then you have an ‘only in this place’ choice where you won’t allow it to spread out of context, share with groups, with friends, and then ultimately just for yourself.
We have gigabytes of documents on our hard drives, and mostly it’s just for us. Need to think about where you want to place yourself in this continuum. Flickr allows people to shift back and forth between public and private, can set privacy for each photo. Del.icio.us does the same thing, can make bookmarks public and there’re real value in that default public setting. Gives a return to the people sharing, which convinces them to share stuff that they might not otherwise.
Introducing privacy to Slideshare. Giving people take control over their stuff.
Q: What about images that have people in them.
We don’t see that as much as photo sharing sites. But relationship between you and the people who took the photo should be able to mediate the situation if they take an embarrassing photo of you.
Q: How do you deal with the temporal issues around privacy, might want to share things now but not later.
If it’s not being indexed by search engines then you can share it and then remove it, so one option for privacy might be “don’t index it by search engines”.
When you think about social networks you architect them to be open, so when you try to put privacy in place it can be hard to think how to do it.
Levels of participation – 1/10/100 rule. 1 creator, 10 synthesisers, 100 consumers. So use all visitors to surface good content, and leverage their participation even if they are quite passive.
Enable social navigation. Have most viewed, most featured, tags, and ‘viral navigation’ – can see what your friends are doing, only applies to registered users, not casual browsers. Three types of navigation: tag-based, popularity-based, virality-based.
Lots of different metrics, most viewed, most commented, can allow content to rise to the top. There is powerpoint porn and people do upload it, and it always ends up in ‘most viewed’, and used to have that most viewed showing up on the front page. Now have good community flagging, and don’t show most viewed on front page anymore. But different metrics reveal different things, so most viewed, commented, favourited, all show different material.
Wisdom of Crowds, James Surowiecki. Talks about getting wise decisions from large numbers of people. Thought a lot about that when doing the system. One thing about this is that there’s no getting away from the users. they are always there, they will email you, call you, post on your blog, write posts about you on their blogs, and it’s great because it tells you this live stream of feedback that you’re constantly getting from your users. Never did usability testing for Slideshares because the users would tell them when there was a problem. They were very involved with the system. Sometimes it can get a bit much, where you don’t want to listen to the loudest users because they are sometimes just a minority and you then miss the needs of those who don’t speak out so much.
Came back form a conference once and put up her slides, and embedded on her blog and got lots of comments and blog posts, but no one really asked it how to embed it. So just launched Slideshare without really telling anyone. Then got Techcrunched, and hadn’t really shown it to many people before that.
Any kind of feedback you get before you launch is to hypotheticals, such as screen shots, so better to put it out there. The risk of failure was not that great, but just wanted to experiment and see what happened. If it didn’t work, it didn’t work. But it did, and now the whole team is dedicated to it. For people who have another revenue stream, that’s a great way to experiment with a product. Another reason for not trying to get feedback before launching because feedback very invidualistic. But you need to see how people interact with each other in the system, and you can’t get that from individual feedback responses.
Q: Did you aim it at particular people.
Mainly people like ourselves, but people used it in ways we didn’t imagine.
Q: What about a closed beta?
That takes a lot of energy to do a closed beta, to get the URL out to people. But better to just put it out there and the people who need it will find it. If you’re quite about it… there are different ways of releasing. For us, we were really bootstrapped, small team, didn’t really have connections, so it was more powerful to put it out there and then people find it. We were not closed but we were invite-only, and that was the biggest mistake we made, because the initial momentum is enough to carry you out there, and invite-only crimps that.
Believe in “launch first, refine later”, don’t do too much, don’t try to do everything, just enough that people can see what you’re trying to do, and then refine after that.
Q: We can’t put stuff out that people depend on if it’s not ready.
Yes, it depends, we are not doing anything that people depend on it. But spending a whole amount of time creating something when you don’t know what people will do… Maybe do a small launch?
Q: Well, we can’t do that, it takes more resource. We are providing corporate apps, and they need to work.
Yes, it’s contexts. But we will end up moving in a more corporate direction slowly, as people use it in that context.
Q: What happens if you release a buggy product and people use it and never come back.
Yes, that’s a problem. We didn’t build too much, but we did try to take the bugs out. We thought about scaling, but not too much, so made sure we got the initial burst and that the site wouldn’t go down. You do have to make sure that it’s not so buggy that people won’t use it.
Q: Observation, if you’re talking non-essential B2C then you’re right, but this doesn’t work for B2B or mission critical stuff.
Well, yes, but i think there are things to be learnt from this. It’s a so much more effective way of building products. If you have an idea, and you’re thinking something, you’re wrong, part of your idea is wrong, that’s just the way it is. So put it out there, find out what’s wrong, be agile. Metrics were a huge part of that, we have a shadow app, which is all the metrics we care about.