Over the last several videos, we've looked at some principles and frameworks for social media, and we've talked about the power of being able to go beyond being there. What I'd like to do in this wrap up video is to be able to distill six principles for things that you can use to make your social media successful. And there's a lot of great books that have come out recently. One that I really like in particular that I recommend is Bob Kraut and colleague's books on building successful online communities. Some of the things that we're motivated to do are intrinsically motivating. I want to go swimming in the ocean, or go spend time with friends. Other things that we do are extrinsically motivating. I need to pay my taxes on time so somebody doesn't come after me. And many things are a combination of the two, like a lot of our professional lives. It's important to think about when you're designing social media, which one you're after, and when these things can compose, or when they can work against each other. One example of a social media system, which is really funny and kind of devious but remarkably effective, was a group of Stanford students in BJ Fogg's class in the fall of 2007 built a Facebook app. This was right when Facebook for the first time had apps that you could build on top of the platform. And their app let students send hotness to each other, which was nothing more than a simple messaging app that said something to the effect of hey, somebody has sent you hotness, or somebody thinks you're hot and to find out who, you need to have at least ten users on the platform. And so of course, everybody wants to know who thinks they're hot, they're curious, so they sign up ten friends to use the platform and then they'd email some of their friends, and some of their friends and some their friends. And this viral approach got them several million users in the span of a single university quarter. And it was so effective that it actually caused Facebook to change the rules about what an app could do on top of their platform. So this is one example, a cynical one, but nonetheless excellent, about how, when users are motivated, you can really drive usage on your platform. And Facebook has done a number of things that make it work so well. And another book that I recommend is Clara Shih on the Facebook Era. Three of the things that Facebook does that work extremely well are first, on Facebook, the people who are sending articles or other bits of information, are people you know. And so it's not just a weird crazy thing on the Internet, or a news article, or something else it's a friend of yours who is interested in that particular thing, and the fact that it's being socially recommended really drives a lot of the usage. Second, on Facebook, our app is socializing, and being curious, and finding out what's going on, and so a lot of the opportunities that we see come from all different spheres, and that's okay because that's in line with the task that we're doing. And there are these, the clicks that we have available to us are hot. They're right there. And that makes it really easy to be able to explore a number of different things. And in particular, one of the things that we see with social media is that, and our friends liking articles is an example of this, is it provides social proof that if my friend Hector recommends an article, that's a kind of socialproof that the article is good. And this evidence of other people's behavior, especially for our friends, but really for anybody, can be extremely effective. And one of my favorite examples of this is a study by Robert Cialdini's group where they were looking at hotel towel reuse. In hotels, they go through an enormous number of towels, mhich of course uses a lot of water and detergent and labor and other things. And so for environmental and other reasons reasons, it is beneficial to be able to limit the number of hotel towels that get used. And so what Cialdini did, is they tried to see which messaging approaches were most effective at getting people to shift their behavior. And one of the things they found to be most effective is this door sign right here that says please reuse the towels. And so, if we're asking you to reuse your towels because of the behavior of others, so most of the guests in our hotel reuse the towels. That actually has a demonstrable effect. If we give a standard environmental message, we get about 35% of people will reuse their towels. That might be something like, it's good for the environment to be able to reuse your towels, and there are people who pay attention to that. Now, if we say something as simple as, most of our hotel guests reuse the towels, then the rate of reuse goes up from 35%, that we see with the standard environmental message, to 45% with a social message. And remarkably, if you say 78% of the guests in your hotel room, room 312 or whatever it is, reuse their towels, then this goes up even higher. And so by being a member of a community, even if it's a nominal community like guests of a particular hotel room, you can drive behavior change through that social channel, much more effectively than simply stating that it's good behavior. On a social network like Facebook or Twitter, some of the people that we interact with are strong ties. Strong ties might be somebody like a family member, mother, father, sister, brother, there are best friends, there are husbands and wives and spouses and boyfriends and girlfriends, and other people that we interact with every day. Online, and in person too, we also have lots of weak ties, so a weak tie might be family friends, or colleagues who I know somewhat, but are not my best friends. So if I've got me here, I have a small number of strong ties. Almost by definition there can't be too many of these, you can't have 1,000 best friends. But I can have a much, much larger number of weak ties. So I have lots of colleagues. I was just at the ACM Chi conference this week in Seoul, Korea, and I know lots of people there. I've been going to Chi every year for 16 years, published papers there. I've written articles with people there. I have strong ties that I see there, good friends, and also lots of weak ties as well. Both our strong ties and our weak ties are useful for being able to do the things that we need to do in life. And this incite comes from sociologist Mark Granovetter who is now at Stanford, where he looked in particular at getting a job. What Granovetter found is that most people who got a job, got it not through their father or sibling or significant other, they got it through one of their weak ties. Maybe a high school classmate, or a friend of a friend, or somebody that their parents know or somebody like that, somebody that you used to work with at a previous job. These are our weak ties. And what you see is that our weak ties have a more different social networks than we do. My best friends are also good friends with each other. [LAUGH] My parents are also strong ties of each other and so because of the greater interconnectivity in our strong ties, there's less additional sensors out there, than there are with our weak ties. Our weak ties give this much more distributed antenna. That's the idea of a social network. And many of the social networks of today are built on the ideas of Granovetter and others of being to enable weak ties, whether it's for getting a job, or dating, or simply socializing. One of the most difficult things to do, as a platform designer of social media, is to make space for reinterpretation that works well. And I'd like to give you some examples of things that are important for people, and are an important part of social media, and others that can cause undesirable behavior. Telling the two apart is not always obvious. About a decade ago, an early social network was called Friendster, it was a predecessor to Facebook. And one thing that happened on Friendster is that, as people joined and added friends, people started creating Fakesters. And a Fakester is a not real person who's part of the social network. So it might be Batman. Or it could be an idea like justice. And then you can become friends with Batman or friends with justice. And this is funny and people like it. And many social networks, of which Friendster was one example, get squeamish when people add fake people to their social network. They say hey, you can only add real people, this person's fake, no you can't add them, and they started shutting down these accounts. This caused a lot of user backlash. Friendster even managed to make it worse actually by allowing Fakesters that were advertising supported. And getting advertising involved in social media can often be necessary, but delicate. So, when users saw that Fakesters were not allowed when they were just for fun, but then they later did get allowed when they were for advertising, that really turned a lot of people off. You see these things today, so this is an example I like on Twitter, this is Karl the Fog in San Francisco, where I lived for a decade. And Karl offers his thoughts on life being the fog of San Francisco. And people even tweet to Karl, so, for example somebody takes a picture and they send it to Karl the Fog. This is the kind of play that makes sense on social media. Other kinds of reinterpretation are more serious and can be extremely profound, both in terms of making the world a better place and in terms of really changing how social media work. One of my favorite examples of this is hashtags were not part of Twitter at the very beginning. It was a user-created convention that then really flourished and kind of became an integral part of the platform. And one thing that you can use the hashtags for is, for example, when the earthquake hit Haiti, several people were using hashtags as a way of coordinating disaster relief. So people on Twitter can provide information like hey, we need this resource here. Or for example, a hospital could say we have a certain number of beds available, using the Haiti hashtag. And several of the volun-Tweeters even acted as intermediaries, reading the streams, and connecting people who had resources and people who needed resources. And this metadata addition that happened over time where people figured out how to manage the flow of what was originally a pretty unstructured stream became important, and over time we've seen that structure evolve and integrate into the platform. Usage on social media really increases when every incremental bit of usage adds more than just that element, and I think of this as the ways that you can have one plus one equal three, where every time you add more information, you get even more benefit out. One example I like of this is the website MovieLens, it's a movie recommendation engine that I use all the time. And, with MovieLens you rate movies that you like, and it will then suggest movies that others who liked those same movies like. So, for example, if I like Fellini films, it may recommend, users who liked Fellini films also liked, and I can find new movies that way. And what's great is that when I add a piece of information, that helps every single other person on the social network, and it helps me more too. And so you get these network effects where the value of every single piece of information increases as the network gets larger. So both for others and for me, everybody's getting more benefit. And finally, the last thing we're gonna to talk about today is how can you make the behavior that you want people to do easy, and behavior that causes problems in your social network to be irrelevant or ignored? One trick for a lot of this is plausible deniability. My favorite example of plausible deniability is that when your phone rings, you don't always answer it, and of course it could be that you're sitting right there and the phone is ringing. You may even be looking at the caller ID to find out who's calling. Or it could be that you didn't hear it, that your phone was on silent, that you were away from your phone. Any of these things are, that you're busy, anything is possible, and what's nice about the plausible deniability is that we can graciously say, I didn't hear your call, I wasn't there, I was busy, and you can decide whether to take the call, and it doesn't always need to line up with exactly what had happened. One thing that's in the Bob Kraut book that I mentioned at the beginning that I think is really clever is that at MIT, when students were doing illegal or undesired behavior MIT would then send a notification to you the next time you logged in and say, hey, somebody has been using your account to do illegal things. And of course, MIT is pretty darn sure that that somebody is you, but having the distance that's added and the user has a little bit of plausible deniability, you are offering them an opportunity for them to say, oh, it wasn't me, and at the same time realizing, somebody's watching me, I better not do that anymore. Very cool. A second way that you can make good behavior easy and discourage bad behavior is the strategy of libertarian paternalism. This is an idea in behavioral economics, and I think it's best illustrated through retirement planning. It's really important to save for retirement. Few of us do as much of it as we should. A big part of retirement planning, in America at least, comes through employee plans. So your employer may automatically deduct a fraction of your paycheck and put that in a retirement savings account for you. It takes a while to get this set up, and it's hard to figure out how much you should do, and people are busy, and so, that's one reason why people save too little for retirement. The libertarian paternalistic approach is to figure out how much people should be saving, and have that be the default option. Traditionally, most employer retirement plans are opt-in. You have to manually set up the retirement savings. They could, by contrast, be opt-out, so that when you join, it subtracts a reasonable portion of your paycheck, saves that for retirement, life is good. You still have complete control. You can change it to be anything you want, but the default option is one that's probably good for you. But what about when somebody actively wants to behave badly? So libertarian paternalism works really well, when people want to do the right thing, but it's too hard, or we're too lazy, or we're too busy, or whatever. What about when people are actively behaving badly, for example trolling online. If you ban trolls immediately, they just create another account and start trolling again, and so actively responding to trolls just invites more trouble. With trolls, what often works the best is shadow banning. And that's where you take the troll's world, and they can still continue to post, only nobody other than they and potentially other trolls see that information. So, this shows up mostly on newsgroup posts and that sort of thing. So, I can post to a thread and none of the other users see it. Eventually, trolls may catch on, but it takes a lot longer and it's a much softer approach. So, in this way, people can still behave badly, but they're shouting into a vacuum. With Wikipedia, it also encourages good behavior and mitigates bad behavior when needed. By default, most articles are easy to edit. Some articles are contentious, and some for just a small period of time. So something flares up in the news, there may be differing viewpoints or people who want to troll, so what we can do is that for a contentious articles, we can lock them from being edited, just for a short period of time. And that way, most of the time, when things are going well, most of the articles, everything's editable, and you get the benefit of the wisdom of the crowds producing this wonderful encyclopedia. But in the key cases where an article is contentious, we can pull that back. These are six questions that you can ask yourself as you're designing a social network, or even engaging with existing platforms. And you can use these questions to help guide your social media strategy and be able to deliver value to users. Happy designing.
