Google UX Researcher Explains the Social Networking Gay Bar Problem | Hacker News Reader

Google UX Researcher Explains the Social Networking Gay Bar Problem

slideshare.net

267 pointslionhearted56 comments

>> If your privacy practices aren't transparent, then you introduce doubt. Doubt leads to lower usage.

So true. Too many apps and services don't understand that as users, when we see "Connect to Facebook/Twitter", our minds are immediately filled with doubt and mistrust. Why? Because we want control over what our friends see about us, and we suspect your app doesn't give a shit and will shamelessly promote itself at our expense. So be clear. If you're going to write on my wall, say you will. And if you're not, then say you won't. Don't leave me guessing.

But is this true of most users? Or only people like us? (I'm really asking, I'm not sure; all I know is that I don't like using Facebook Connect.)

I think its most users. Remember also that a lot of facebook users don't use it much and don't understand all of what it does.

http://foursquare.com/

http://www.kingdomsatwar.com/

http://vimeo.com/

http://tripadvisor.com

(just examples, on the top of my head, that are using Facebook Connect).

Some friends of mine had posted, on Facebook, the results of some online autism-spectrum awareness test they had taken—or rather, the test app had posted their results and a link to its own page. So I followed the link, and the first thing I got was a dialog box asking me to give the app permission, not just to post to my wall, but to access my private Facebook information. I declined.

I will assume that my attention to boundary management proves that I am not autistic. The jury is still out on the authors of that app.

"I will assume that my attention to boundary management proves that I am not autistic. The jury is still out on the authors of that app."

Maybe I'm being overly PC, but two things:

1. Lack of boundary management is more of an issue of antisocial personality disorder or narcissism, not autism as far as I know.

2. Don't use autism as an insult.

Correct, and correct. My apologies.

I mitigate this by doing my main browsing in Firefox but always doing Facebook in Chrome, in Incognito mode. No cookie cross-pollination.

I've always considered/wanted to do this, but the bother of having to actually type in my credentials every time and changing my habits has put me off. IIRC, Chrome won't remember my logins/passwords in incognito mode, correct?

Based on your comment I assume the hot new FF/Chrome plug-in for 2011 will sandbox personal information between sites.

You could use a Facebook Disconnect[1] to block all Facebook connect iframes from websites.

[1]: https://chrome.google.com/extensions/detail/ejpepffjfmamnamb...

I logged in once in regular mode, then I logged out, enough for Chrome to remember the credentials. When I go to Facebook in incognito I get the login page with credentials filled in, I have to click the login button but that's it.

I really enjoyed this presentation.

Aside from all the interesting ways of defining privacy I never thought of before, the implications of the presenter's goal (making everyone happy to use social networking for almost all of their life) struck me the hardest.

Let's say some radical changes came about so that almost any reasonable person would feel free to post all their personal data on facebook, knowing that sensitive parts would be kept to specific groups. That still wouldn't take care of what I see as the biggest lurking concern; how our data will be handled in the future.

It's certainly paranoia, but we've all seen good companies turn to shit, and people have even seen stable countries turn bad. If a database of everyone's un-self-censored private data existed, it would have untold implications. Even now, it's staggering to think about. I don't know, I guess it's getting late and I'm sleepy. But I think about FB's extremely high valuation and wonder if something like that could be a seed of instability that would allow for decision makers from unknown vectors to influence how FB uses its data and stuff like that.

I don't think it's paranoia at all. It doesn't even take bad motives, just a lapse of judgment or faulty assumption to ruin privacy -- like Google Buzz integration into GMail.

I think the solution requires more than feel-good policy statements and trust building. And God forbid someone mention legislation.

Ultimately the users need full localized control (hosting, or decrypting) over their private identities, content and social connections. Private groups should be darknets. And it should be as easy as automatic to set up and use. Zooko's Triangle be damned.

This presentation is somewhat long, but extremely informative. If you're designing any sort of application or content that includes sharing or privacy, I'd consider it almost required reading. It's that informative - the best I've seen on the topic.

Aside: Does anyone else have problems with left/right keys not going to the next slide as normally happens on slideshare? I got a bit annoyed having to use my mouse 223 times.

The entire UI was horrific for me. I had to use full screen since the slide didn't show up fully otherwise, and even in fullscreen Page up/down keys didn't work, arrow up/down keys moved terribly slow (like 1 page/minute slow), and the only way for me to navigate was to use the mouse scroll wheel (or the scrollbar on the right). I'd really like to know if I missed something or it was this way for everyone.

The flash slide viewer is awful.

Incredibly informative presentation. It helped visualize and conceptualize things I was already doing, like occasionally restricting my status updates to certain people: e.g. when I complain about my living situation and don't want those I'm living with to see it or when I hesitate to post a crude joke because I don't want my mom to see it.

I would suspect that to build something that comes close to matching our offline social networks you would have to start from the ground up, but I also think Facebook is in a position to offer some of the benefits.

One thing I would love to see? When I write a status update, post a photo or note, share a link, etc - I would like Facebook to tell me how many can see it. Just as LinkedIN can tell me how many people are in my network, Facebook could easily tell me that this photo marked "friends only" will be visible to my 128 "friends", but this status update marked "friends of friends" will be visible by ~15,000 people. Similarly, in Debbie's case, when viewing other people's photos, Facebook could tell her: your friends can see this photo and offer her tools to disallow that. Just because the Gay Bar employees allowed "friends of friends" to see their photos and I'm friend, doesn't immediately mean I want my friends to see them or see that I commented on them. I think Facebook could tackle these scenarios given the structure they already have.

Has anyone made an effort yet to do the evident thing, and simply build a social service that lets you have separate groups of friends for all features?

This is the main user-visible feature of Diaspora.

The problem is that nobody wants to present their users with a gigantic list of contacts on first login and ask them to sort and organise them into groups. Instead, everybody just lumps all the contacts together so that the user is least inconvenienced at first, only to later experience problems. Google is in a pretty good position to tackle this, since a fair number of people who already have Google contacts probably also have their contacts sorted into groups for email, and that's a good representation of their real-life social network.

I don't sort my contacts into groups, and I really doubt whether 'a fair amount of' other people actually do.

Distinguishing between for example work and personal contacts seems pretty hard, I'd say even for google.

You sort your contacts in gmail? I honestly would never bother, really don't see the point.

I very much doubt Google are, but Facebook might be in a good position to figure this out themselves now anyway. If I'm friends with Em, Bob, Barry and Jill, and Em and Bob are friends and Barry and Jill are friends, the friendship groups are fairly obvious aren't they?

Obviously I've never seen their data so don't know, but I wouldn't be surprised that they do have the information. Just looking at 'Mutual Friends' of people I know on Facebook it seems to be fairly well defined.

http://en.wikipedia.org/wiki/Clique_(graph_theory)

Facebook goes further than the mutual friendships and takes into account all sorts of interactions like "Like"'s, comments, etc. and probably re-sharing of contents, common events, common pictures, common places... I've been told they use that to figure out what to show you in your feed, so pretty much to know whom of your FB friends you're actually interested in.

But I'd be very surprised if they don't use that in a "group" perspective as well. Which would be very interesting because groups evolve over time but you going back to clean up your lists doesn't happen very often, if ever.

From a privacy standpoint, it is a bit scary. But it's very cool from a tech/CS point-of-view.

Pretty much what ycombinator funded "the fridge" tries to achieve. The problem is the best tool is probably a hybrid between full segregation and full connectedness. Problem is that in this kind of territory your application becomes harder for the user who has to think more about who is going to see which piece of content.

Is it really that hard? Why not just let me check the box near "high school" and uncheck the box near "parents' friends"...

Anyone who isn't naive about the implications of a "like" or a comment or a photo in everyone else's news feed will think twice about whether to post at all, which I'd argue uses more cognitive energy than simply opting in the correct group of friends to receive the information.

I have to say that any application intended to allow people to manage, say, four intersecting sets of items of any sort is going to back rather hard to use, unless someone has come with some incredibly clever UI metaphors I don't know about.

Will there be four (or five) checkboxes next to every normal action? It would seem to complicate things by a few orders of magnitude. Remember adding elements is often a multiplicative rather than an additive affair - just adding a few extra checkboxes to a given page can take it from clear to "head-spinning".

Is it really that hard?

Actually it is. Think about how few people actually categorize people in their IM client, or bookmarks, or virtually any other service.

There are a couple of ways to try to deal with this:

1) Put the onus on the contactor to specify the relationship. This is what LinkedIn does.

2) Actually have several different mechanisms that users can use. I really like free form tagging. I'd like to be able to send status and then have a tag field where I just list the tags of the various groups I want to send this status update to.

The oddly nice thing about how Facebook does it today though is that since everyone is so used to getting random weird status updates, no one cares. Imagine if you got an email from a friend that said, "Eating pie, watching Glee. Bored, might go to sleep early." You'd be like, "WFT? Is this guy gonna kill himself or something? Why did he send me this?" But on Facebook, you don't even notice. Heck you might even say, "Yeah, Glee is kind of weak in season 2".

I don't know. How people interact on the internet is weird.

There are sites that addresses this by creating completely independent social networks for separate interests. This seems to actually work for certain sensitive topics (look up "FetLife" for an example).

We're working on exactly that issue for our product (Windsoc). We're working on algorithms that attempt to create social circles, while providing an interface that should make manual adjustment and categorization easy. When someone posts using our service, they'll be able to select specific circles of contacts, or even individuals, to Facebook or other services.

I've noticed that people have started to give up on the idea of contact management across services, which is usually a good indication that it's the perfect time to make it happen.

(If anyone is curious, we'll be starting private beta in December.)

There used to be a tcp/ip stack called winsock. Thanks for the memory.

The d was silent for me when I read that sentence.

Come for the "gay bar", stay for the slides

For anyone else who can't stand SlideShare, here it is in Google Docs: https://docs.google.com/fileview?id=1vcckzZhBmWzGSalZV_yaJcQ...

That's why I keep three separate social networks: one on facebook (friends), one on Twitter (colleagues and strangers) and one on LinkedIn (professional contacts).

It's hard to keep them separated though, especially since now I feed my twitter stream to the other two, seriously considering to remove the feed.

It's also interesting to think about how to use social networks/networking if you're the owner of a store or service that people may not want to promote across all of the networks.

eg: Debbie probably wouldn't have any trouble recommending a new burger joint she found (except if she was telling everyone she was on a diet, perhaps). But now that Debbie has realised her contributions to social networks leak out further than her intended audience, it's going to be really really hard to convince her to 'Like' that gay bar were her friends work.

As someone involved with a few different 'gay bars' (ie, services people like using, but in a lot of cases don't want all(/any!) of their friends/family to know about) it's an interesting problem... And I have no idea what the solution is.

Debbie's problem is exactly what makes me use the Friend Lists feature of Facebook. It might not protect your data that well, but at least it is able to keep things semi-private from the (less tech-savvy) friends of yours from other social circles.

But, of course, Facebook screws you in the end. Even you choose exactly what friends see that photo of yours, and what friends see that wall post, you happen to comment on a friend's photo and then everybody sees it.

And it also does nothing about friends of yours from some circle deliberately accessing the profiles of the people from other circle and seeing everything you didn't want them to see (like tagged photos of yours that they couldn't see on your profile because you forbidden them to).

It's also surprisingly difficult and annoying to add and switch people around to Friend Lists.

Strange. I figured that Facebook handled this by default. If I make a comment on Friend A's status or post, and Friend B is my friend but not a friend of A, I assumed that Friend B could see neither A's post or my comment. That's a pretty big hole...

It depends on the privacy setting for the status/post.

The good news of this presentation is that vast improvements in the social network space are possible, improvements that might just be able to tumble the incumbent. The hard part is finding them.

Could the multiple-kinds-of-contacts issue be handled if social-networking sites explicitly allowed one user to have multiple handles? E.g., a user could be “Bess Smith” to her family members, “B. J. Smith” to her college friends, and “Elizabeth Smith Jones” to her clients, and only her five closest friends would have “friend”-level access to all three identities.

Great presentation but his claims about IM are not solid IMO.

IM is personal messaging i.e. we communicate directly with people in the way that's appropriate. Nothing get's broadcasted out. If anything IM is a perfect list.

I have friends, groups of friends, colleagues, clients etc. on my list but they don't bleed into each other.

Also in most IM you can decide to be invisible for selected people and not for others.

This talk raises some questions for me.

Email doesn't seem to have the kinds of grouping problems discussed in the presentation. So what is it really that isn't in email that can then be solved by something Facebook-like that doesn't do "social networking" like Facebook does it? Is it just that the interface is simpler?

email = status update, social group = mailing list, contacts db = friends network

what else?

photo sharing. video sharing. email is suboptimal for this. if you don't want to see all photos in full resolution, you still pay the response cost.

Passive observation. I'm vaguely interested in what my acquaintances are up to, but they would never impose so much as to email me.

I think email falls into the same category as the real life social graph. You communicate seperately with each group of friends. Other groups don't see your comments. The swimming kids don't see you reply to the friends from the gay bar talking about pictures.

Say, you want to let people know that you are going to be in their city. Email will be too intrusive. So indirect communication is advantage (which social group allows). A contact in the city can see it and choose to ignore it. Another one, may jump at the opportunity, and welcome you and want to meet you.

But we are still left with the problem of broadcast to all, which the PPT articulates very nicely.

This is four months old:

http://news.ycombinator.com/item?id=1479761

So you are surprised that a title like, "Google UX Researcher Explains the Social Networking Gay Bar Problem" gets more votes/views than one with this title, "The Real Life Social Network v2"? And four months ago - wasn't that around the time of the Facebook movie w/ a similar title?

I like how pseudonymity works: you go by one or more pseudonyms, on various web sites. It lets you have multiple identities, and figuring out which identities are the same person is non-trivial. It's an easy hack. You can post your erotic furry artwork under one identity, use another professionally, another with your friends, and everything will be more or less okay.

Alternately, maybe people will get used to lack of privacy, and acknowledge that everyone has more than one group of acquaintances, and that information on the web is permanent even if people themselves change. You'll know this is happening when someone finds naked pictures of a female teacher on the Internet and she doesn't get fired. I'm not holding my breath, though.

I think he is simply wrong that web is only going to get more and more social.

I would posit that an opposite movement will also happen. Social networks become more ubiquitous but that very ubiquity will teach people what not to share and raise the value of the "anonymous web".

For example, if a merchant can understand their customers exactly, it introduces an information asymmetry into the market game that the merchant enjoys but I, as customer dislike. I don't want a merchant to know exactly what I'd be willing to pay for X. I want the merchant offering prices anonymously and then for me to be able to pick the best of those prices.

Social media will revolutionize everything, says man working in social media.

This is effectively solved by the new Facebook groups.

Oh, it's the presentation I read 4 months ago.

Facebook Groups and Friends Lists are a step in the right direction but they haven't been implemented well in other aspects of Facebook.

If Google Me ever became a reality (and not in the limited "only existing product enhancements" fashion) this would be an interesting discussion. Facebook's model isn't friendly to the privacy oriented sharing schemes.

Reply on news.ycombinator.com