Google Groups Pain in the Neck [en]

[fr] Google Groups trouve qu'il n'est pas raisonnable de vouloir ajouter plus d'une dizaine de personnes à la fois à une newsletter nouvellement créée.

I’ve used Google Groups to set up a newsletter for Going Solo.

Here it is, with added proof (if needed) of my hopeless lack of design sense.

When I set up the group, I did what most normal newsletter creators would do: went through my contacts to invite those who might be interested in joining. I selected 30 or so people to start with.

My action triggered a flag for review, as I might be a potential spammer:

Your request to invite X new members has been flagged for review by our staff.

In order to protect our members from unsolicited email, Google manually reviews invite requests which meet various criteria. Your request will not be reviewed unless you provide us with more information in the form below. Reviews generally take 1 – 2 business days.

Please provide an explanation for where these new members come from and why they would want to be part of your group. Note that Google takes a very dim view of Spam. The people you invite must know you and be expecting your message. If they complain, you will be banned from our service and your group will be deleted.

Great.

Well, I wrote up an explanation, saying I was setting up this newsletter so that people could stay informed about Going Solo (registration is closing soon btw), and that I was going through my address book to let people know about it.

Anything wrong with that, in your opinion? I think not, and Google obviously didn’t think there was anything wrong either, because they let my invitations go through after a few hours.

BUT.

Now, each time I invite even one single person, my request is flagged.

Google Groups: Threatening!

What a pain! I’m going to be inviting people many times a day over the next week, as I dig out e-mail addresses. And obviously, just announcing the existence of the newsletter is not enough to get people to sign up — ever heard of lower the barrier to entry? If I’m creating this newsletter, it’s because I’m finally coming to my senses (!) and realising that not everybody follows Twitter, subscribes to blogs, hangs out on Facebook or upcoming, and that good ol’ e-mail still has some good days before it when it comes to getting information out to people.

I am really annoyed at Google Groups for making this so difficult. Shouldn’t there be a way for me to get the limit “lifted” for my group, by offering proof I’m not a nasty spammer, but a businesswoman (OMG!) who is very much aware that she will very quickly use up her social capital if she spams her network with irrelevant stuff? And therefore, that I actually need to send out invites to a few hundred people?

Also, look at this form:

Google Groups invite members

Don’t you think that “e-mail addresses” field invites a reasonably large number of addresses?

I went through the help, and it wasn’t very encouraging, but I did learn a few useful things:

So, please. If you have friends working on Google Groups, please draw their attention to this post and issue. It’s a bloody pain in the neck.

Oh yeah — and please sign up for the newsletter. I’m going to have trouble inviting you 😉 — [email protected] also works.

Groups, Groupings, and Taming My Buddy List. And Twitter. [en]

[fr] Long, long billet sur la notion de "groupe" en social software et les différentes formes que peut prendre cette notion. Trop raide pour traduire ou résumer, navrée.

Warning: very long post. Not proof-read. Hope it makes sense. Mostly dictated, so if you see funky stuff that isn’t a typo and really looks weird, try reading out loud.

“Group” is a word which is thrown around a lot in the social software/social tools/social networking/social thingy arena. Flickr has groups. Google has groups. So does Yahoo!, of course. CoComment is working on groups (and have been for ages). Twitter is being advised against them (I second that). YouTube, Facebook, Orkut, Last.fm — “groups” seem to be a compulsory feature for any 2.0 service today. It’s very natural, too: we need to break down large communities in order to be able to function within them (see The Dunbar Number as a Limit to Group Sizes for some thinking around this issue). Unfortunately, it’s also a result of all the 2.0 “community” buzz stuff floating around: “implement groups, and your tool/app will have communities!”

Like many overused words, “group” is actually used in different contexts to mean different things, and this brings about quite a lot of confusion. “How to implement groups” is a theme that I’ve had a few exchanges about with both the coComment and the Twitter people, and I think it’s an impossible question to answer unless we have cleared up the vocabulary a little to start with.

I would like to distinguish between three types of “groups”, which are often all called “groups”, but which have different characteristics and different uses:

  • “groups” or “shared-interest groups” (“Flickr-groups”)
  • “groupings” (“ad hoc assemblages of people with similar interests” — Stowe Boyd)
  • “contact groups” (organising my contacts)

Shared-Interest Groups

This is usually what people think of when they say “group”. It is a set of people who come together to (hopefully) form a community around a shared interest. Usually, one chooses to join such groups. Belonging to the group gives you some kind of special connection to other members (which you might not know, but you now have one thing in common with), and allows you to “do things” you would not be able to do if you were outside the group. (For exemple: send a message to all the people in the group, or post a photo to a shared album.)

Typical examples of this kind of group are Yahoo! Groups or Flickr Groups. People join these groups to be able to build something, share something, or simply hang out with the other members of the group. However, if you look at the way people use this kind of group in communities which are more “social networking”-oriented, like Facebook or Orkut, you will see that they tend to not be that active inside the groups, but that they use them a bit like “tags” to advertise their interests. These groups are therefore not only a way of connecting with other people, but also a way of saying something about yourself. And in some communities, the latter is clearly more important.

Groupings

Shared-interest groups are a bit limited when it comes to making your application truly “social”, as I heard Stowe Boyd point out during his Building Social Applications Workshop at the LIFT conference earlier this year. Now, I’ve been through Stowe’s blog to try to serve you with a nice citation that explains exactly what he means by “groupings”, and haven’t really found anything that satisfied me. (As far as I can see, Stowe first talks about groupings in In The Time Of “Me First”: IBM Slowr?, and explains a bit more in In The Time Of “Me-First”: Stikkit.)

Here’s the definition Stowe gives in his workshop slideshow, slide 24:

Groupings: ad hoc assemblages of people with similar interests.

Stowe Boyd

As I understand it, groupings are things that “happen” rather than things that people elect to join or build. Groupings emerge within a social network because of the way people are using it. Groupings are things that occur naturally and all the time inside networks, but the tricky part will be to decide which groupings to make visible to the users and how.

The first time I really encountered this type of automatic grouping of users based on their behaviour was in Last.fm. Last.fm tells you who your “neighbours” are, by picking out people who have similar music-listening habits as yours. So, in last FM, not only can you see my contacts or “friends”, people I have elected to be connected to in some way in the online world of last FM even though our musical tastes may have little in common, but you can also see my neighbours, people I probably do not know and definitely have not chosen to be connected to, but which I am inevitably connected to because we share similar musical tastes.

Isn’t this a more interesting way of interconnecting people than having them explicitly join groups saying “I like this or that artist”? CoComment also has a neighbours feature (I like to think that I’m for something in its existence, as it was one of the first suggestions I made and pushed for about a year ago), but unfortunately you can’t see other people’s neighbours or do much with your neighbourhood. The value groupings will add to your tool or service will depend greatly on which groupings you decide to make visible to your users, what doors being part of a given grouping opens up for the user, basically, what you choose to do with these groupings (display them? Nice, but not enough in most cases).

With all this in mind, if you are trying to figure out “the best way to implement groups” for your application/tool /2.0 service, here is what I would recommend. Start by taking a long hard look at how your application already organises users into possible groupings. What can you make visible? What is interesting? What doors could you open to people who are inside the same grouping? What are your users going to want to do with these groupings?

Some examples of groupings could be:

  • people who have listened to a particular song regularly over the last six months
  • people who favourite my photographs on Flickr
  • people who subscribe to a given blog
  • people who have commented on a given post or blog
  • people who have marked me as a contact
  • people who use a given tag
  • people who comment on posts or photographs tagged “cat”
  • people who ordered this or that book on Amazon
  • people who have been marked as a contact by somebody
  • people who have joined a certain group…

As you can see, the definition of “grouping” is much wider than the definition of “group”. “Groups” are a small subset of “groupings”, which have a performative flavour, as you become part of them by the simple act of stating that you desire to be part of them.

The example before last is a little bit problematic in my sense. Most of the time, a user ends up belonging to a grouping because of the way he or she uses the system. It is your actions which make you part of a grouping. Here, you are not part of a grouping because of something you have done, but because of what somebody else has done to you (added you to her contacts). I have been hesitant for this reason to consider “being somebody’s contact” as a grouping, but if you look at it from the point of view of the social network, it is still a way in which “usage” organisers to people who are part of the network.

The existence of these “passive groupings” (from the point of view of the user who is part of the grouping) invites us to go through the looking-glass and examine what goes on from the perspective of the user creating the groupings by making his connection to other users explicit.

Contact Groups

I hope that we have now come to accept that networks are asymmetrical. It is not because I have marked you as a contact, that you have to mark me back as a contact too. I think that a great source of confusion is the general use of the word “friend” in social networks. There is an emotional component in there that makes it rather difficult to say “well, you might think I’m your friend, but I don’t.” Friendship is supposed to go both ways. “Contact” is a much more neutral word, which is easily understood as meaning “you are, in some way, part of my world here.”

“In what way?” is the big question here. In what way is John part of my world? In what way am I part of his, if at all? I will leave the second of these two questions completely aside in this discussion, for I consider it to be a psychological, emotional, and relational minefield. In our offline relationships, we don’t usually get to know exactly how important we are for our friends or acquaintances, or even love interests. We are treading on eggs, here. And to make things even more delicate, different people use different words to describe the people who are part of their world. These are, in my opinion, human relational issues which are way too delicate to be formalised in a social network without a lot of serious thinking, if they are to be respectful of people’s feelings and meaningful in any way.

The first question, however, is a crucial one. I personally think that it is also the key to managing many privacy issues intelligently. How do I organise the people in my world? Well, of course, it’s fuzzy, shifting, changing. But if I look at my IM buddy list, I might notice that I have classified the people on it to some point: I might have “close friends”, “co-workers”, “blog friends”, “offline friends”, “IRC friends”, “girlfriends”, “ex-clients”, “boring stalkers”, “other people”, “tech support”… I might not want to make public which groups my buddies belong to, or worse, let them know (especially if I’ve put them in “boring stalkers” or “tech support” and suspect that they might have placed me in “best friends” or “love interests”… yes, human relationships can be complicated…)

Flickr offers a half-baked version of this. I say “half-baked” because it does allow me to introduce some organisation in my contacts, but it is not quite satisfying. And regarding what has been said above, this classification is made public — so inevitably, there is no way that it can be satisfying to the person making the classification. It has to remain politically correct. Basically, what Flickr does is allow you to single out certain contacts as “friends” or “family”. This is tame enough, particularly given that the word “friend” has been emptied of much of its meaning by social networks which use it as a synonym for “contact”. What is interesting here is how Flickr uses this classification to help users manage privacy. I can make certain photographs visible only to my friends or my family. I can decide to allow only my contacts to comment. But this kind of control remains quite coarse, because the groups are predefined and may not map well to the way I view my social world and want to manage my privacy.

A more useful way to let a user organise his contacts is simply to let him tag them. Xing does that. Unfortunately, it does not allow one to do much with the contact groups thus defined, besides displaying contacts by tag, which is of course nice, but about as useful as making groupings visible without actually doing anything with them.

Use more precise vocabulary than “group”

Have you noticed how I’ve been using the word “groups” to speak of this way of classifying one’s contacts? Well, instant messaging software uses the word “group” (“buddy groups”, “contact groups”), and that’s what people are used to. Now, imagine the confusion if somebody says “Twitter needs groups”, meaning “contact groups”, and the person listening understands it as “shared-interest groups”? These are two very different kinds of groups. They are organised differently and serve a different purpose. See why I think we need to stop speaking about “groups” in general and be much more precise with our vocabulary?

  • Shared-interest groups are groupings that we actively choose to be part of, they are generally public, or at the least, we know who the other members are, and the point of being part of such a shared-interest group is to be able to do certain things with the other members, or get to know them.
  • Contact groups (normally) passive groupings that somebody puts us into, they are generally private, to the extent that one does not know exactly what grouping one is in, and the interest of such contact groups is mainly for the person creating them, who can choose to treat the people inside them differently (mainly regarding privacy).
  • Groupings, defined by Stowe Boyd as ad hoc assemblages of people with similar interests, can actually be understood as a very generic expression, including the two previous ones, to refer to “ad hoc assemblages of people emerging through social network/software/tool usage.” When it is one’s actions which bring him/her into a grouping, we can speak of “active groupings”, and when it is another’s actions, “passive groupings”.

One could probably say that the way in which a social application implements groupings (which are made visible and how, and which actions, features, permissions or characteristics are associated to them) — shared interest groups and contact groups being two particular species of groupings — is going to play an important role in how successful it is, because groupings in general are the key through which users will interact with each other.

Maybe somebody could start working on a taxonomy of sorts for groupings? We already have active and passive, the weird performative ones that are the similar-interest groups, all the contact group stuff, but we could imagine classifying and analysing groupings by looking at what brings one into a grouping: is it interaction of some type with other users? Quantity of something? Centred around one object, or a collection of objects? Is there a time component? Does it involve reciprocity? What kind of pattern of usage is it linked to? We could go on, and on…

Case-study: Twitter

Even though this post has been ripening in my head (ew!) since February, the reason I am writing it today is the following twitter from Tara Hunt:

Advising Twitter (Britt) AGAINST groups (gameable/spammable) and FOR personal lists (solves group messaging)

twitter from Tara

I have blogged about Twitter quite a few times already, spoken with the Twitter people when I was in San Francisco and sent them a bunch of feedback and ideas that I haven’t got around to blogging yet (I wonder when I will). This should make pretty obvious that I really really like this service. (So that’s the disclaimer: fangirl.)

If you’re still reading this, your head is probably full of groupings/similar-interest groups/contact groups ideas and concepts. Let’s see how they apply to Twitter. The nice thing about Twitter is that it’s a rather simple application, feature-wise (and that’s one of the things that makes it so nice). So, where are the groupings? Here are some:

  • users who are friends with John
  • users John is friends with (not the same grouping!)
  • users John is following (still another grouping, because of the distinction twitter makes between friends/contacts and the act of “following”)
  • users who are following John but he is not following (fans/stalkers, depending on how you look at it)
  • users who answer John’s twitters (with @John)
  • users who use the word “LIFT07” in their twitters

What makes Twitter great? Well, besides the great online/offline integration through the use of mobile phones, the clean, usable interface, the great people using it and the cats in the servers, one of the things that makes Twitter Twitter (if I may say) is what it does with the grouping “users John is friends with”. Well, it’s pretty simple, in fact, and you’ll probably think I’m pointing out the obvious (but that, in my mind, simply indicates how good a job Twitter have done with it): they display all the twitters of those users in that grouping on one page. Well, yeah, I guess that was the Obvious thing to do with that grouping.

Amongst the other types of groupings, one can wonder if Twitter needs to introduce similar-interest groups, or contact groups. I don’t see much of a case for the former, as Twitter is centred around people and relationships rather than the content of their interactions. Twitter is not really about what I’m saying to people. It’s about who I’m talking to. Twitter is precious because it gives me a space in which I can share a little things about my life with anybody who has decided that these little things had some value to them (and that can include non-Twitter users). Twitter it is equally precious because it provides me with a space (and this is where the “what they actually did with that grouping” thing comes in) through which I can stay informed of the little things in lives of others that I have decided were meaningful for me.

Which brings me to contact groups. Contact groups could have two purposes for twitter:
– privacy management
– twitter overflow management, particularly on mobile devices.

Without getting into the technicalities involved (and I’m aware they are not straightforward), let’s imagine that I can tag my Twitter contacts. This allows me to give some structure to my online world in Twitter. I can use that structure in two ways: make certain messages visible only to certain people I have chosen (privacy), receive messages on a given device only from certain people (overflow).

Tagging is the best way to create these contact groups. It leaves each user completely free to organise their world how they wish. It allows multiple classification of contacts. Keep the tags private, and personal dramas are avoided. Multiple classification requires establishing rules for when conflicting orders are given. Interfaces (web and mobile) need to be devised to tag contacts, to set message privacy (default, message by message, on/off style), and following behaviour. Not straightforward, of course, but can certainly be done.

Remains the basic question: does this kind of feature address a real need? (For me, it does.) How is it going to change Twitter if it is implemented? (If this can be predicted…) What might happen if it is not implemented? Well, you know, the usual stuff when making a decision.

Disturbed About Reactions to Kathy Sierra's Post [en]

[fr] Comme cela avait été le cas lors de l'affaire SarkoWeb3, la blosophère s'est maintenant emparée de la triste histoire des menaces reçues par Kathy Sierra, telle une meute affamée et sans cervelle. Hypothèses présentées pour faits, coupable car non prouvé innocents, noms, déformation d'information, téléphone arabe, réactions émotionnelles trop vite bloguées et sans penser... tout y est.

Encore une fois, je suis déçue des gens.

Since I read and posted about Kathy Sierra’s latest post, and stayed up until 3am looking at blog post after blog post pop up on Technorati and Google Blogsearch, I’ve been growing increasingly uneasy about what I was reading in the blogosphere.

Like many other people I suppose, I was hit with this “tell me it ain’t so” feeling (denial!) that makes one sick in the stomach upon reading that Kathy had cancelled her ETech appearance out of fear for her safety. My heart went out to her. Of course, I felt angry at the people who had cause her such fear, and I also felt quite a bit of concern at seeing known blogger names appear in the context of this ugly affair.

And then, of course, there was the matter of getting the word out there. I blogged it (and blogged it soon — I’ll be candid about this: I realised it was breaking news, heck, I even twittered it before Arrington did!), and although I did use words like “horrible” and “unacceptable” (which are pretty strong in my dictionary, if you are familiar with my blogging habits), I refrained from repeating the names mentioned in Kathy’s post or demanding that the culprits be lynched.

One of the reasons for this is that I had to re-read some parts of Kathy’s post a couple of times to be quite certain to what extent she was reporting these people to be involved. Upon first reading, I was just shocked, and stunned, and I knew I’d read some bits a bit fast. I also knew that I had Kathy’s side of the story here, and though I have no reasons to doubt her honesty, I know that reality, what really happened, usually lies somewhere in between the different accounts of a story one can gather from the various parties involved. So I took care not to point fingers, and not to name names in a situation I had no first-hand information about, to the point of not knowing any of the actors in it personally.

In doing this, and taking these precautions, I consider that I am trying to do my job as a responsible blogger.

Unfortunately, one quick look at most of the posts coming out of Technorati or Google Blogsearch shows (still now, over 15 hours after Kathy posted) a collection of knee-jerk reactions, side-taking, verbal lynching, and rising up to the defense of noble causes. There are inaccurate facts in blog posts, conjectures presented as fact, calls to arms of various types, and catchy, often misleading, headlines. I tend to despise the mainstream press increasingly for their use of manipulative headlines, but honestly, what I see some bloggers doing here is no better.

Welcome to the blogmob.

The blogmob is nothing new, of course. My first real encounter with the mob was in May 2001, when Kaycee Nicole Swenson died (or so it seemed) and somebody dared suggest she might not have existed. The mob was mainly on MetaFilter at that time, but there were very violent reactions towards the early proponents of the “hoax” hypothesis. Finally, it was demonstrated that Kaycee was indeed a hoax. This was also my first encounter with somebody who was sick and twisted enough to make up a fictional character, Kaycee, a cancer victim, and keep her alive online for over two years, mixing lies and reality to a point barely imaginable. I — and many others — fell for it.

Much more recently, I’ve seen the larger, proper blogmob at work in two episodes I had “first-hand knowledge” about. The first, after the LeWeb3-Sarkozy debacle, when bad judgement, unclear agendas, politics and clumsy communication came together and pissed off a non-trivial number of bloggers who were attending LeWeb3. There were angry posts, there were constructive ones and those which were less, and then the blogmob came in, with hundreds of bloggers who asked for Loïc’s head on a plate based on personal, second-hand accounts of what had happened, without digging a bit to try to get to the bottom of the story. Loïc had messed up, oh yes he had, but that didn’t justify painting him flat-out evil as the blogmob did. In Francophonia it got so bad that this episode and its aftermath was (in my analysis) the death stroke for comments on Loïc’s blog, and he decided to shut them down.

The second (and last episode I’ll recount here) is when the whole blogosphere went a-buzz about how Wikipedia was going to shut down three months from now. Words spoken at LIFT’07 went through many chinese whisper (UK) / Telephone (US) filters to turn into a rather dramatic announcement, which was then relayed by just about anybody who had a blog. Read about how the misinformation spread and what the facts were.

So, what’s happening right now? The first comments I read on Kathy’s post were reactions of shock, and expressions of support. Lots of them. Over the blogosphere, people were busy getting the news out there by relaying the information on their blogs. Some (like me) shared stories. As the hours went by, I began to see trends:

  • this is awful, shocking, unacceptable
  • the guilty must be punished
  • women are oppressed, unsafe
  • the blogosphere is becoming unsafe!

Where it gets disturbing, and where really, really, I’m disappointed and think bloggers should know better, is when I read headlines or statements like this (and I’m not going to link to all these but you’ll find them easily enough):

  • “Kathy Sierra v. Chris Locke”
  • “Kathy Sierra to Stop Blogging!”
  • “Kathy Sierra hate campaign”
  • throwing around names like “psychopath” and “terrorist” to describe the people involved
  • “Personally I am disgusted with myself for buying and recommending Chris Locke’s book…” and the like
  • the assumption that there is a unique person behind the various incidents Kathy describes
  • taking for fact that Chris Locke, Jeneane Sessum, Alan Herrell or Frank Paynter are involved, directly, and in an evil way (which is taking Kathy’s post a step further than what it actually says, for the least)

In my previous post, I’ve tried to link to blog posts which actually bring some added value. Most of the others are just helping the echo chamber echo louder, at this point. Kathy’s post is (understandably) a little emotional (whether it is by design as

I’d like to end this post with a recap of what I’ve understood so far. (“What I’ve understood” means that there might be mistakes here, but I’m giving an honest account of what I managed to piece together.) I’m working under the assumption that the people involved are giving honest accounts of their side of the story, and hoping that this will not unravel like the Kaycee story did to reveal the presence of a sick, twisted liar somewhere.

Please, Blogosphere. Keep your wits. This is a messy ugly story, and oversimplications will help nobody. Holding people guilty until proven innocent doesn’t either. (Trust me, I’ve been on the receiving end of unfounded accusations because somebody didn’t hear my side of the story, and it sucks.)

The problem with bullying is that perceived meanness isn’t the same on both sides. Often, to the bully, the act is “just harsh” or “not to be taken seriously” (to what extent that is really believed, or is some kind of twisted rationalisation is not clear to me). To the bullied, however, the threats are very real, even if they were not really intended so. Bullying is also a combination of small things which add up to being intolerable. People in groups also tend to behave quite differently than what they would taken isolately, the identity of the individual tending to dissolve into the group identity. Anonymity (I’ve blogged about this many times, try a search) encourages people to not take responsibility for what they say, and therefore gives them more freedom to be mean. Has something like this happened here?

If you have something thoughtful to say, then say it. But if all you have to say has already been said out there ten times, or if you won’t take the trouble to check your sources, read carefully, calm down before blogging, avoid over-generalisations, and thus avoid feeding the already bloated echo-chamber — just go out for a walk in the sun and let the people involved sort themselves out.

The word is out there, way enough, and I trust that we’ll get to the bottom of the story in time.

Update: I’m adding new links which actually add something to this story to my first post as I find them, so check over there for updates.

Group or Author Blogs? [en]

[fr] Conversation hier avec Suw au sujet de la meilleure approche pour initier les gens au blog dans un cadre professionnel. Contrairement à mon intuition, elle recommande d'éviter les blogs "communautaires" à moins que le thème en soit très clairement défini. Il vaudrait mieux donner à chacun son blog, si possible avec une période d'essai sur l'intranet pour repérer qui "capte" et qui ne "capte pas", afin d'encourager les futurs blogueurs à se sentir responsables du blog. A plusieurs, on tend à rencontrer le syndrome "les autres blogueront". Et vous? Qu'en pensez-vous? Des expériences à partager dans le domaine?

Interesting and thought-provoking conversation yesterday with Suw, about group blogs vs. author blogs to get people to start blogging. Group blog can work with newbie bloggers if they have clear focus (ie, we are going to get together to blog about things we’ve come upon in Lausanne).

If there is no clearly-defined topic, then it is better to get people started on their own blogs, so that they take responsability for it. Otherwise you get the “somebody else will post” syndrome that I’ve noticed on a couple of multi-author blogs I participate in (or try to direct). If necessary, make them start blogging on the intranet before going “public”.

Your experiences with newbie bloggers in more-or-less corporate environments? Tried group blogs? Prefer author blogs? Got theories?