Are You Ready For Social Software?

[Originally published in Darwin, January 2005. I am reprinting because of a request from a reader that led me to search for this piece. Thank goodness someone reprinted in its entirety, because Darwin content has been offline for several years.]

Years ago, a logic professor beat it into my bony head that Sherlock Holmes had it all wrong when he consistently claimed to use deduction in solving his cases. It turns out he (or better, Arthur Conan Doyle) was using induction, which is, according to Webster’s, “the act or process of reasoning from a part to a whole, from particulars to generals, or from the individual to the universal.” In working from a paltry collection of clues to a full understanding of the actions and motives of the butler and his victim, Holmes/Doyle was, basically, developing a picture of the universe surrounding the crime from a few hints.

The same sort of confusion — the difference between induction and deduction — seems to be at work in the rapidly escalating debate about “social software:” its meaning, relevance and purpose.

What is Social Software?

People naturally tend to use software as a means to advance personal interests and to interact socially. As a result, the most broadminded consider the “cc:” line on e-mail the starting point of social software; others restrict the term a bit more. In fact, you may be tempted to ask, “what isn’t social software?”

I believe the phrase social software should be more helpful, and can distinguish software built around one or more of these premises:

Support for conversational interaction between individuals or groups — including real time and “slow time” conversation, like instant messaging and collaborative teamwork spaces, respectively. This is also supported by the interplay always going on in blogs, where one blogger riffs on something another has said, and a third jumps in with more commentary, and the next thing you know, 40 others chime in, and someone suggests creating a groupblog to pursue the theme, whatever it may be. A big freewheeling discussion, with snippets of the interaction spread all over the place.

Support for social feedback — which allows a group to rate the contributions of others, perhaps implicitly, leading to the creation of digital reputation. Digital reputation — also known as karma (from the Slashdot web community model) or whuffie (from Corey Doctorow’s science fiction novel, Down and Out In the Magic Kingdom) — will turn out to be an area of great importance. Consider the lengths that eBay sellers go to to maintain a good reputation.

Support for social networks — to explicitly create and manage a digital expression of people’s personal relationships, and to help them build new relationships. These usually involve some sort of “six degrees of separation” system. One example is the Friend Of A Friend (FOAF) proposed standard, an XML-based approach to define your interests, phone number, e-mail, and the degree and kind of relationships you have with others, including creating explicit links to their FOAF specifications (which, of course, refer to others’ FOAF definitions, and so on).

The heady interest in Web-based services like Ryze, Friendster, LinkedIn and others, which are explicitly social (or business) networking systems, is being driven by a growing awareness of the fluidity and flexibility of networking through the Internet.

Adina Levin, author of BookBlog, recently suggested that social software could be defined as “tools that depend more on social convention than on software features to facilitate interaction and collaboration.” But I think this stops short of what is going on: Social software allows us to create new social groupings and then new sorts of social conventions arise.

Kenneth Boulding, the economist, humanist and social scientist, once wrote: “We make our tools, and then they shape us.” That is what social software is doing. It is changing the way that we socialize.

So What’s The Big Deal?

On the other hand, social software has aroused the ire of some well-known cyber-culture vultures, such as blogger Dave Winer (the founder of RadioLand, a blog technology company), who recently opined:

Social Software? I’ve been in the software biz for 2.5 decades, so I’ve seen this kind of hype over and over. Take something that exists, give it a fancy new name, and then blast at reporters and analysts about it. Every time around the loop it works less well. In the ’80s, it worked very well. In the early 21st century, there aren’t enough analysts with credibility to make such a pig fly.

P2P was the last gasp. I remember getting breathless invitations to keynotes where this or that luminary was going to finally tell us what it is. In the end it wasn’t the technology that made a difference, but ironically, the people. Apparently the promoters of Social Software were listening.

It’s wrong. We don’t need this. Weblogs are about punching through the hype machine of idiot analysts and reporters who go for their BS. Social Software has existed for years. What’s the big news? A few people are looking for a pole to fly their flag on. Pfui!

I disagree with Dave (which isn’t unusual), as do others who think the term has legs (or wings). David Weinberger (Darwinmag.com’s Swift Kick columnist) has weighed in saying,

First, I consider social software actually to be emergent social software. That narrows the field to software that enables groups to form and organize themselves…. Second, it doesn’t much matter to me whether the software is new or old. I’m excited about the fact that that type of software is now being recognized (i.e., “hyped”) as important.

Social Software: Bottom-up

Social software is likely to come to mean the opposite of what groupware and other project- or organization-oriented collaboration tools were intended to be. Social software is based on supporting the desire of individuals to affiliate, their desire to be pulled into groups to achieve their personal goals. Contrast that with the groupware approach to things where people are placed into groups defined organizationally or functionally.

One good metaphor is worth a thousand words, so I suggest the following: Social software works bottom-up.

People sign up in the system (for example, by downloading an IM client and registering an ID there) and then they affiliate through personal choice and actions (I add you to my buddy list, and you decide to remove me from yours).

Traditional software approaches the relationship of people to groups from a top-down fashion. In the corporate setting, its hard to imagine a person existing without being specifically assigned membership to top-down groups: your team, your division, the budget committee and so on.

Over time, more sophisticated social software will exploit second and third order information from such affiliations — friends of friends; digital reputation based on level of interaction, rating schemes and the like. And this new software will support David Weinberger’s notion of enabling groups to form and self-organize rather than have structure or organization imposed.

Blogging is a good example of this dynamic, and perhaps is the primary irritant pushing us today to grope our way towards new terms and tools. The group interactions around blogging arise in many ways: authors post thoughts, others comment and still others add their opinions. Likewise, social software starts with individuals: People start with their own interests, biases and connections, and these become reflected in social relationships, from which a network of groups emerge from the interchange. And the blog developers add more features to blogs to support this group interaction.

A contemporary example is the blog concept of Trackback — a means to automatically post at your blog any comments made on other blogs regarding something you have written.

Traditional groupware puts the group, the organization or the project first, and individuals second. As a member of a Lotus Notes group, for example, you are provided specific access to specific sorts of information based on the administrator’s settings. It’s all about control. It’s deductive: enforcing the general conditions upon each specific individual. The individual is fractured into a number of unintegrated group personas. The fact that you are involved in other groups, that you have had a long history with others in the groups, etc., is secondary to the fixed purpose of the group, whatever that is.

Social software reflects the “juice” that arises from people’s personal interactions. It’s not about control, it’s about co-evolution: people in personal contact, interacting towards their own ends, influencing each other. But there isn’t a single clearly defined project, per se. It’s a sprawling, tentacled world, where social dealings are inductive, going from the individual, to a group, to many groups and, finally, to the universe. Or at least the itty-bitty universe of all people using the Internet.

Why Now?

There are hundreds of millions of people connected through the Internet, using all manner of media — real time/transient, slow time/persistent and the various hybrids — to form groups. Online business or personal network systems like Ryze, Friendster, Meetup and LinkedIn are exploding in use, often adding tens of thousands of new users every week, because they provide the key elements of social software: conversational interaction, social feedback leading to digital reputation and explicit representation of “equaintance,” as blogger Gary Turner styles digital relationships.

The answer to nearly all “why now?” questions is technology and money, and that is true here. The availability of low-cost, high bandwidth tools like blogs or systems like Ryze, when coupled with the critical mass of millions of self-motivated, gregarious and eager users of the Internet, means social software is certain to make it onto “the next big thing” list.

Investment groups are eager to find a successful business model in social software, and I am certain that there are many to be discovered in each of the three key areas that define social software.

Despite the wet blankets and the naysayers, we are witnessing the appearance of a new crop of inductive, bottom-up social software that lets individuals network in what may appear to be crude approximations of meatworld social systems, but which actually are a better way to form groups and work them.

Perhaps just as interesting as the way that social software is transforming group interaction, across different time zones or in the same room. Social software is destined to have a huge impact on how businesses get at their markets. So the essential elements of social software will be incorporated into more conventional software solutions, changing the way collaboration and communication is managed within and across businesses, and ultimately transforming how companies sell and interact with customers.

Unlinking From Social Networks: Part 2

My project to unlink myself from the dozens or more social networking apps I have registered with is gaining momentum, and a lot of heat. The back channel — where dozens of people have emailed me asking me what the hell I’m up to — has been four or five times as active as the public interchange here at Get Real and over at Operating Manual for Social Tools (a project now closing down).

Questions range from “why are you dropping out of social networks in general and LinkedIn in specific” to statements of support and agreement with my general comments. Here, in a nutshell, are my motivations:

• I have participated in the various public social networks only passively — responding to others requests to connect, and occasionally passing along a request to connect to some contact.
• Because of the investments I have made in existing modes of networking — particularly social media based networks — I have not spent any significant time trying to exploit the SNAs.
• I have had a couple of disquieting interactions with those trying to aggressively promote themselves, their products or services through SNAs, as recounted here and here. I don’t really want to be prey to that sort of predator.
• I have wound up getting dozens of requests each month in the various networks by people more than two degrees away trying to reach people more than two degrees away, where I have little social capital involved, and I uniformally have been turning down those requests. In essence, these are a form of spam, although one that is allowed by the ‘rules of engagement’ surrounding the SNAs.
• I am annoyed that the SNAs don’t provide opt out at every juncture: please don’t involve me in requests like this, please don’t allow this person to contact me. please don’t contact me ever. The services vary widely in this regard. I was able to drop out of LinkedIn within a 24 hour period, although it does require sending a message to customer support.
• And I have an abiding interest in the creation of an interoperable basis for social networks. My experience in the instant messagingworld — where we have several large public networks that do not interact easily — demonstrates the problems inherent in pushing ahead with a fragmented model, where several large players will grow without any obvious incentives for interoperability, although it may well be in the public interest. (See the recent story about SocialPhysics.org, as an alternative.)

I set up a poll at www.votations.com that has just about a 100 respondents. Although my poll is flawed (for example, the first two questions are really the same, stated slightly differently), I am still interested in the results.

• Of the respondents (which are primarily my contacts at LinkedIn), roughly one third are passive users, not initiating activities but just responding to requests from others.
• One third have considered dropping out, because of lack of acitivity or too many requests.
• 75%+ of respondents believe they have been socially spammed (“someone trying to use the SNA in a way that does not line up with my goals or profile”).
• Roughly one third state that SNAs “are lacking critical features” — a lot of missing features — that would make them usable.

There is a sizable group, perhaps even more than half who find SNAs beneficial.
The remainder have serious issues and questions. My read is that these technologies are immature, have a long way to go, and probably have not assumed the form that will in the long run be the ‘killer app’ for SNAs.

My bet is that a deep integration of an open platform for social networking that easily integrates with social media is the best bet for future success. I would appreciate any other pointers to SNA research or development in this area: that’s where I think the missing critical features lie.

I plan to rework the poll, and press on with my retreat from SNAs. Next are ZeroDegrees, Spoke, Orkut, Friendster, Tribe.net, and so on. More to follow.

The Ten Commandments Of Social Networking

[originally published at Get Real, 22 August 2004]

Clay reprises my recent comments about Multiply and its email invitations, and does a very good job of making my argument more clear than I did, I think.

Clay Shirky

Stowe, reading my earlier Multiply rant, responds saying Multiply isn’t spam, and says that we need a statement of purpose for social networks to adhere to.

I’m more pessimistic than he; I believe that Multiply join messages are spam. Now spam has the “I know it when I see it” problem, so to talk carefully about it requires a specified definition. Here’s mine — spam is unsolicited mail, sent without regard to the particular identity of the recipient, and outside the context of an existing relationship.

Anyone sending me mail because I am on a list I haven’t asked to be on; without having a reason to think that I, in particular, would want this mail; and without us already knowing one another, is spamming me. In particular, ads sent to me as a member of a category, no matter how targeted, count, in this definition, as spam. You could be advertising a new brand of gin specially brewed for Brooklyn-dwelling Python hackers who like bagpipe music and that mail would still be spam.

If you adopt this definition, even just for the sake of argument, it’s pretty clear that Multiply fails the first and second tests. I did not ask for mail from them, and they are not sending me mail because they know me — they simply have my address on a list furnished by my friends. […] I think where Stowe and I may disagree is in point #3: do I have an existing relationship with the sender of the mail?

This is, I admit, a judgement call, and to re-phrase what I think Stowe is saying, Multiply is operating in good faith as a proxy for its users. My friends have furnished my address to Multiply, and authorized the service to contact me on their behalf. Thus the incessant messages from Multiply should be thought of as coming from my friends, and not from Multiply itself.

I hope I have characterized Stowe’s view correctly; in any case, I think Multiply fails this test as well, because I think they are engaged in a new form of targeted marketing. Jon Lebkowsky’s farewell to Multiply message includes this observation: “…next thing you know, Multiply was spamming all my Orkut contacts with a brainless marketing letter supposedly written by yours truly, only I didn’t see it until someone said no, no way, and noted the cheerful Muzak inanity of the message sent in my name.”

Clay has exactly defined the boundary cases in the ethical quagmire we are struggling with here:

1. I have assumed that the individual adding me to their contact list at Multiply (or elsewhere) is actually an individual known to me, and therefore I would not be surprised at getting an email invitation from them. Alternatively, if the invitation is coming from a party “outside the context of an existing relationship” it *should* be considered spam. But such an activity would be spamming on the part of that party (individual or group), and not necessarily on the part of the service. For example, someone could join LinkedIn for the purpose of spamming, which would not be the fault of LinkedIn, per se.
2. If a SNA coopts the contact list of its users and sends unknown, uneditable, and unannounced email invitations or (even worse) unsolicited advertisements for its or other services, that should be considered spam. This is what seems to have happened in Jon Lebowsky’s case, when he used the Multiply feature to invite his Orkut contacts… or so he thought. (I found that at least one of the Orkut or Friendster invitation features was not working yesterday when I was fiddling at Multiply — maybe they are revamping while this debate rages?)

I am totally opposed to parties spamming through SNAs as in case #1, and just as opposed to SNAs that meet case #2. I stated that SNAs try to make legitimate invitation of known contacts by email easy, to increase the acceptability of use. Clay argues that social connectedness should come at a slower rate, at a higher cost:

I think the growth of Friendster, one user at a time, undermines this notion, but however hard it makes it, that is a good amount of hard. Getting rapid growth one user at a time is difficult because it is supposed to be difficult. Social systems are, by definition, inefficient, and attempts to make them high throughput end up destroying them.

This last comment can be interpreted almost as a condemnation of the teflon slick feel of social networking applications, across the board, and I think gets into the guts of the problem: when social networking applications are targeted toward supporting human scale (not mass database) social networking for appropriate (non spam) purposes within the context of existing social (not commercial) relationships, things are fine. When you stray outside of any of those modifiers, it’s immoral, wrong, and possibly illegal under the CAN Spam Act.

Finally, Clay doesn’t hold with my push for a code of ethics that all should accord with (along the lines of what Duncan Work at LinkedIn recently pushed in his “Bill of Rights”), arguing for a more Darwinian solution, where the malefactors will just die off. I don’t know; I think the idea has legs, so I am going to try to boil down a short list of “do’s and don’ts” for SNAs, and promulgate it as the Ten Commandments of SNAs.

For example, Clay suggests that every email invitation from an SNA should include an explicit and easily discovered opt-out button. I strongly agree. The SNAs may want to qualify it in various ways (opt-out only for invitations from the specific sender; for a specific period of time; or for all invitations, ever), but there should be a way to opt-out, both at the SNA’s website and in every email invitation or other communication.

The Ten Commandments of Social Networking Applications (Part 1):

1. Social networking applications shall provide explicit and easily used opt-out features; specifically, every message sent by a social networking application on behalf of users, as marketing, or for whatever purpose shall provide a mechanism for complete opt-out, as well as a means to opt-out by email and at the SNA website.
2. SNAs shall not send messages to any user’s contacts without the explicit consent of the user, and without first displaying both the list of contacts to which the message is to be directed, as well as the complete content of the message.
3. SNAs shall not expose any user’s contact information or the information associated with the user’s contacts to anyone other than the user without the explicit permission of the user.
4. SNAs shall prohibit unsolicited commercial messages through their systems, and should bar or block users that try to send such messages.
5. SNAs shall provide means so that users can block messages from specific users.
6. SNAs shall provide users an “unlisted” capability, so that their use of the system can be undiscoverable if they wish.

Well, that’s a start. Other recommendations are cheerfully accepted.