Tuesday, October 27, 2009

Social search and advertising: Google's endgame?

A few weeks back, Jeremiah Owyang wrote a piece "Revealing Google's Stealth Social Network Play." In it he detailed the tactical benefits of a combined of Google Reader, Wave, and Sidewiki in a back-door strategy aimed at social networking. And more to the point, to realizing the advertising opportunities around social networking.

Google has been neither a leader nor a even a decent case study in social networking. It's home-grown social network, Orkut, is popular elsewhere but not here. Open Social is still very real, but is largely invisible to the public. And when it comes to making use of the social graph, Google profiles are a distant cousin to Facebook and even Linkedin profiles. Google's products seem to betray a distinct affinity for information over the more popular and user-friendly experiences that have resulted in the conversational turn in social networking: Facebook status and activity feeds, and twitter.

But with Google Wave, Jeremiah's observation looks spot on. Wave not only facilitates a potentially game-changing departure from old-school email, but also supports the export and re-embedding of "wavelets" outside the Wave experience. These wavelets function as apps, and some of the early extensions featured have already begun to spark interest among developers who see Wave as an application platform turbo-charged by access to Google search, contacts/address book, and distributability. If successful, Google Wave is poised to serve as a platform for distributed social networking.

Brynn Evans writes today about Google social search, in Why There's Nothing to Fear in Social Search. Social search may seem innocuous enough, and the video posted on the company's blog contains a not-so-subtle pitch for Google profiles (the more you related sites and services you add, the better Google can serve you!), but the flip side of an improved search experience is of course advertising. Namely, social advertising.

Now, this is a nut that many have failed to crack, try as they have. But Facebook's failed Beacon was a sign of things to come. There's money in the feed. Feed-based advertising, which I liken to product placement in mainstream media, promises (for now) to leverage the rich social context and realtime conversational power of activity feeds and twitter. Now that twitter has offloaded its advertising problem to Microsoft's Bing and Google, it can worry about making twitter a richer experience, while delegating advertising to the search engines. But reconstructing the conversation, as Adina Levin notes in her post Search the conversation, and as many of the semantic, sentiment, and influence relevance companies I've spoken with will attest, is all the more difficult the shorter the message and the thinner the relationship.

Which is possibly where Wave might create more than a ripple for Google's alogorithmagicians and data miners. Google has lacked access to the information that can be extracted from mined social actions. Wavelets, embedded on end-user and brand blogs, sites, and elsewhere (eg phones, participating social networks), could be used to create an index of social activity. For wave interactions are captured by Google (which hosts the original wavelet and sees all interactions that occur on it).

A social action index built on the back end of Wave could be combined with search indexes of conversational messages from twitter (and possibly other activity feeds: Myspace, Facebook?). Add to those, indexing of blog comments and sidewiki, Google reader subscriptions and its comments, likes, and shares, plus the rich social graph information provided by Google contacts, and you have what looks to me like a distributed, decentralized, gold-mine of search queries, documents, conversations, relationships, and activities. All built on an advertising platform.

If Google could auction off ads in realtime, for printing to the page around conversations, filtered and qualified by social interaction data and constrained perhaps by relationships, it could conceivably personalize targeted advertising and also push a new class of social sales and offers to the user's social graph. That is, reaching friends through those most trusted and respected for their influence in their areas of expertise.

The grail of advertising is not one to one relationships with customers, but access through the right person to a whole network of friends. In or around their own words and at the time most likely to get attention. Realtime is solving the attention problem by capturing it when it's being paid. But it takes a company with a lot of social data to connect the dots and provide social relevance. Google is looking a lot smarter of late.

Labels: , , ,

Thursday, October 08, 2009

Realtime streams: now and then

All social media involve a dislocation that de couples the act of communication or interaction from its artifact, which is a text or recording. This is a shame, in some respects, but one that creates possibilities that wouldn't exist if it weren't for the medium. The medium allows us to be always here and now but visible elsewhere anytime. It has a built in "anyplace, anytime."

This anyplace, anytime is brought into focus by each of us when we use social media. For us it's always now. When I use twitter, I use it now. If I read your tweet, it's now. Your now, which is now "then," is again "now" for me. In reading your tweets I experience them in my own time, even though they were written by you in your time. On your time.

These different times become irrelevant to the medium, for each user's activity makes them present. But the differences do have consequences for some of the medium's particular capabilities. One of these being its way of focusing and harnessing our attention.

Media theory makes the observation that media, or mediated experiences, amplify along some axes of experience while bracketing out others. The phone: voice, and talk. Tv: the eye, and watching. Twitter: the now?

If each of us is in the now but in our own now, then the dislocation and de coupling of a tool like twitter is exacted on the time dimension. We don't experience it that way, because we're always "in time." But we do experience the temporal artefacts, if you will, of the dislocation.

For we in being on twitter, now, we're paying attention to other people, seeking attention from other people, who are not there now, or not in our "now," even though the tool makes it seem so. There's a temporal illusion, if one may mix metaphors ontologically. And I think this may have something to do with the residual practices that develop around attention and which contribute to the attention economy.

I am on twitter now, and for all intents and purposes you seem to be too, or rather, I'm experiencing you now (even though it's now "past" and "then" for you). If I pay attention, by tweeting, tweeting to you, retweeting you, or even simply by reading/observing (which is paying but not giving attention), then I'm being social. I'm engaging in a social act. That social act connects us virtually, because I'm paying attention to you. And if I tweet, some part of that attention wants to close the loop with you. It wants a response.

All social action, mediated especially, intrinsically seeks a return look, a response, if not from you then some other person. It's a tacit social principle and basic social binding mechanism, meaning that it goes without saying.

"Goes without saying." Communication, because it has other people in mind, does a lot that goes without saying. The return is what we want from twitter — and the reason that so many new users drop it. The simplest return is the follow — and the reason so many use following strategies. But talk intrinsically begs the question, makes the appeal, and suggests the response. Talk is structured so that every linguistic statement suggests appropriate, valid, responses. That's how language and meaning work.

The dislocation of all these attention flows, for we are all in the flow of attention, from the streams that result from them, creates a fundamental social "desire" for relocation, or connection. All these mediated forms of talk are looking for ways to make communication more probable, more successful, and more valuable.

The dimension of time is a hidden dimension but one that we know is there, and which operates at a deep level, because twitter is a tool of now. We may see the streams of others, but we experience them in the flow of our own.


This post in continuation of a thread:
Synchronic and diachronic readings of activity streams
The Flow Past Web: even better than the RealTime thing
Activity Streams: Realtime and Streamtime
Activity Streams: Content and Flow

Labels: , ,

Tuesday, October 06, 2009

Brands, and putting twitter word of mouth in context

An interesting study of twitter's viability for eWom, or electronic Word of Mouth marketing, has been making the rounds (Twitter Power:Tweets as Electronic Word of Mouth
). The research involved analysis of 150,000 tweets, treated as natural language expressions, or "talk". The aim of the research was to study tweets in which brands are mentioned for a number of attributes relevant to brands, including sentiment, purpose, frequency, and so on.

I found this interesting for several reasons. First was that I've been arguing of late that the conversational turn in social media (twitter, status updates, et al) makes everyday speech into a commodity. That the medium's translation of talk into a form that can be captured, saved, studied, mined, and so on only points to the further use of consumers for marketing purposes. (While I don't personally like this, it has a whiff of inevitability about it. The frontier having shifted from what we consume to what we say.)

This research is a rich study of the tweet in its commodity form: removed from the context of twitter user relationships and from any kind of transactional or conversational context. (Tweets used were extracted for their mention of a brand names studied.)

Secondly, the research finds that "most tweets that mention a brand do so as a secondary focus." I described this in much less precise terms last week, arguing that brands might focus less on how they are reflected in consumer sentiment and more on how the consumer seems to identify with and through brands in online social contexts. The research seems to have found, in other words, that brands are not the sole object of tweets that mention them. Brands are mentioned in passing, in conversation, yes, but not with the intent of soliciting interaction with the brand.

Interestingly, the research cites an assumption examined elsewhere that "consumers engaged in relationships with brands in a manner similar to the personal relationship they formed with people," adding that in online branding "These brand relationships may be the result of participation in brand communities."

I think there are nuances here worth some investigation. A brand's significance to a consumer may in fact have little in common with human relationships. Of course this changes if the brand community manager and consumer interact online. But the "brand" seems to me more likely to involve values, interests, and personal as well as social meanings associated with a brand but not directly caused by it.

Perceptions, reputation, trust, admiration, coveting... these are aspects of human relationships but are not in themselves relationships (to me, at least). And I think they are shaped socially, not in direct reflection on the brand's messaging and image-making.

Also of interest to brands in this study would be the preponderance of positive sentiments expressed in tweets that mention brands: "more than 80% of the tweets that mentioned one of these brands expressed no sentiment. This indicates that people are using Twitter for general information, asking questions, other information-seeking and -sharing activities about brands or products, in addition to expressing opinions about brands or products. Of the 268,662 tweets expressing sentiment, more than 52% of the individual tweets were expressions of positive sentiment, while ≈33% of tweets were negative expressions of opinion."

If I were a brand manager I would want to see these tweets in context. A research or monitoring tool able to show me context of conversation and something of the relationships that leap to life in the course of that conversation.

And I think it's important here to note that "relationships" can be fleeting, transient, and as they often are in conversational media, a sign of the medium's "coincidensity" and speed.

Referring to the brand model of Esch, Langner, Schmitt, & Geus, the authors write of online consumers, that "current purchases were affected by brand image directly and by brand awareness indirectly."

This will be obvious to a brand manager, but current twitter and social media analytics tools can derail the most disciplined analyst. Mentions are the most easily captured signs of social media relevance to branding. But "indirect awareness," which I read as "socially-mediated branding," is harder to track and quantify. Lest the ROI debate threaten to rear its head here, I still think that a softer, more subjective, "sociability" review belong to the social brand's marketing efforts.

Of the four types of brand-relevant tweeting listed here, for example, it would be interesting to know who sentiments were shared with; who was information solicited from; who was it provided to; and in what was the brand comment a reference to?

Furthermore, and I know that these questions aren't yet supported by tools, and so don't scale well: can the brand learn from how it is identified with, whether its social standing is increasing or slipping, or what kind of person the band information is sought from? Are users with social status, fame, success, knowledge, credibility as experts or reputations as critics, solicited or offered brand-relevant tweets?

I suspect that the types of expression listed here would need to be read closely for how they are addressed, and for how they might reflect on their authors. For tweets that mention brands are often a reflection of social relevance. A tweet asking for ticket information on a band is also a sign of an excited concert-goer: a sign of support and interest as much as the need for information.
  • "Sentiment: the expression of opinion concerning a brand, including company, product, or service. The sentiment could be either positive or negative.
  • Information Seeking: the expression of a desire to address some gap in data, information, or knowledge concerning some brand, including company, product, or service.
  • Information Providing: providing data, information, or knowledge concerning some brand, including company, product, or service.
  • Comment: the use of a brand, including company, product, or service, in a tweet where the brand was not the primary focus."

It's good to see research on this, and especially good to see research that regards tweets as utterances. If we are ascending the ladder of meaning and complexity from the word through the search phrase, on to the utterance, then perhaps it's not so far out to hope we will reach the rung of conversation in the not-so-distant future.

More from Twitter Power:Tweets as Electronic Word of Mouth:

They report that: "Of the 14,200 random tweets, 386 tweets (2.7%) contained mention of one of the brands or products from our list (Table 1). There were 2,700 tweets (19.0%) that mentioned some brand or product, inclusive of the brands that we used in this study."

And of greater interest to brands, would be the preponderance of positive sentiments expressed: "more than 80% of the tweets that mentioned one of these brands expressed no sentiment. This indicates that people are using Twitter for general information, asking questions, other information-seeking and -sharing activities about brands or products, in addition to expressing opinions about brands or products. Of the 268,662 tweets expressing sentiment, more than 52% of the individual tweets were expressions of positive sentiment, while ≈33% of tweets were negative expressions of opinion. This is in line with prior work such as that of Anderson (1998), who showed that there was a U-shape relationship between customer satisfaction and the inclination to engage in WOM transfers. This suggests that extremely positive and satisfied and extremely negative customers are more likely to provide information relative to consumers with more moderate experiences."

"As can be seen from Table 7, most tweets that mention a brand do so as a secondary focus. These tweets account for just under half of the branding tweets in this sample. Users expressed brand sentiment in 22% of the tweets. Interestingly, 29%of the tweets were providing or seeking information concerning some brand. This shows that there is considerable use of microblogging as an information source. This would indicate several avenues for companies, including monitoring microblogging sites for brand management (i.e., sentiment), to address customer questions directly (i.e., information seeking), and monitoring information dissemination concerning company products (i.e., information providing)."

Labels: , , ,

Social media, converging streams?

One of my favorite books about community is a work by Nobel Prize winner Elias Canetti called Crowds and Power. It's a beautiful and thoroughly insightful study on people assembled in different ways and for a kaleidoscopic set of reasons. I turn to the book often when thinking about how social media both separate and connect us, using it as an imaginary frontier of sorts for what mediated crowds might or could do.

A piece by Tim Leberecht reminded me of Canetti this morning. Got me thinking about converging streams and how conversational media sometimes produce that effect of being together at the same time.

Which is really a matter of paying attention at the same time, more than of being together, for the medium only connects across our individual spaces and times. The Germans have a nice word for the sense of being with others: "Mitsein." "Being with" is contrasted with contiguity, or being "next to" or adjacent to one another. We're not in one another's stream of consciousness when we are just next to one another; we are when we are "with" one another.

There is no "Mitsein" online, but there is a sense of something that approximates it. But it comes not through being together. It comes through talk. Talk that indicates we are here and now, paying attention. The response is its signal flare.

In a medium so perfectly suited for a kind of self-talk, or talking aloud in front of others, it might be strange that there are occasions when we get a sense of Mitsein. Approximated, of course, in the medium's own peculiar kind of proximity, or proximate intimacy. An "approximity" perhaps. A blend of the real and the imagined, of memory and expectation.

Verbal communication, not the language of bodies sharing space as in Crowds and Power, produces this approximation online. The kind of talk that appeals for a response. The kind of talk that runs out a line with hooks.

Hooks are important for conversation. I much prefer dialog to monolog. Hooks, in the form of "and you?" strung out along the thread of a good conversation are what call me into the world of people. I listen, I pay more attention, when conversation is drawn by the two of us. I like interruptions and clipped sentences, finishing one another's thoughts, and mutual effort of threading out a good line together.

I wonder if the brief moments of simultaneity that pass now and then across our webbed social spaces will result in stream convergence. If the community of talk media might lie not in distributing messages but in the sense of sharing time. And if the point of doing more to make streams — of messages and update and activities — more interesting is also to create more hooks by which to connect them. If streams, like people, not only want the greater flow of the river but also the shared flow of time.

Labels:

Monday, October 05, 2009

Social Interaction Design: Ratings

I had other things in mind for this morning until a client sent me an article in today's Wall Street Journal about online ratings. She, like many others running review and ratings-based sites, is "suffering" from excessively generous end user ratings. The article, which surveys a number of online properties, cites the tendency to 4.3: On the Internet, Everyone's a Critic But They're Not Very Critical.

Offering up a number of anecdotes as reasons for the broken state of online ratings, the article's authors pretty much capture what many of us get intuitively about why online ratings really don't work.

I thought I'd break this down from a social interaction design perspective to get at some of the causes of this.

First and foremost is the fact that most online systems built to capture user tastes, preferences, and interests engender bias. And online media amplify bias, for a number of reasons.

This bias originates with the user's intention, which goes unknown and is not captured in the rating system itself. The reasons a user may have for rating something can be many: a mood, attitude, a personal interest, a habit of use, interest in getting attention, building a profile, promoting a product, and so on.

Social media, because they provide indirect visibility in front of a mediated public, amplify any distortion baked into the selection itself (a selection being the act of rating something). This amplification is explained in part by the de-coupling of selective acts (rating) from consequences and outcomes.

Selections are de-coupled from personal consequences, which excuses a certain lack of accountability and responsibility. Selections are de-coupled from their context of use, which range from personal utility to social promotion. And selections are de-coupled from social implications, which removes the user from his or her contribution to a social outcome (eg, highly-rated items look popular).

Consider the reasons a user may have for making a selection (rating something). They include:
  • personal recollection (like favoriting)
  • to inform a recommendation engine (so that it can make better personal recommendations)
  • because the item is a favorite (sharing favorites)
  • because the social system has no accountability
  • because it always creates the possibility of recognition for the user
  • because it promotes the item
  • because it's nice (socially; possibly karmic)
  • because it's a gesture about how the user felt

Social selections are thus encumbered by ambiguity: of intent, of meaning, of relevance, and of use.

Can these be addressed and resolved by better system design? Or can they only be resolved by social means?

It might be possible to couple ratings with outcomes. This would involve new sets of selections and activities made available to other users and used to create consequences. Users would then consider these consequences when making a rating selection.

Contexts of use could be distinguished, so that users rate with greater purpose. This would involve creating new views of rated content, such as "rate your favorite item this wk," "rate your favorite genre," "rate your personal favorite," "rate which you think is the best," and so on. Each of these distinctions, if followed by users (!) would specify the selection by means of a different social purpose.

It might be possible to reduce ambiguity by means of some cross-referencing achieved by algorithms and relationships set up in the data structure. Without detailing these, they would probably include means by which to distinguish:
  • the bias of the user him or herself, measured in terms of personal tastes
  • the domain expertise of the user, as demonstrated by ratings provided by the user on other items and in which categories/genres/domains
  • the social communication and signaling style of the user, which would reveal some of his/her relation to the social space
  • use by other users and the public, as a measure of relevance

Cross references could then be applied when aggregating ratings, used to filter and sort the ratings sourced for averaged results. Theoretically, the system would be able to identify experts, promoters, favoriters, and others by their practices.

Social solutions might be created to supply distinctions among the different kinds of social capital involved in ratings. Such as:
  • the user's expertise (domain knowledge)
  • trust capital, or the user's standing within his/her social graph
  • credibility capital, or the user's believability, as measured in loyalty perhaps
  • reputation capital, or the tendency of the user's ratings to be referred to and cited beyond his/her immediate social graph

Finally, ratings systems can diversify possibilities for making selections, and separate communication from ratings selections so that ratings are used less for visibility and attention-seeking reasons (eg users who rate a lot).

There are too many kinds of socially-themed activities and practices in which ratings play a part for me to delve into this here. But each theme could be examined for the social benefits of ratings, for how they attribute value to the user, add value to content, and distinguish social content items to result in shared social and cultural resources. Those distinctions could be used to isolate different rating and qualification systems so that they are tighter and less biased.



Recent related posts:

Labels: ,

Friday, October 02, 2009

Twitter, Google Wave, and Online Talk

This post is a reflection on some questions raised by Adina Levin in a post on Google Wave dated July. I haven't myself used the product, so this is not a product review but is instead a continuation of some of the thoughts Adina raised around Wave's social models. I'll speak here more to the ongoing innovation in conversation tools rather than attempt even educated guesses as to Wave itself. I should also say that this post is un-premeditated and off the cuff.

Wave is a communication tool. In that, it will be compared to twitter. But there seem to be substantial differences between the two, as many (some) will no doubt have experienced from using Wave. Regardless of Google's strategic interest in launching Wave (as a response to twitter or not), they seem to bear resemblance only in their contributions to the conversational trend in social tools and social media.

Twitter creates a mediated public, and this means that twitter users are not only using it for communication but for social reasons also. As I tried to show in a recent post (Social media: the attention economy explained), the user's awareness of this public results in incidental social effects, byproducts, and outcomes.

Communication in twitter is not just a matter of talking to people but of being seen talking by the public — or at least being aware that one's communication may be seen. The tweet itself thus takes on two forms. One, the statement itself, which may be described as communication (what a person says). And secondly, the commodity form of the tweet, which is an artifact of digitally mediated communication and which results in statements being re-distributable.

As I've written elsewhere, this makes many conversational tools a means of production: of the self, of relationships, of visibility, presence, status, and so on. In the communication age, these tools are an intrinsic part of the attention economy and of the manufacture, if you will, of a mediated self: one that is extended across time and space, represented and captured online.

Note that the "Self" is always extended across time and space through relationships; but media offer the possibility of representing this extension. This means that distribution becomes as important a factor in a social tool's use as communication (talking with the purpose of reaching understanding with somebody about something).

The tweet as commodity form plays into and allows many social and cultural practices involving social visibility, status, reputation, and other aspects of individual identity and social position. Redistribution of the commodity form of the tweet, as seen in retweeting "influencers," sharing news, linking to blog posts and sites, announcing one's activities, engaging in social rituals such as #followfriday, social and event pics, declarations of gratitude — all these social activities are achieved using direct and indirect acts of communication.

They use both forms. Loosely coupled or un-coupled linguistic statements; and the commodity form of the tweet, whereby the tweet is essentially a social object, and the act of distributing it supplements its meaning.

The meaning of an act of communication, the stated meaning of the tweet in other words, is the communicative act. The commodity form is the meaning the tweet has that's not in the statement but obtains from its use. A retweet is a statement retweeted and thus the act of retweeting is a social act which has its own social meanings above and beyond what the retweeted tweet actually says.

Twitter thrives on the supplementary meanings that are produced in the wake of its unique discontinuities: de-coupled conversational turns, out of synch and time, each experienced in a view particular to the user's own selection of followers. In contrast to Wave, twitter is a disaggregated social space. Each of us has his or her own window onto a social world taken in through stretches of thin but durable attention streamtime.

These social acts, which are virtually unlimited in possibility given twitter's open structure and lack of social design (no groups, virtually no functional syntax, no navigation besides chronological, etc) result in a highly inefficient social space rich in ambiguities that are as compelling and engaging as they are frustrating.

My position on this is that ambiguities of social action and intent, as well as of linguistic meanings, are the fuel of conversational media. For the greater the ambiguity of intent and meaning, the more social relationships and interpersonal handling (interaction) has to do. The more it has to do, the richer the social possibilities.

Social conventions and practices supply understanding to compensate for design inadequacies. In short, loss of context is addressed by social action and emerging practices. Practices provide a different type of context, one not of design but of interaction.

Google Wave seems intended to capture conversation in its context. It is not a public social tool, and not likely to engender the types of social visibility, identity, status, and so on that have made twitter what it is. As such, it seems interested in providing a functional improvement to conversation, by means of design, by means of containing the audience, by means of capturing and offering playback of past conversation.

Adina writes: "Wave is a toolset with even more flexibility than a wiki, with even more interactive content. This poses even greater challenges to help people understand how to use it and be productive."

I would object somewhat to the suggestion that conversational tools ought to be designed with productivity in mind. And to the idea that the tool has a way of being used. Conversation itself has structure and organization, both internally (linguistic statements have grammar, syntax, and semantic stabilities) and pragmatically (conversations involve moves, turns, and many selections that expose how participants interpret what's going on).

Adina addresses Wave's threading: "When there are comments interspersed between paragraphs in email/forum threads, it can be difficult for newcomers to get the gist of what has occurred. But there is a time-honored way to bring people up to speed — summarize the conversation to date. The summary has a social purpose, too, it steers the discussion toward a state of current understanding." I beg to differ, again, on the last point. I don't believe there is such a thing as "current understanding."

Conversation is itself an action system. Communication, say a statement, that is not answered is only an observed act of communication. Communication that is picked up is social action. The act of responding to, or picking up, a statement is an act. It has linguistic meaning (what's said in the response) and it has social meaning (to those participating). So from the perspective of mediated social interaction, conversation is more than reaching consensus ("current understanding") about what's been said so far.

Many participants, in fact, will have relational interest in the conversation to date. Not just what has been said but who said it, to whom, how, and so on. This is the drama and performance of talk, and has a great deal of social relevance to participants as well as to those who use the playback feature.

Adina raises good questions about Wave's social models, and directly poses the matter of groups: "The differences between these models make a vast difference between how the tools are used and what they are good for." Again, I wonder whether groups are even the right design approach in conversational tools. It could be that we need to think in terms of social action, interaction, and conversation rather than groups.

Groups aggregate an audience (participants included in the group), capture attention, provide social inclusion (and exclusion), and create a place or context for communication. Wave might make some of this irrelevant (I would need to use it to better understand design implications and models).

I have a preference for thinking in terms of frames of experience and interaction, over abstracted social models, and particularly those that imply containers. For in conversational tools, the interactions can have order and organization (for example, they have temporal order: a matter clearly addressed in Wave) without need for audience containers (groups, pages, place). Context can be created ad hoc as messages are threaded, arranged, re-aggregated, and so on.

That's about all I wanted to say. Conversational tools are in part such a rich turn in social media's evolution because the act of talking in front of others, in a form that can be redistributed and stored, will always engage social interests. All statements have a double meaning. That of the statement and that of the act of making the statement. It belongs to communication itself that we can tell the difference between the statement and its production (utterance and the uttering of the utterance). So for this reason, open public social spaces always enjoy the social play of ambiguity: of intended meaning and of the social act of making, circulating, referencing statements (and their authors!).

Wave seems to want to streamline the conversational experience. I cant see how it would possibly relate then to the unique sociality of twitter. That it might offer up possibilities not only for use of conversation, but for meta functionalities derived from the observation, visualization, navigation and analytics of conversation seems, however very clear. But that would be a different post entirely!

Labels: , ,

Thursday, October 01, 2009

Foursquare vs Yelp: Recommendations and Reviews

Foursquare and Yelp are each sites that capitalize on user-contributed reviews and recommendations. Users contribute their favorite places and things to do, spotlighting best-kept secrets and customer favorites. Users get visibility and even some amount of notoriety for their contributions. Their enthusiasm for, or against, a merchant can have substantial repercussions for businesses. In the age of social media, might sometimes makes right, whether the customer is "right" or not.

Social interaction models: Yelp and Foursquare
Yelp and Foursquare offer an interesting comparison in the use of social interaction models. For each of them has had to create a compelling and engaging social experience, and has done so with some degree of success.

Yelp has done it with a slight twist on reviews. Yelp's reviews may be associated with a business, but are in fact as much about their authors as they are the business reviewed. On Yelp, users can profile their tastes, interests, habits, and opinions through the places they frequent. In this way, Yelp makes it easy for users to talk about themselves without having to fill in the "about me" box so common to un-themed profiles.

Foursquare does it with a combination of recommendations and offline activity check-ins. Users leave short posts recommending things to try or do at a location, and then separately check in to locations they visit. In a sense, Foursquare extends the practice of reviews by going mobile: Foursquare can be used to find friends on the go. But it substitutes the recommendation for the review, and in its focus on messaging over review writing, seems more closely aligned to social interactions and relationships than to reviewer taste profiling and publishing.

Yelp's interaction models: extracting the value add
From a social interaction design perspective, the differences between Yelp and Foursquare are interesting. Each site is designed to capture users interested in real places.

Yelp captures interests in particular places and makes connections to other similar places: it turns the individual user's subjective interest into an objective "type" of interest, and constructs relationships that are then surfaced as a directory of sorts.

For example, interest in one Chinese restaurant can be used to create links to other Chinese restaurants. Yelp can get as specific with this as it's able to subdivide interests. Theoretically, it could get down to specific dishes, to service, price, ambience, and so on. It could do this (and does in some attributes, like price) by means of structuring form input at the review, or by extracting meta data by mining text (less reliable).

This is the essential practice of end-user review sites, and rests on the assumption that subjective review content can be translated into common social values. I call this "taste making," for it by-and-large corresponds to the role played by media in our culture, relying in this case on local and "authentic" experts over accredited or branded (mass media) experts.

Bias in the model
The transformation of subjective interests (values held by the individual user) into some form of socially valid tastes and opinions is undermined, however, by the introduction of bias in the social practice of reviews. Bias enters the system because reviews not only serve to describe a business, but to express individual user personality also.

In any social system, the user's interest in making an impression, and being seen (popularity, respect, credibility, or other form of social rank), introduces a second incentive to the core activity. If the core activity is the "review," then motives corresponding to the system's social architecture distort behavior. And indeed, popularity, leaderboard rank, visibility, follower count, and any number of similar social effects can be motivating to users for whom online interactions serve personal and psychological interests (which is not only commonplace, but deeply sticky).

Whether the user is interested and motivated by trust, reputation, celebrity, credibility, intellect, experience, or something else, will factor into his or her habits and online social participation styles. Engaging with these motives is essential to participation, but also contributes to the social bias and distortion of social content. No amount of filtering, sorting, or ordering user contributions can eliminate bias if it has been introduced by the motivating attributes of a social system.

Separating social interaction from content production
What Yelp has done, and which was smart (if unintended, I don't know), was to offer symbolic and gestural tokens and icons to users for use in communicating with each other. This not only had the effect of building social relationships (compliments are great ice breakers) — it also offloads social interaction and communication into a separate social system. Users need not speak to each other in their reviews, but can do this by means of tokens. Reciprocity, as a social norm, then comes into play and encourages positive social behaviors. And exchange and gift economies come into play as a social mechanism governing the use of these tokens.

(Note that in many social systems, these tokens are an unlimited social resource; if there were limited numbers of tokens available to users, competition for possession of tokens for social rank would govern the dynamic.)

Foursquare's interaction model: social activity
Now let's look at Foursquare. In contrast to Yelp, Foursquare users profile themselves by where they have been, and to some degree by what they have done (insofar as they post a statement about it.) A look at Foursquare posts shows that consensus seems to emerge quickly around points of interest. Users may be more inclined to agree with one another on what makes a place good. But that's not likely the reason for the uniformity of their posts. More likely is that the form here is the recommendation, not the review.

Recommendations are intrinsically more social: they are directed at an audience. And on Foursquare, the audience is those who are going to a venues, not those who are comparing venues (by review shopping). Not only are recommendations addressed to people (reviews being written for a public), they are most likely to cite the best thing to do.

And indeed, Foursquare seems more interested in cultivating social activity than in building a community of experts. Social activity benefits Foursquare by motivating users to check in to a venue when they are there, which in turn provides presence and location information useful to the mobile user.

Foursquare was built in the era of twitter, and takes inspiration more from tweeting than from writing. It serves communication and social connectedness; this, again, is clear from the site's emphasis on friends and followers.

To help serve this purpose, the game-like aspect of Foursquare has been implemented well. A variety of badges provide two social functions: differentiating individual users from the user population overall (users differentiated by having a badge), and identifying user interests (by what the badge means). As with Yelp, the ambiguity involved in what a badge means can be compelling in itself (in Foursquare: is she a "player," or does she just travel with male friends? Did she mean to look like a player or is that Foursquare's doing?)

The interest here is a social interest. Foursquare attracts users who enjoy playing: for mayor, for stats, for badges, and to a lesser extent, for friends. Because social gaming and games suspend the normal conventions of social interaction while at the same time putting real relationships into play, there are endless variations Foursquare can roll out in the future.

For example, the site could embrace interests of users to whom pure social games are less appealing, and instead address their inclination to be taste makers, demonstrate expertise, display their depth of local experience or knowledge, and more. Foursquare could provide modalities to end users to bring attention to these other user personality types. Photographers might twitpic scenes and situations grabbed on location. Contests could be staged for "best of" category, including discoveries, best-kept secrets, and the more obvious local favorites. City walks could be extracted from local mayors for tips on a great first date, things to do on a family visit, or bartenders and service staff who are fun to talk to.

Frames of social activity
The advantage earned by Foursquare obtains from channelling social activity into social games. These games generate participation, offer a compelling engagement model, are fast and relatively quick and easy, and can be used as an interaction system for many different kinds of content.

From a social interaction design perspective, games are frames: interaction and user experience are framed by the game. All social situations involve a frame of some kind, whether mediated online or not. This frame supplies participants with an idea of What's going on and How to proceed, both critical aspects to social interaction.

Use of a frame other than that intrinsic to the content itself provides other, and new, things to do. Therein lies the innovation of socially-mediated experiences: experience frames that leverage and extend relationships, forms of talk (questions, recommendations, etc), interactions with tokens (eg social gifts), gestures (eg compliments), and so on. Social interaction designers can use frames to organize social interaction around content, and thereby offload some of the social motives from content left behind, improving its value to non-participating users. Or the opposite: to concentrate social motives into communication in order to thicken a system's social sticky. Every frame brings with it new ways to capture user interests and motivations.

Conclusion and implications
Interaction models that directly relate to users and what they find interesting, and not concepts like "community" or the "social graph," are in my opinion the more precise approach to designing and leveraging social media. Since all social media involve some variation on talk and talking, interactions can be structured and organized by design and their outcomes ordered and presented to lay emphasis and focus on the aspects and social dynamics that propel a social system forward. We do this best, I think, not by abstracting models but by aligning them closer to user experiences. The richer our understanding of what users are like and what they do, the better our interaction models will be.

Labels: ,