Chapter 13. Social Media Junkies Unite!
Some commons-based peer production efforts are less
self-conscious on the part of the users, and emerge more as a function
of distributed coordinate behavior, like del.icio.us or Flickr. The
critical defining feature of these “enterprises” is that they rely
primarily on social information flows, motivations, and relations to
organize the group. Individuals self-identify, mostly, for tasks, and
through a variety of peer-review mechanisms contributions get
recognized by the group and incorporated into what emerges as the
collaborative output.
—Yochai Benkler interview in OpenBusiness (about his book
The Wealth of Networks)
As with “Web 2.0” and “synergy,” the buzzphrase “social media”
has taken on a life of its own, and has already spawned its own mutant
bastard: “social media marketing.” As with those other terms, and much
Internet jargon, “social media” means different things to different
people. As noted at the beginning of this book, we prefer to use the
term to refer to social production and consumption of media objects,
and not merely as a lazy synonym for social networking in general.
This entire chapter is devoted to the collecting, sharing, and
creation of social objects. As Hugh MacLeod put it in a post on his
weblog, called “Social Objects for Beginners” (http://www.gapingvoid.com/Moveable_Type/archives/004390.html):
The Social Object, in a nutshell, is the reason two people are
talking to each other, as opposed to talking to somebody else. Human
beings are social animals. We like to socialize. But if you think
about it, there needs to be a reason for it to happen in the first
place. That reason, that “node” in the social network, is what we
call the Social Object.
So how do these social media objects swirling around us add up
to an ecosystem or even a marketplace? How do people discover them,
subscribe to them, and share them across networks? (See Figure 13-1.)
See Chapter 17 for a further
discussion of microformats and semantic markup in general.
Social media is a two-way street: read/write. In addition to
providing tools for sharing and publishing media, you can provide your
users with interfaces for zeroing in on the streams they’re interested
in and then sifting through them for the most interesting and relevant
objects (Figure 13-2).
Also known as asymmetric following (and
explored more thoroughly in Chapter 14), following is a way of
expressing interest in someone else’s activities and objects and
subscribing to them. It does not require reciprocation, and although
it might correspond with acquaintanceship or friendship, it does not
necessarily imply a reciprocal relationship between the follower and
the followed (Figure 13-3).
Related patterns
Add/Subscribe in Add/Subscribe
“One-way following (aka asynchronous following)” on page
364
Updates in Updates
What
As Randy Farmer is fond of reminding us, “context is king.” As
human beings, we rely on context to derive meaning from our sensory
inputs. One of the unfortunate side effects of augmented universal
oversharing is that we get these streams (torrents, really) of updates
and objects from all of our connections across multiple social facets,
usually with most or all of the originating conceptual context
stripped away.
This dissolution of context is alienating and disorienting for
most normal people. Even those of us who are at times capable of
surfing these unrestrained information feeds usually grow weary of the
onslaught eventually.
The first resort for most people is “social filtering,” which
means relying on the pointers of friends and those we follow for
deciding what to pay attention to (Figure 13-4). The ordinary
follow and subscribe interfaces suffice for enabling users to “tune
in” to the recommendations of others, but you can use this pattern to
give people additional handles on which to filter for
context.
Use when
Use this pattern when the potential for information overload and
jumbling together of unrelated contexts grows intolerable (Figure 13-5).
How
Provide affordances for restoring (or, if necessary, imposing)
contextual filters on datastreams so that they can be parsed in more
manageable groupings (Figure 13-6 and Figure 13-7).
Filtering can also be achieved by giving users a way to hide
people or specific types of objects. Instead of singling out a context
and showing just items in that context, which tends to be a temporary
choice, hiding involves singling out a context and filtering items in
that context out of view (Figure 13-8).
People will also use leaderboards (see Leaderboard in Leaderboard), favorites
(see Favorites in Favorites),
and other “best of " tools as an attempt to filter on quality (Figure 13-9).
Why
Giving people the ability to filter incoming information based
on various contexts (type of content, closeness of relationship to the
sender, timeframes) enables them to establish a stable point of view
from which to explore the rich, never ending stream of new objects and
information.
As seen on
Facebook (http://www.facebook.com/)
FriendFeed (http://friendfeed.com/)
Google Reader (http://www.google.com/reader/)
What
In the search for relevancy and quality, people have a difficult
time zeroing in on satisfactory content (Figure 13-10).
Use when
Offer recommendations when you have a sufficient body of data
about your user’s selfdeclared and implied interests as well as a rich
enough social graph to be able to identify similarities and make
helpful guesses about likely interesting content (Figure 13-11 and Figure 13-12).
How
Offer a call to action inviting the user to explore
recommendations. Educate the user about how to obtain better
recommendations (for example, by rating content).
Display recommendations as a list, or if there is a large
number, in a carousel or scrollable window.
Why
Recommendations push objects toward people rather than relying
on them to be passively discovered. If you can provide value to your
users by making educated guesses about the type of objects they are
interested in, then you may be able to capture their loyalty. The
benefit to users is more readily finding the information and media
they need without having to hunt around for it quite so hard.
Related patterns
Testimonials (or Personal Recommendations) in
Testimonials (or Personal Recommendations)
As seen on
Amazon (http://www.amazon.com/)
Digg (http://digg.com/)
The Filter (http://www.thefilter.com/)
Netflix (http://www.netflix.com/)
SeeqPod (http://seeqpod.com/)
StumbleUpon (http://www.stumbleupon.com/)
Twitter (http://twitter.com/)
Social search is an emerging phenomenon, and there are a number
of different aspects of search that can be enhanced with a social
dimension (are you searching for people? are your searches facilitated
by social behaviors? are you searching for social objects?). The
phenomenon of finding content by searching on user-contributed tags is
perhaps one of the most familiar social search experiences available
online today.
The two most interesting forms of social search I’ve seen are
real-time search and conversational search.
What
People can’t always find breaking news and current topics of
public conversation with ordinary keyword searches of indexed web
resources, and already get frequent pointers to current information by
the electronic equivalent of word of mouth (Figure 13-13).
Also known as “The Notificator” (http://www.borthwick.com/weblog/2009/02/05/creative-destruction-google-slayed-by-the-notificator/).
Use when
Use this pattern with an activity stream service to enable
people to find concepts in up-to-the-minute status updates and
activities.
How
Provide the familiar elements of a search interface (a text
box and a search button), and make it clear to the person
searching that the results will be ordered by recency
(reverse-chronological order) and not by relevancy (Figure 13-14).
Optionally, give the user hints about the sort of things
that can be profitably searched for in a real-time search
interface. For example, Twitter Search lists the current top
trending topics (Figure 13-15).
Optionally, offer the ability to subscribe to search
results, most commonly in the form of an RSS feed, to give people
the ability to track a term or phrase and be notified almost
immediately whenever it appears (Figure 13-16).
Why
The world is moving too fast for “old-school” search engines to
keep up with the leading edge. Real-time search tools that capture
signals from the social web provide a method for finding extremely
current information and news.
Related patterns
Tools for Monitoring Reputation in Tools for Monitoring Reputation
As seen on
Google Alerts (http://www.google.com/alerts)
Technorati (http://technorati.com/)
TweetNews, a mashup of Twitter and Yahoo! Boss running on the
Google App Engine (http://tweetnews.appspot.com/)
Twitter Search (http://search.twitter.com/)
Yahoo! Alerts (http://alerts.yahoo.com/)
What
People sometimes want information or advice that can’t be found
in a neutral, objective reference guide, and they would ask another
human being directly if they could find someone interested in or
knowledgeable about the topic of their question (Figure 13-17).
Also known as “Subjective Search.”
Use when
Use this pattern when you wish to foster communication and
cooperation among the people using your social application.
How
Provide a large, inviting text-entry box to encourage
questioners to write full sentences (like a human being) instead
of query strings or Boolean operators, and label the form button
with a word such as “Ask.”
At the same time, expose open questions to people as a way
of inviting them to answer (or route questions to likely, willing
answerers based on affinities you derive from the meta data in
your social graph). See Figure 13-18.
Alternately, as Aardvark does, rely on existing
conversational channels (in the case of Aardvark, primarily IM)
for capturing questions, routing them to potential respondees, and
delivering answers (Figure 13-19).
Optionally, embrace a reputation system to help ferret out
the best contributors and the most helpful answers.
Why
Directly querying an index of data is a great way of searching
for information, with historical roots going back to the earliest
libraries, archives, and repositories, but people have always gathered
information in other ways as well. In fact, most people in the real
world ask other human beings for information as a starting
point.
As seen on
Aardvark (http://vark.com)
LinkedIn (http://www.linkedin.com/)
Yahoo! Answers (http://answers.yahoo.com/)
Most mailing lists everywhere
Some people like to browse, others prefer to search, but most
use a combination of both. And no one ever says to themselves, “Today
I’m going to only browse and do no searching,” or vice versa. A person
may start by searching for information and then when he finds
something juicy, begin browsing from there to related content.
Similarly, browsing may lead to search and then back to browsing
again.
Providing your users with ways to “pivot” between one form of
discovery and another offers them the richest possibilities.
To do so, provide a persistent search box that is always in the
same location (most commonly the upper-right or upper-left of the
screen), and when displaying search results, offer related links,
“more like this,” and other opportunities for lateral
exploration.
Chapter 13. Social Media Junkies Unite!
Some commons-based peer production efforts are less self-conscious on the part of the users, and emerge more as a function of distributed coordinate behavior, like del.icio.us or Flickr. The critical defining feature of these “enterprises” is that they rely primarily on social information flows, motivations, and relations to organize the group. Individuals self-identify, mostly, for tasks, and through a variety of peer-review mechanisms contributions get recognized by the group and incorporated into what emerges as the collaborative output.
Keeping Up
As with “Web 2.0” and “synergy,” the buzzphrase “social media” has taken on a life of its own, and has already spawned its own mutant bastard: “social media marketing.” As with those other terms, and much Internet jargon, “social media” means different things to different people. As noted at the beginning of this book, we prefer to use the term to refer to social production and consumption of media objects, and not merely as a lazy synonym for social networking in general. This entire chapter is devoted to the collecting, sharing, and creation of social objects. As Hugh MacLeod put it in a post on his weblog, called “Social Objects for Beginners” (http://www.gapingvoid.com/Moveable_Type/archives/004390.html):
So how do these social media objects swirling around us add up to an ecosystem or even a marketplace? How do people discover them, subscribe to them, and share them across networks? (See Figure 13-1.)
Social metadata today
Much has been made of this trend labeled as Web 2.0, which places social web tools and services in the forefront of open, collective, and collaborative interaction in and around content and other media objects. One of the benefits of the shift is the by product of semi structured information and metadata in and around the objects and content. Information on the Web prior to this era was, at best, loosely structured, or heavily structured through the process of entry using form-field-heavy interfaces.
Semistructured information is created through the application of the HTML markup structures, human tagging, machine tagging, URLs, page titles, navigation structures, and inbound links. All of this creates metadata that gives some structure. Social tools, such as Yahoo! Answers and Wikipedia, all play a vital part in augmenting metadata around objects.
Much of the social metadata is not created by the people using the services, but is a mixture of user input and smart tools. The exception to this is tagging, but the value of this goes beyond metadata for the sake of metadata; it adds personal value for retrieval and contextualizes information for retrieval and aggregation. Most blogging and wiki tools use proper semantic HTML structures, and many augment these tools with microformats under the surface.
Today’s metadata and future uses
Much of the future lies in an idea that has been around many years. This future is embracing the Semantic Web and similar tools, building on semantic information. Semantically relevant metadata improves relevance in information and media retrieval. The Semantic Web has had a chicken-and-egg problem, as it has the tools to do fantastic things with structured information, but it has been held back by the lack of that structured information at a scale that will make a difference. Today’s social semi structured information gives enough of a boost that Semantic Web tools can begin to provide their long-promised power.
The Semantic Web is based on triples of information: subject, predicate, and object. Using what we have today, this turns into: Thomas uses the resource tag; Thomas’s resource tag points to web page X. This pairing of two triples gives us fairly good information that becomes a building block for future uses.
Using the social tools of today, we are setting up a very nice semi structured foundation for tomorrow. We do this in a variety of ways, but all are built through the patterns required for interaction on social tools. The primary element in social architectures is identity within the service. This allows an understanding of the subject, or who is making a statement. Most services are focused around some sort of social object (link, photo, video, statement, document, etc.) that is shared, pointed to, described, and/or central to a collective effort through conversation or collaborative capturing of a description. These social environments offer rich fodder for the subject (who) and predicate (descriptors) about an object (social object).
Additionally, many of the social tools are capturing ancillary metadata and exposing it. This ancillary data is something Flickr calls machine tags, which include things such as the date the photo was taken, the date the photo was added to the Flickr service, the camera type that took the photo, the geographic location where the photo was taken, etc. This metainformation is not something that is generated by the individual using the service, but by the tools the person uses. This is not only prevalent in tools like Flickr, but also in blogs and community services that have profile information associated with an identity, as well as what others in the service have stated about that identity or activities that identity has participated in.
This richness of metadata can then be used by the tools to serve up better understanding of the objects to improve search, surface highly probable similar items, build highly probable synonyms, as well as help discern meaning of ambiguous terms and statements. This will help us understand, for example, the probability that the Apple Macintosh being discussed in a podcast in the second week of January 2008 in San Francisco is likely different than the Apple Macintosh photographed in roughly the same place seven months later. Our tools can discern that the January 2008 Macintosh is a computer-related discussion because the MacWorld convention is held in San Francisco then, but the July 2008 Macintosh is a fruit at a gathering of an American agricultural association’s large meeting in San Francisco. The ancillary metadata around these social objects give clues to this discernment, but clues are also available in an aggregation of the interests of the people using the identities in the services.
Summary
In short, the social tools we are using today are letting us focus on what we care about, and through the use of lightweight connections and light form fields, are capturing and building a web of semi structured information. The web of semistructured information working as metadata provides enough of a foundation to be used as structured elements, which are the fodder for using Semantic Web tools. This use of the Semantic Web tools leads to better relevance and discernment providing drastically better search to find exactly what the seeker wants, not just what is good enough. This also provides much better capability for aggregating information people care about and would like to keep closer to them.
—Thomas Vander Wal, Principal & Sr. Consultant, InfoCloud Solutions (http://infocloudsolutions.com)
See Chapter 17 for a further discussion of microformats and semantic markup in general.
Tuning In
Social media is a two-way street: read/write. In addition to providing tools for sharing and publishing media, you can provide your users with interfaces for zeroing in on the streams they’re interested in and then sifting through them for the most interesting and relevant objects (Figure 13-2).
Following
Also known as asymmetric following (and explored more thoroughly in Chapter 14), following is a way of expressing interest in someone else’s activities and objects and subscribing to them. It does not require reciprocation, and although it might correspond with acquaintanceship or friendship, it does not necessarily imply a reciprocal relationship between the follower and the followed (Figure 13-3).
Related patterns
Add/Subscribe in Add/Subscribe
“One-way following (aka asynchronous following)” on page 364
Updates in Updates
Filtering
What
As Randy Farmer is fond of reminding us, “context is king.” As human beings, we rely on context to derive meaning from our sensory inputs. One of the unfortunate side effects of augmented universal oversharing is that we get these streams (torrents, really) of updates and objects from all of our connections across multiple social facets, usually with most or all of the originating conceptual context stripped away.
This dissolution of context is alienating and disorienting for most normal people. Even those of us who are at times capable of surfing these unrestrained information feeds usually grow weary of the onslaught eventually.
The first resort for most people is “social filtering,” which means relying on the pointers of friends and those we follow for deciding what to pay attention to (Figure 13-4). The ordinary follow and subscribe interfaces suffice for enabling users to “tune in” to the recommendations of others, but you can use this pattern to give people additional handles on which to filter for context.
Use when
Use this pattern when the potential for information overload and jumbling together of unrelated contexts grows intolerable (Figure 13-5).
How
Provide affordances for restoring (or, if necessary, imposing) contextual filters on datastreams so that they can be parsed in more manageable groupings (Figure 13-6 and Figure 13-7).
Filtering can also be achieved by giving users a way to hide people or specific types of objects. Instead of singling out a context and showing just items in that context, which tends to be a temporary choice, hiding involves singling out a context and filtering items in that context out of view (Figure 13-8).
People will also use leaderboards (see Leaderboard in Leaderboard), favorites (see Favorites in Favorites), and other “best of " tools as an attempt to filter on quality (Figure 13-9).
Why
Giving people the ability to filter incoming information based on various contexts (type of content, closeness of relationship to the sender, timeframes) enables them to establish a stable point of view from which to explore the rich, never ending stream of new objects and information.
As seen on
Facebook (http://www.facebook.com/)
FriendFeed (http://friendfeed.com/)
Google Reader (http://www.google.com/reader/)
Recommendations
What
In the search for relevancy and quality, people have a difficult time zeroing in on satisfactory content (Figure 13-10).
Use when
Offer recommendations when you have a sufficient body of data about your user’s selfdeclared and implied interests as well as a rich enough social graph to be able to identify similarities and make helpful guesses about likely interesting content (Figure 13-11 and Figure 13-12).
How
Offer a call to action inviting the user to explore recommendations. Educate the user about how to obtain better recommendations (for example, by rating content).
Display recommendations as a list, or if there is a large number, in a carousel or scrollable window.
Why
Recommendations push objects toward people rather than relying on them to be passively discovered. If you can provide value to your users by making educated guesses about the type of objects they are interested in, then you may be able to capture their loyalty. The benefit to users is more readily finding the information and media they need without having to hunt around for it quite so hard.
Related patterns
Testimonials (or Personal Recommendations) in Testimonials (or Personal Recommendations)
As seen on
Amazon (http://www.amazon.com/)
Digg (http://digg.com/)
The Filter (http://www.thefilter.com/)
Netflix (http://www.netflix.com/)
SeeqPod (http://seeqpod.com/)
StumbleUpon (http://www.stumbleupon.com/)
Twitter (http://twitter.com/)
Social Search
Social search is an emerging phenomenon, and there are a number of different aspects of search that can be enhanced with a social dimension (are you searching for people? are your searches facilitated by social behaviors? are you searching for social objects?). The phenomenon of finding content by searching on user-contributed tags is perhaps one of the most familiar social search experiences available online today.
The two most interesting forms of social search I’ve seen are real-time search and conversational search.
Real-Time Search
What
People can’t always find breaking news and current topics of public conversation with ordinary keyword searches of indexed web resources, and already get frequent pointers to current information by the electronic equivalent of word of mouth (Figure 13-13).
Also known as “The Notificator” (http://www.borthwick.com/weblog/2009/02/05/creative-destruction-google-slayed-by-the-notificator/).
Use when
Use this pattern with an activity stream service to enable people to find concepts in up-to-the-minute status updates and activities.
How
Provide the familiar elements of a search interface (a text box and a search button), and make it clear to the person searching that the results will be ordered by recency (reverse-chronological order) and not by relevancy (Figure 13-14).
Optionally, give the user hints about the sort of things that can be profitably searched for in a real-time search interface. For example, Twitter Search lists the current top trending topics (Figure 13-15).
Optionally, offer the ability to subscribe to search results, most commonly in the form of an RSS feed, to give people the ability to track a term or phrase and be notified almost immediately whenever it appears (Figure 13-16).
Why
The world is moving too fast for “old-school” search engines to keep up with the leading edge. Real-time search tools that capture signals from the social web provide a method for finding extremely current information and news.
Related patterns
Tools for Monitoring Reputation in Tools for Monitoring Reputation
As seen on
Google Alerts (http://www.google.com/alerts)
Technorati (http://technorati.com/)
TweetNews, a mashup of Twitter and Yahoo! Boss running on the Google App Engine (http://tweetnews.appspot.com/)
Twitter Search (http://search.twitter.com/)
Yahoo! Alerts (http://alerts.yahoo.com/)
Conversational Search
What
People sometimes want information or advice that can’t be found in a neutral, objective reference guide, and they would ask another human being directly if they could find someone interested in or knowledgeable about the topic of their question (Figure 13-17).
Also known as “Subjective Search.”
Use when
Use this pattern when you wish to foster communication and cooperation among the people using your social application.
How
Provide a large, inviting text-entry box to encourage questioners to write full sentences (like a human being) instead of query strings or Boolean operators, and label the form button with a word such as “Ask.”
At the same time, expose open questions to people as a way of inviting them to answer (or route questions to likely, willing answerers based on affinities you derive from the meta data in your social graph). See Figure 13-18.
Alternately, as Aardvark does, rely on existing conversational channels (in the case of Aardvark, primarily IM) for capturing questions, routing them to potential respondees, and delivering answers (Figure 13-19).
Optionally, embrace a reputation system to help ferret out the best contributors and the most helpful answers.
Why
Directly querying an index of data is a great way of searching for information, with historical roots going back to the earliest libraries, archives, and repositories, but people have always gathered information in other ways as well. In fact, most people in the real world ask other human beings for information as a starting point.
As seen on
Aardvark (http://vark.com)
LinkedIn (http://www.linkedin.com/)
Yahoo! Answers (http://answers.yahoo.com/)
Most mailing lists everywhere
Pivoting
Some people like to browse, others prefer to search, but most use a combination of both. And no one ever says to themselves, “Today I’m going to only browse and do no searching,” or vice versa. A person may start by searching for information and then when he finds something juicy, begin browsing from there to related content. Similarly, browsing may lead to search and then back to browsing again.
Providing your users with ways to “pivot” between one form of discovery and another offers them the richest possibilities.
To do so, provide a persistent search box that is always in the same location (most commonly the upper-right or upper-left of the screen), and when displaying search results, offer related links, “more like this,” and other opportunities for lateral exploration.
Further Reading
“Creative destruction ... Google slayed by the Notificator?”, by John Borthwick, http://www.borthwick.com/weblog/2009/02/05/creative-destruction-google-slayed-by-the-notificator/
“Do your friends make you smarter? Exploring social interactions in search,” by Brynn M. Evans, http://www.slideshare.net/bmevans/do-your-friends-make-you-smarter-exploring-social-interactions-in-search
“Social Objects for Beginners,” by Hugh MacLeod, http://www.gapingvoid.com/Moveable_Type/archives/004390.html TweetNews, http://tweetnews.appspot.com/
“Why social search won’t topple Google (anytime soon),” by Brynn M. Evans, http://brynnevans.com/blog/2009/01/30/why-social-search-wont-topple-google-anytime-soon/