Can Facebook Change Its DNA?

by Thomas Vander Wal in , , , , , , , ,


The posting on Facebook for Business or LinkedIn Gets More Valuable received a lot of feedback and discussion. The e-mail discussions has urged me to post about the Facebook DNA and why it will be tough to transition Facebook into a really valuable tool for business without changing the services to a great degree. A social platform like Facebook is desperately needed for business, within enterprise, and organization that have information flows between the people interested in or that are members of an organization. The platform for this type of community has slightly different needs that Facebook does not provide.

Facebook and Its Initial User-base

It is not secret that Facebook started as a services that was by and for students and members of a university setting. The focus was connecting to people you have run across or will run across as friends, acquaintances from social events, or classmates. The services is quite conversational and does a decent job pulling together a person's shared digital lifestream (many others have done this for years and do it much better, Jaiku (now owned by Google) is one example of a much better approach).

Facebook as a service tended to focus on the current conversation as students largely are focussed on the now or short-time frame planning, e.g. it is Tuesday and what are people doing on Thursday through the weekend or who is going skiing this weekend and putting together a carpool. The time span of conversation and interest is etherial and Facebook very much reflects the short term nature of conversation. The focus is also students in networks framed as campuses, which enables this Local InfoCloud of familiarity to be used for memory, which is not needed in the service. This means if I see something in Facebook, say a good restaurant for a special date night, I can track down the person who suggested the restaurant or tap friends in the network (on campus) to recall the restaurant. This foundation of building a digital space for people in a common physical network does not need strong search or capability to hold-on to information that may have future value.

The reliance on building a digital services that replicates the frailties of physical communication (all hallway, classroom, or party conversation is lost to the ether and is reliant of human memory) is problematic when the people using the service are physically distributed.

Future Information Value

Facebook's largest failure is the lack of using digital conversation for the great value digital conversation provides, which is capturing conversation and information. This capturing of digital conversation has great value as the information that is shared often has much greater future value than current value. Take the same restaurant example from before, when a friend recommends a restaurant I may not have a need for that romantic restaurant now, but I most likely may need that information at sometime in the future. There are two things that I can do with that information: 1) Know the information is in the service and can search for it at a later time ("restaurant" and "romantic" in search, which may also find others have made the same recommendation); or 2) Mark the information in some manner for my ease of refinding (flag, favorite, or bookmark it with annotation from my perspective and context).

The Challenge Facebook has Changing

Making this shift in Facebook will be a huge challenge, not for the developers or technology (although Facebook has not modified its interface to handle the scaling of information flows of applications or increase in traffic, but that is another real problem that needs addressing) as they seem capable to make the minor technology changes that would be needed to use the service for improved search or holding onto information. The big pain point will be making a change that is counter to the understanding and use of those who still make up the core of its use-base, university students. When Facebook opened its doors beyond university students it created panic (well deserved) as their closed society was open to all viewers. The difficulty in holding on to information and searching through the services did not keep people from finding party photos or profiles that were meant for just friends and not future employers or others outside their immediate network. A large number of people left, or more accurately stopped using the service as much as they did (after removing the potentially problematic images and content).

There is an incredible mass of information in Facebook, most of it nearly impossible to find and hold onto at this point. Many of the people who have used the service have relied on this obscurity to keep from removing the (at the moment funny, ironic, humorous, or even not fully informed) statements made on walls, groups, status updates, and other places. Changing the core platform to create value for the people using the service as a digital information platform, that is similar to what is expected outside Facebook could have dire repercussions for Facebook as a service. But, not making these changes makes Facebook a really poor platform for business and knowledge sharing.

Valuable Lesson in Foundations

Having build many web base services, intranet, and networked services over the many years I have been doing this it continually comes back to the initial foundations are the ones that are hardest to shake. The initial people using a service shape it for the future and the initial technology decisions of a service often are the ones that are hardest to change. This change is difficult as the practices are socially ingrained into the understanding of the mores of the system and the social interactions are predicated on these understandings. All design and all interaction aspects of systems are political and drive people's motivation for use as they set the rules and social scripts for interaction. The components are not intentionally political, but they have do trigger reaction and alter motivations no mate how slightly. Not only does the medium have a message, but how the services works or does not work has a message that triggers a response for if and how it is used.


A Stale State of Tagging?

by Thomas Vander Wal in , , , ,


David Weinberger posted a comment about Tagging like it was 2002, which quotes Matt Mower discussing the state of tagging. I mostly agree, but not completely. In the consumer space thing have been stagnant for a while, but in the enterprise space there is some good forward movement and some innovation taking place. But, let me break down a bit of what has gone on in the consumer space.

History of Tagging

The history of tagging in the consumer space is a much deeper and older topic than most have thought. One of the first consumer products to include tagging or annotations was the Lotus Magellan product, which appeared in 1988 and allowed annotations of documents and objects on one's hard drive to ease finding and refinding the them (it was a full text search which was remarkably fast for its day). By the mid-90s Compuserve had tagging for objects uploaded into its forum libraries. In 2001 Bitzi allowed tagging of any media what had a URL.

The down side of this tagging was the it did not capture identity and assuming every person uses words (tag terms) in the same manner is a quick trip to the tag dump where tags are not fully useful. In 2003 Joshua Schacter showed the way with del.icio.us that not only allowed identity, upon which we can disambiguate, but it also had a set object in common with all those identities tagging it. The common object being annotated allows for a beginning point to discern similarity of identity─Ás tag terms. Part of this has been driven on Joshua's focus on the person consuming the content and allowing a means for that consumer to get back to their information and objects of interest. (It is around this concept that folksonomy was coined to separate it from the content publisher tagging and non-identity related tagging.) This picked up on the tagging for one's self that was in Lotus Magellan and brings it forward to the web.

Valuable Tagging

It was in del.icio.us that we saw tagging that really did not work well in the past begin to become valuable as the clarity in tag terms that was missing in most all other tagging systems was corrected for in the use of a common object being tagged and the identity of the tagger. This set the foundation for some great things to happen, but have great things happened?

Tagging Future Promise

Del.icio.us set many of out minds a flutter with insight into the dreams of the capability of tagging having a good foothold with proper structure under them. A brilliant next step was made by RawSugar (now gone) to use this structure to make ease of disambiguating the tag terms (by appleseed did you mean: Johnny Appleseed, appleseeds for gardening/farming, the appleseed in the fruit apple, or appleseed the anime movie?). RawSugar was a wee bit before its time as it is a tool that is needed after there tagging (particularly folksonomy related tagging systems) start scaling. It is a tool that many in enterprise are beginning to seek to help find clarity and greater value in their internal tagging systems they built 12 to 18 months ago or longer. Unfortunately, the venture capitalists did not have the vision that the creators of RawSugar did nor the patience needed for the market to catch-up to the need in a more mature market and they pulled the plug on the development of RawSugar to put the technology to use for another purpose (ironically as the market they needed was just easing into maturity).

The del.icio.us movement drove blog tags, laid out by Technorati. This mirrored the previous methods of publisher tagging, which is most often better served from set categories that usually are derived from a taxonomy or simple set (small or large) of controlled vocabulary terms. Part of the problem inherent in publisher tags and categories is that they are difficult to use outside of their own domain (however wide their domain is intended - a specific site or cross-sites of a publisher). Using tags from one blog to another blog has problems for the same reason that Bitzi and all other publisher tags have and had problems, they are missing identity of the tagger AND a clear common object being tagged. Publisher tags can work well as categories for aggregating similar content within a site or set of commonly published sites where a tag definition has been set (but that really makes them set categories) and used consistently. Using Technorati tag search most often surfaces this problem quickly with many variation of tag use surfacing or tag terms being used to attract traffic for non-related content (Technorati's keyword search is less problematic as it relies on the terms being used in context in the content - unfortunately the two searches have been tied together making search really messy at the moment). There is need for an improved tool that could take the blog tags and marry them to the linked items in the content (if that is what is being talked about - discerning predicate in blog tags is not clear yet).

Current Tools that Advanced

As of a year ago there were more than 140 social bookmarking tools in the consumer space, but there was little advancement. But, there are a few services that have innovated and brought new and valuable features to market in tagging. As mentioned recently Ma.gnolia has done a really good job of taking the next steps with social interaction in social bookmarking. Clipmarks pioneered the sub-page tagging and annotation in the consumer tagging space and has a really valuable resource in that tool. ConnectBeam is doing some really good things in the enterprise space, mostly taking the next couple steps that Yahoo MyWeb2 should have taken and pairing it with enterprise search. Sadly, del.icio.us (according to comments in their discussion board) is under a slow rebuilding of the underlying framework (but many complaints from enterprise companies I have worked with and spoken indepth with complain del.icio.us continually blocks their access and they prefer not to use the service and are finding current solutions and options to be better for them).

A Long Way to Go

While there are examples that tagging services have moved forward, there is so much more room to advance and improve. As people's own collection of tagged pages and objects have grown the tools are needed to better refind them. This will require time search and time related viewing/scanning of items. The ability to use co-occurance of tag terms (what other tags were used on the object), with useful interfaces to view and scan the possibilities.

Portability and interoperability is extremely important for both the individual person and enterprise to aggregate, migrate, and search across their collections across services and devices (now that devices have tagging and have had for some time, as in Mac OS X Tiger and now Vista). Enterprises should also have the ability to move external tagged items in through their firewall and publish out as needed, mostly on an employee level. There is also desire to have B2B tagging with customers tagging items purchased so the invoicing can be in the customers terminology rather than the seller terminology.

One of the advances in personal tagging portability and interoperability can easily be seen when we tag on one device and move the object to a second device or service (parts of this are not quite available yet). Some people will take a photo on their mobile phone and add quick tags like "sset" and others to a photo of a sunset. They send that photo to a service or move it to their desktop (or laptop) and import the photo and the tag goes along with it. The application sees the "sset" and knows the photo was transfered from that person's mobile device and knows it is their short code for "sunset" and expands the tag to sunset accordingly. The person then adds some color attribute tags to the photo and moves the photo to their photo sharing service of choice with the tags appended.

The current tools and services need tools and functionality to heal some of the messiness. This includes stemming to align versions of the same word (e.g. tag, tags, tagging, bookmark, bookmarking). Tag with disambiguation in mind by offering co-occurrence options (e.g. appleseed and anime or johnny or gardening or apple). String matching to identify facets for time and date, names (from your address book), products, secret tag terms (to have them blocked from sharing), etc. (similar to Stikkit and GMail).

Monitoring Tools

Enterprise is what the next development steps really need to take off (these needs also apply to the power knowledge worker as well). The monitoring tools for tags from others and around objects (URLs) really need to fleshed out and come to market. The tag monitoring tools need to become granular based on identity and co-occurance so to more tightly filter content. The ability to monitor a URL and how it is tagged across various services is a really strong need (there are kludgy and manual means of doing this today) particularly for simple and efficient tools (respecting the tagging service processing and privacy).

Analysis Tools

Enterprise and power knowledge workers also are in need of some solid analysis tools. These tools should be able to identify others in a service that have similar interests and vocabulary, this helps to surface people that should be collaborating. It should also look at shifts in terminology and vocabulary so to identify terms to be added to a taxonomy, but also provide an easy step for adding current emergent terms to related older tagged items. Identify system use patterns.

Just the Tip

We are still at the tip of the usefulness of tagging and the tools really need to make some big leaps. The demands are there in the enterprise marketplace, some in the enterprise are aware of them and many more a getting to there everyday as the find the value real and ability to improve the worklife and workflow for their knowledge workers is great.

The people using the tools, including enterprise need to grasp what is possible beyond that is offered and start asking for it. We are back to where we were in 2003 when del.icio.us arrived on the scene, we need new and improved tools that understand what we need and provide usable tools for those solutions. We are developing tag islands and silos that desperately need interoperability and portability to get real value out of these stranded tag silos around or digital life.


Open Conversations and Privacy Needs for Business

by Thomas Vander Wal in , , , , , , ,


I thought I would share the latest press bit around this joint, Thomas Vander Wal was quoted in Inc Magazine What's Next: Shout it Out Loud (or in the August 2007 issue beginning on page 69). The article focuses the need and desire for companies to share and be open with more of their data and information. Quite often companies are getting bit by their privacy around what they do (how their source their products/resources, who they donate money to, etc.) and rumors start. It is far more efficient and helpful to be open with that information, as it gets out anyway.

Ironically, in the same paper issue on page 26 there is a an article about When Scandal Knocks..., which includes a story about Jamba Juice and a blog post that inaccurately claimed it had milk in its products, which could have easily been avoided if Jamba Juice had an ingredients listing on its web site.

The Flip Side

There are two flip sides to this. One is the Apple converse, which is a rare example of a company really making a mythic organization out of its privacy. The second is companies really need privacy for some things, but the control of information is often too extreme and is now more harmful than helpful.

Viable Privacy

I have been working on a much longer post looking at the social software/web tools for and in the enterprise. Much of of the extreme openness touted in the new web charge is not a viable reality inside enterprise. There are a myriad of things that need to be private (or still qualify as valid reasons for many). The list include preparations for mergers and acquisitions, securities information dealings (the laws around this are what drive much of the privacy and are out dated), reorganizations (restructuring and layoffs, which organizations that have been open about this have found innovative solutions from the least likely places), personal employee records, as well as contractual reasons (advising or producing products for competitors in the same industry or market segment). Out side of these issues, which normally add up to under 30 to 40% of the whole of the information that flows through an organization, there is a lot of room for openness in-house and to the outside world.

Need for Enterprise Social Tools Grasping Partial Privacy

When we look at the consumer space for social software there are very few consumer tools that grasp social interaction and information sharing on a granular level (Ma.gnolia, Flickr, and the SixApart tools Vox and LiveJournal are the exceptions that always come to mind). But, many of the tools out there that are commonly used as examples of social web tools really fall down when business looks at them and thinks about privacy and selective sociality (small groups). The social web tools all around really need to grow up and improve in this area. As we are seeing the collaboration and social tools evolve to more viable options we start to see their more glaring holes that do not reflect the reality of human social interaction.

Closing the Gap

What we need is for companies to be more open so the marketplace is a more consumer and communicative environment, but we also need our still early social web tools to reflect our social realities that not everything is public and having tools that better fit those needs.


Following Friends Across Walled Gardens

by Thomas Vander Wal in , , , , ,


Phil Gyford makes his plea for a single social network sign-on or as it was stated last week by Jeremy Keith on Twitter, "portable social network with XFN"  [Jeremy posted his More thoughts on portable social networks on his blog.]

The single sign-on is an often heard request these days. Nearly as much as a dashboard, er central interface to monitor all one's friends across the various walled social gardens.

I have been watching ClaimID and Marc's People Aggregator to make huge strides on this front.  Some recent conversations seem to point to others possibly providing this solution.  So far it is the early adopters that are needing these tools, but a recent post by Scott Andrew about people moving from MySpace to Facebook makes me think the cross-garden social tools will have a mainstream appeal in the very near future.  Well, those of us that are veterans of the on-line services of the early 90s will recognize the need as our friends jumped from one closed service to another when there was a 5 dollar a month fee hike.  Your buddies left Prodigy and went where? AOL, Compuserve, Delphi, usenet groups, etc.?

The current web environment makes the keeping up with your friends (if they grant you permission - this really should be the case) easier across closed social services.  Digital identity will be essential as will our trusted list (or grouped list) of friends we want to keep in they loop with our life (lives).


Local InfoCloud as a Responce

by Thomas Vander Wal in , , , , , ,


Ed Vielmetti posted about neighborhoods, networks, communities, online+offline and I had the following comment. My comment seems to fit in as a follow-up post to the Local InfoCloud post (linked below). The online and offline is very important, but so is the individual and the individual interests we have.

There is a huge need for tools that can connect in the neighborhoods. The neighborhood listserve is not the solution, even if some have been successful. The UK's Up My Street was an interesting take on this. There should be potential in something like Yahoo Local, but the people connecting to people is not there.

I have been doing a fair amount of thinking around this as part of the Local InfoCloud (more than just location, but location is very important) as in the Exposing the Local InfoCloud. Each of the components of the Local InfoCloud can be mixed with others and should be mixed.

This summer I have been to more neighborhood cookouts than any time in the past. But the commonality is our kids are around the same age and they interact at the local preschool just up the block. It is the similar/common interests that bring us together. It is the "location", "near in thought" (kids interests), and "affiliation" (school) components that are the aggregation/attraction points.

Part of the problem with every social networking site is they are broad-line friend based and not focussed on facets of our lives. The social network waters are muddied by the broadlines an make it difficult to identify common bonds with people whom we may not yet know, or know from other life contexts. The digital life tools need to start bubbling up the individuals and focuss less on the popularity engines based on people with dissimilar interests.


Location? Location? Where am I?

by Thomas Vander Wal in , , , , , , , , , , ,


I have been traveling more than usual this year to places in the United States and Europe. Some I have been to before and others I have not. Many of the trips are to places for only a few days and are set around meetings, conferences, or speaking engagements. I am often making plans at the last minute or having to make arrangements on the fly as ancillary meetings (not the prime reason I am there) get moved or cancelled. I am often looking for food, coffee, wifi, electronic stores, hardware stores, etc. in a location I am not completely familiar with. I am needing services of the local businessman, but I am not local.

The "Local Services"

You say, "there are many local services". Yes, there are Yahoo Local, Google Local, A9 Yellow Pages search, and other more local guides. But, none of them work on a mobile. There are Google SMS search and Mobile Yahoo, which has search that can tie to your local info, but if I am traveling I most likely have not save where I am looking for options.

Most modern phones know your location, they have to by law in the United States for emergency service calls. The phones do not provide easy access to that location software because the carriers providing the service do not want you to have it for free, they want somebody to pay for that information. If I call information they are not going to tell me where I am, nor the type of service or store I am seeking.

A Hack Finds "Where"

My current hack is to stand in front of a store, which I know the street name and I send the request for information about the place to Google SMS (ritual coffee. san francisco, ca) and I get one important piece of information back, the zip code. The zip code in the United States is the key to getting location information. There is nothing when driving (or actually riding as a passenger, because one never text messages while driving) or walking around that tells you the zip code (I have given up asking strangers on the street the zip code as it is more often than not incorrect). Once I have the zip code I can ask the mobile services for "coffee 94110" and get another place to get coffee and sit down because Ritual Coffee Roasters is utterly packed and already has seat vultures hovering.

Ministry of Silly Steps

Doing this little dance I get options, but it is a few steps that I should never have to take. The information most needed in a local search when mobile is location

Zip It, Zip, Z..

With the zip code I can dump that into my Mobile Yahoo! "new location" and get results. But, even because Yahoo! Mobile knows it is me (they offered me my stored locations (such as Home and Work)) it does not use that information to give me things I have reviewed and stored in Yahoo! Local. In the online version of Yahoo! Local I get reviews from people in my "community" (that really really needs to get a firm understanding of the granular social network), which is often helpful (if I know the person and can adjust my perception because I know how close that person's preferences are to mine on that subject). Sometimes I need an extension cord or an Apple Store (or a good substitution).

Elsewhere: Missing Even Partial Solutions

Additionally, this only works in the United States. The global local versions of Yahoo don't have fleshed out local services that are anything close to what is available in the United States and my "community" (as imperfect of an approach as it is at the moment) is still more helpful at filtering than nothing and I know I have many people in my "community" that have not only been to the same locations I am in, but have reviewed restaurants, local stores, etc. on the web and I want to be able to pull that information back in. Yes, this means the services need to grasp and embrace digital identity to make this work (or just build a social network capable address book that knows who my friend's identities are on various other services and social networking tools where this information may be sitting - not rocket science by any means). I heard some native language services were around, but those would not be fully helpful to me (I think I could get through it however), but if I tried a service that did not work it is not pointing me to one that does (now that would be insanely helpful and I would likely go to the kind service people for everything first as they would point me to just the right place every time).

Ya Beats Goo

Well at least Yahoo! understands there are places outside the United States. Google's services are not there, or any where on the mobile front it seems. In my last trip to Europe nobody knew that Google offered these services, which it seems they do not, in one of the most mobile use intensive cultures in the Western Hemisphere.

Enough

I know, enough. I agree. We need mobile information that works. WiFi is not here everywhere. Even if it were I am not foolish enough to pull out my laptop to try and get a signal and then get the information I need. I have a mobile device with the perfect capability to do just this. Actually there are more than double (if not triple - can not put my fingers on this info) the users with this capability on their mobile than laptop users in the United States (foolishly most laptops do not have locative hardware in them to ease this possibility if it was your last possibility). The technologies are here. Why are we not using them?


Who Are You

by Thomas Vander Wal in , , ,


The person is the center of focus in the Personal InfoCloud, but each person has to interact with others and others' information. It is at this point of interaction where identity becomes important. We use many services around the web. From Epinions to LinkedIn to our personal blogs and beyond we all have many identities and usernames. These "can" be manageable for ourselves, but often our friends value what we say and some of our opinions. We are also interested in what others say, for those opinions and taste that run similar to ours. We also are a verbose bunch writing in volumes over time. We also comment and review on many different services, like restaurant reviews on Yahoo Local, Amazon A9, and other sites.

Now lets say we are going to another city on travel. How do we leverage our friend's opinions and their network of friends reviews? In part we need granular social networks, but we also need digital identity that works.

The digital identity light went off thanks to Brett Lider and it has been a deep interest in the many months since. This may have also been triggered by the Identity Gang and the Gilmore Gang's December 30, 2004 digital identity show. Identity has been a solution fraught with schisms and opposing camps, but it seems to be coming together to some degree at this point. The more I read about digital identity from the Identity Gang and in particular Kim Cameron I realize it is central to the Personal InfoCloud.

At this point we need to modify our address books or use their notes forms to dump in our friends identities and usernames we know about across the various systems we interact with them on. This is unmanageable and not easily usable at this point. But if we want to follow John Doe and his reviews we have to know he uses jdoe, doe, paininmaximus, and 33482aad across various systems. We have to know what systems he is what identifier/username. Now lets say he gave us a card/key that he managed what we knew and he could update as he changed his information. We could then use that information in our own aggregation of information he has posted