Personal InfoCloud

Lessons To Learn

Sadly, the new social functionality has broken much of worked well as an ambient social tool. More problematic was LinkedIn did not seem to grasp what it had: so to build on top of very good start, but it seemingly looked at Facebook for inspiration, but Facebook does not seem to be aware of good social interaction design practices.

When building social tools for broad audiences (more than 3,000 people) — which open services on the web are — there is a progression of 3 things that must be accounted for in the planning stages: 1) Velocity; 2) Volume; 3) Relevance.

As social tools start getting used they go through a progression one of them are these three stages of concern. Velocity of information is how quickly information is added by the community and has turn over on the in the frameworks. Volume is the mass of information that accumulates over time that will force how information is shared, found, and used. Relevance becomes essential when there is large volume and filtering is needed for information flows and for allowing people using the service to have a manageable stream of information that is relevant to their needs.

Many social services can go through these three stages in a few short months if they have 50,000 users or more. LinkedIn does not seem to have considered getting to and beyond the first stage in their planning.

Social Interaction at Scale & Volume

As LinkedIn has added social features they have created more streams of information in their flow. More streams lead to more velocity of information. This can be good if the basic concepts for understanding monitoring these streams as well as providing methods for moving things out of the flow so they can be acted upon or set into a personal task flow.

It seems as if the new social features are aimed at the roughly 80 percent that have 100 or fewer connections, not the moderate or heavy connectors who are the unpaid evangelists that have helped LinkedIn grow. Not understanding the value various segments bring to a service and how to satisfy those groups is rather short sighted

Social Tsunami on Homepage

The one thing that started the frustration with LinkedIn’s shift was the flood of unrelated social items on the homepage. Much of the social content shared is personal ID focused and not group or work focused (even when shared in groups or work related settings - a quick look at activity summaries regularly shows this).

One of the task flows I had with LinkedIn was to accept a connection or get notification of a connection then go to their profile page and download their vCard. The social tsunami that took over the front page of LinkedIn made that task all-but impossible. Part of it is the velocity of information running through the front page for connections increased, velocity the design did not account for.

Additionally, the new social components started eating up valuable real estate on that page and had no simple interaction design convention for minimizing, hiding, or turning off that module of functionality from the page.

Eventually the ability to turn off notifications to the social tools was added to the Settings page, but there is no notification of that functionality on where the problem exists, the pages where this container shows up (we learned this in software design in the early 90s). Also problematic is the social elements are clustered by task/tool relevance and not person or subject. Including pivots could greatly improve this as well as allow for shifting context by the person using LinkedIn.

Focal prioritization is essential to include in initial planning, as this becomes critical when dealing with the relevance stage or even handling a scaling volume of information. Each person using a service is going to have a slightly different set of priorities for relevance and focus. This is going to require some malleability of the system interface to allow for personal optimization of their relevance and streams.

This is not emergent behavior but the reality of what happens when systems scale. LinkedIn is built “;like a classic chamber meeting where networking is orchestrated”;, as stated by Margaret Rosas. Sadly, LinkedIn is not built for flexibility that is needed as systems scale to or beyond the volume stage. It is built as if this was a surprise, which prior to 18 months ago LinkedIn’s careful approach was much smoother with their growth of features and functionality.

LinkedIn changed its layout and structure of its pages to account for the coming new functionality, which is quite smart. But it did so in a manner that seemed to consider all notifications and functionality should have the same focus.

If you remove notifications, there is no ambient notification to let you know there is really any activity. The front page is part portal and part dashboard, but the distinct concepts around these two approaches seem to have baffled the interaction designers and developers.

LinkedIn: Social Node or Social Hub?

LinkedIn also seems quite schizophrenic as to its social purpose. It has built part of the social framework as if the rest of the web only allowed limited interaction with it, which it would make it just a node on a network. This destination framework does not account for people having any other service that provides social features that could easily be shared in or out.

The other side of LinkedIn is a hub, which information flows though. Inbound status messages from other services show up in LinkedIn’s pages as to the “applications”, but using connecting identity in a manner that permits not having Twitter messages I read elsewhere show up in LinkedIn would be more than helpful (yes, part of this is OAuth, which Twitter and many other have not deemed valuable yet (come on Twitter this is not rocket surgery). The applications and information it allows in is limited to a relatively small number of services. Having a small number of services integrated should allow for contextual relevance of the objects, but that would be assuming again LinkedIn was well thought through. This interaction with services would also benefit form LinkedIn offering OpenID as well as OAuth integration to ease the pain and security.

LinkedIn does not have an open API as of yet (this should have happened when they launched status and some other social elements). The LinkedIn API for status would allow LinkedIn to be a sharing out hub as well as the partially capable in-bound hub it already is. LinkedIn is a business focussed social environment, but has not realized its DNA is business based and there are task flows and workflows to enable that would make a lot of sense.

LinkedIn Forgot the “Me” in Social

All social begins with me. Social interaction is about an individuals intentions, actions, and their activities. What things a person wants to share with others and how interact with is one part of the social framework. Another other is consumption and working with the flows of content generated by others. LinkedIn did a decent job with flow until it started adding the more social features in the last 18 months. What LinkedIn did know (focus and purpose) they now show little grasp of understanding as their features have created more flow and more velocity for the information ebbing through the service with no planning for it. It takes very little understanding of social tools to know that this will likely happen and there are interaction elements that are going to be required to handle this, for example moving things out of the flow.

Many people want to see those they have just connected with, things they just published/shared and responses. There is also the desire to hold on to things that are relevant to the individual. This holding on to things requires a means to favorite or put it in place where things can be collected and worked on later. These things could be single comments in group discussions, people’s names/profiles who are surfacing, notifications, etc.

With the velocity of information increasing in LinkedIn the capability to perform a task and drop back into the flow where you were is gone. Any decent interaction designer for social tools knows this reality and had a stack of solutions to set in place from the outset.

Social Context in Groups

The math of social software for people is the mostly the one-to-one relationships and being able to see those. But, social software occasionally is about communicating to groups.

LinkedIn added group discussions, but did this as if the last 10 to 15 years in forums and groupware platforms never existed. The group discussions are not threaded nor do they offer the option to turn on threading for the discussions (this has been default for off-the-shelf forums for over 8 years at least). Also lacking is the capability to hold on to and collect valuable items found in discussions, let alone a means to personally contextualize them.

Another thing LinkedIn fails to grasp is contextual relationships to people in the discussions. For example, if someone I know has started or commented in a group discussion the service should highlight this. There is a potentially higher social contextual relevance for that piece of information. When information starts turning from a stream into a flood this becomes insanely important.

Once this reality of contextualizing is realized, there are a couple of options that are likely to be needed quite quickly after. One is adding new people in the discussion that we interact with; this context could be surfaced in the discussion or used to augment the rational surfaced in the recommendations.

Email from groups should not be from the organization name of the group as that looks like it is from the organization. I get official information from organizations, but lacking the understanding of contextual information for e-mail makes an even greater mess of e-mail and group interactions when this is lazily designed. The “from” should begin with LinkedIn group or some other notation.

Context for Events

When LinkedIn added events, I started getting invitations to attend them. But, the wording of the invites made it sound like they were personal invitations, which is not the context they were intended. It took quite a few rather embarrassing e-mails for many events, if they were really requesting my attendance or if it was just an announcement of the event. Understanding a modicum of social interaction and social etiquette would have saved those embarrassing e-mails.

Events also launched with many bugs (many have been ironed out, but most were of the rather blatant variety). One downside of events is there are already an over abundance of event tools, which work rather well (this is a really tough tool set to get right and build). Nearly everybody I talk to has wondered why LinkedIn did not use something like Confabb to license it or buy it (there are many event services available), rather than using their own resources on something that is not up to the level of competing products. Lastly, with regard to events, while the recommendation for connections is good in LinkedIn, the recommendations for events is absolutely horrid. If that is who LinkedIn thinks I am I need a new service now.

Models for Messaging Flows

One of the things that has been flawed in LinkedIn for quite some time is messaging flows. I liked that they pushed messaging out into e-mail and I could respond to a person from my e-mail. One thing that is missing is LinkedIn not updating their messaging flows. Looking in LinkedIn it is quite often impossible to sort out. When I stated I continually have this problem, Jess Leccetti stated, “I’ve had that exact problem! I thought it was my comp being buggy!” Messaging across various media channels is tough and most often fractured. But, when offering a solution it is important to get it right.

Profile Comments Go In...

Finally LinkedIn added the capability to for people using the service to add their own private comments on to other’s profiles. This is a great addition as it allows the means to add context to files. Sadly, it does not seem to surface that information in any other manner other than going to the individual pages.

This lack of functionality outside each profile page is really mind blowing, as it leaves out the capability for using it for tagging, contextual grouping, search aggregation, and use these aggregations for sharing up dates or filtering what is shared. There are emergent activities that could evolve out of these functionalities, but again this seems to not be well thought through.

One approach is a nice simple personal tagging or labeling interaction layer with clickable aggregation interface option. This would allow simply applying glue to personally thread items together through light aggregation. The current comment system only creates islands of context that have chasms between it and other relevant or related items.

Next Steps

LinkedIn needs to get some people in that grasp social interaction design. They purportedly have some, but I am not sure they have influence or the depth of knowledge needed (either is problematic). The LinkedIn service seems to be proof something is horribly wrong along these lines.

LinkedIn also seems to be a victim of not sorting-out what it wants to be. If it wants to be a Facebook for business, the route they are taking is not going to work well for the business users as it is greatly lacking solid functionality and cohesive interaction design with task flows enabled. LinkedIn needs to be LinkedIn and not a Facebook for business.

As many on Twitter have stated, one seemingly viable option is LinkedIn’s social additions of the last 18 months should all be thrown out and simply start over. The only piece that seems to have much positive feedback is the Q&A section, which is not something that I have interest in, but seems to work passably for others.

More coherently, a real reality check is needed at LinkedIn. They must to stop adding features and functionality until they learn to fix what they have added. They need to begin with understanding how social interaction happens, how it scales, and how people need it to work at scale. Stop looking at Facebook for what features to add. LinkedIn has some deep value as a work and business focussed social site, but that is going to require a different focus that what has been applied in the last year to 18 months.

I have deep fear that LinkedIn views what is happening is emergent (emergence happens when things are used in an unpredictable manner: whether wholly unpredictable or unpredictable in that context). What is happening in LinkedIn is not emergent. It is quite predictable: This is what happens at scale with social systems and their information flows.

A grasp of social systems and their uses at various levels of scale (and potential for various interactions and needs) is really needed at LinkedIn. The slowness to act (or, sadly, react as if this was an unknown potential) and fix what they have is not a great sign of encouragement for the organization. Hopefully having Reid Hoffman back as CEO and with Jeff Wiener as President can pull this into focus and set things on a sane path.

Social Tools Need to Embrace Granularity

What we have is partial likes in others and their interests and offerings. Our social tools have yet to grasp this and the few that do have only taken small steps to get there (I am rather impressed with Jaiku and their granular listening capability for their feed aggregation, which should be the starting point for all feed aggregators). Part of grasping the problem is a lack of quickly understanding the complexity, which leads to deconstructing and getting to two variables: 1) people (their identities online and their personas on various services) and 2) interests. These two elements and their combinations can (hopefully) be seen in the quick annotated video of one of my slides I have been using in presentations and workshops lately.

Showing Granular Social Network

Granular Social Network from Thomas Vander Wal on Vimeo.

The Granular Social Network begins with one person, lets take the self, and the various interest we have. In the example I am using just five elements of interest (work, music, movies, food, and biking). These are interest we have and share information about that we create or find. This sharing may be on one service or across many services and digital environments. The interests are taken as a whole as they make up our interests (most of us have more interests than five and we have various degrees of interest, but I am leaving that out for the sake of simplicity).

Connections with Others

Our digital social lives contain our interests, but as it is social it contains other people who are our contacts (friends is presumptive and gets in the way of understanding). These contacts have and share some interests in common with us. But, rarely do the share all of the same interest, let alone share the same perspective on these interests.

Mapping Interests with Contacts

But, we see when we map the interests across just six contacts that this lack of fully compatible interests makes things a wee bit more complicated than just a simple broadline friend. Even Facebook and their touted social graph does not come close to grasping this granularity as it is still a clumsy tool for sharing, finding, claiming, and capturing this granularity. If we think about trying a new service that we enjoy around music we can not easily group and capture then try to identify the people we are connected to on that new service from a service like Facebook, but using another service focussed on that interest area it could be a little easier.

When we start mapping our own interest back to the interest that other have quickly see that it is even more complicated. We may not have the same reciprocal interest in the same thing or same perception or context as the people we connect to. I illustrate with the first contact in yellow that we have interest in what they share about work or their interest in work, even though they are not stating or sharing that information publicly or even in selective social means. We may e-mail, chat in IM or talk face to face about work and would like to work with them in some manner. We want to follow what they share and share with them in a closer manner and that is what this visual relationship intends to mean. As we move across the connections we see that the reciprocal relationships are not always consistent. We do not always want to listen to all those who are sharing things, with use or the social collective in a service or even across services.

Focus On One Interest

Taking the complexity and noise out of the visualization the focus is placed on just music. We can easily see that there are four of our six contacts that have interest in music and are sharing their interest out. But, for various reasons we only have interests in what two of the four contacts share out. This relationship is not capturing what interest our contacts have in what we are sharing, it only captures what they share out.

Moving Social Connections Forward

Grasping this as a relatively simple representation of Granular Social Networks allows for us to begin to think about the social tools we are building. They need to start accounting for our granular interests. The Facebook groups as well as listserves and other group lists need to grasp the nature of individuals interests and provide the means to explicitly or implicitly start to understand and use these as filter options over time. When we are discussing portable social networks this understanding has be understood and the move toward embracing this understanding taken forward and enabled in the tools we build. The portable social network as well as social graph begin to have a really good value when the who is tied with what and why of interest. We are not there yet and I have rarely seen or heard these elements mentioned in the discussions.

One area of social tools where I see this value beginning to surface in through tagging for individuals to start to state (personally I see this as a private or closed declaration that only the person tagging see with the option of sharing with the person being tagged, or at least have this capability) the reasons for interest. But, when I look at tools like Last.fm I am not seeing this really taking off and I hear people talking about not fully understanding tagging as as it sometimes narrows the interest too narrowly. It is all an area for exploration and growth in understanding, but digital social tools, for them to have more value for following and filtering the flows in more manageable ways need to more in grasping this more granular understanding of social interaction between people in a digital space.

Selective Sociality of Villages

The piece that resonated along similar threads to Robert's post is Susan Mernit's "Twitter & Friend Feed: The Pleasure of Permissions". Susan's post brings to light the value of knowing who you are sharing information with and likes the private or permission-based options that both Twitter and FriendFeed offer. This selective sociality as known Local InfoCloud of people and resources that are trusted and known, which we use as resources. In this case it is not only those with whom we listen to and query, but those with whom we share. This knowing who somebody is (to some degree) adds comfort, which is very much like Robert Patterson and Luis Suarez#039; villages where people know each other and there is a lot of transparency. Having pockets where our social armor is down and we can be free to share and participate in our lives with others we know and are familiar to us is valuable.

I am found these two pieces quite comforting as they reflect much of what I see in the physical community around me as well as the work environments I interact with of clients and collaborators. The one social web service I have kept rather private is Twitter and I really want to know who someone is before I will accept them as a connection. This has given me much freedom to share silly (down right stupid - in a humorous way) observations and statements. This is something I hear from other adults around kids playgrounds and practices of having more select social interactions on line in the services and really wanting to connect with people whom they share interests and most often have known (or followed/listened to) for sometime before formally connecting. Most often these people want to connect with the same people on various services they are trying out, based on recommendation (and often are leaving a service as their friends are no longer there or the service does not meet their needs) of people whom they trust. This is the core of the masses who have access and are not early adopters, but have some comfort with the web and computers and likely make up 80 to 90 percent of web users.

Inline Messaging

Here is where things are horribly broken. The closed services (except Twitter) will let you know you have a message on their service on your choice of communication channel (e-mail, SMS, or RSS), but not all offer all options. When a message arrives for you in the service the service pings you in the communication channel to let you know you have a message. But, rather than give you the message it points you back to the website to the message (Facebook does provide SMS chunked messages, but not e-mail). This means they are sending a message to a platform that works really well for messaging, just to let you know you have a message, but not deliver that message. This adds extra steps for the people using the service, rather than making a simple streamlined service that truly connects people.

Part of this broken interaction is driven by Americans building these services and having desktop-centric and web views and forgetting mobile is not only a viable platform for messaging, but the most widely used platform around the globe. I do not think the iPhone, which have been purchased by the owners and developers of these services, will help as the iPhone is an elite tool, that is not like the messaging experience for the hundreds of millions of mobile users around the globe. Developers not building or considering services for people to use on the devices or application of their choice is rather broken development these days. Google gets it with Google Gears and their mobile efforts as does Yahoo with its Yahoo Mobile services and other cross platform efforts.

Broken Interaction Means More Money?

I understand the reasoning behind the services adding steps and making the experience painful, it is seen as money in their pockets through pushing ads. The web is a relatively means of tracking and delivering ads, which translates into money. But, inflicting unneeded pain on their customers can not be driven by money. Pain on customers will only push them away and leave them with fewer people to look at the ads. I am not advocating giving up advertising, but moving ads into the other channels or building solutions that deliver the messages to people who want the messages and not just notification they have a message.

These services were somewhat annoying, but they have value in the services to keep somebody going back. When Pownce arrived on the scene a month or so ago, it included the broken messaging, but did not include mobile or RSS feeds. Pownce only provides e-mail notifications, but they only point you back to the site. That is about as broken as it gets for a messaging and status service. Pownce is a beautiful interface, with some lightweight sharing options and the ability to build groups, and it has a lightweight desktop applications built on Adobe AIR. The AIR version of Pownce is not robust enough with messaging to be fully useful. Pownce is still relatively early in its development, but they have a lot of fixing of things that are made much harder than they should be for consuming information. They include Microfomats on their pages, where they make sense, but they are missing the step of ease of use for regular people of dropping that content into their related applications (putting a small button on the item with the microformat that converts the content is drastically needed for ease of use). Pownce has some of the checkboxes checked and some good ideas, but the execution of far from there at the moment. They really need to focus on ease of use. If this is done maybe people will comeback and use it.

Good Examples

So who does this well? Twitter has been doing this really well and Jaiku does this really well on Nokia Series60 phones (after the first version Series60). Real cross platform and cross channel communication is the wave of right now for those thinking of developing tools with great adoption. The great adoption is viable as this starts solving technology pain points that real people are experiencing and more will be experiencing in the near future. (Providing a solution to refindability is the technology pain point that del.icio.us solved.) The telecoms really need to be paying attention to this as do the players in all messaging services. From work conversations and attendees to the Personal InfoCloud presentation, they are beginning to get the person wants and needs to be in control of their information across devices and services.

Twitter is a great bridge between web and mobile messaging. It also has some killer features that add to this ease of use and adoption like favorites, friends only, direct messaging, and feeds. Twitter gets messaging more than any other service at the moment. There are things Twitter needs, such as groups (selective messaging) and an easier means of finding friends, or as they are now appropriately calling it, people to follow.

Can we not all catch up to today's messaging needs?

Now With More Complexity

But, what we have today is even more complex and difficult as we are often holding conversation across many of these technologies. The communication streams (the back and forth communication between two or more people) are now often not contained in on communication channel (channel is the flavor or medium used to communicate, such as AIM, SMS, Twitter, e-mail, mobile phone, etc.).

We are seeing our communications move across channels, which can be good as this is fluid and keeping with our digital presence. More often than not we are seeing our communication streams fracture across channels. This fracturing becomes really apparent when we are trying to reconstruct our communication stream. I am finding this fracturing and attempting to stitch the stream back together becoming more and more common as for those who are moving into and across many applications and devices with their own messaging systems.

The communication streams fracture as we pick-up an idea or need from Twitter, then direct respond in Twitter that moves it to SMS, the SMS text message is responded back to in regular SMS outside of Twitter, a few volleys back and forth in SMS text, then one person leaves a voicemail, it is responded to in an e-mail, there are two responses back and forth in e-mail, an hour later both people are on Skype and chat there, in Skype chat they decide to meet in person.

Why Do We Want to Stitch the Communication Stream Together?

When they meet there is a little confusion over there being no written overview and guide. Both parties are sure they talked about it, but have different understandings of what was agreed upon. Having the communication fractured across channels makes reconstruction of the conversation problematic today. The conversation needs to be stitched back together using time stamps to reconstruct everything [the misunderstanding revolved around recommendations as one person understands that to mean a written document and the other it does not mean that].

Increasingly the reality of our personal and professional lives is this cross channel communication stream. Some want to limit the problem by keeping to just one channel through the process. While this is well intentioned it does not meet reality of today. Increasingly, the informal networking leads to meaningful conversations, but the conversations drifts across channels and mediums. Pushing a natural flow, as it currently stands, does not seem to be the best solution in the long run.

Why Does Conversation Drift Across Channels?

There are a few reasons conversations drift across channels and mediums. One reason is presence as when two people notice proximity on a channel they will use that channel to communicate. When a person is seen as present, by availability or recently posting a message in the service, it can be a prompt to communicate. Many times when the conversation starts in a presence channel it will move to another channel or medium. This shift can be driven by personal preference or putting the conversation in a medium or channel that is more conducive for the conversation style between people involved. Some people have a preferred medium for all their conversations, such as text messaging (SMS), e-mail, voice on phone, video chat, IM, etc.. While other people have a preferred medium for certain types of conversation, like quick and short questions on SMS, long single responses in e-mail, and extended conversations in IM. Some people prefer to keep their short messages in the channel where they begin, such as conversations that start in Facebook may stay there. While other people do not pay attention to message or conversation length and prefer conversations in one channel over others.

Solving the Fractured Communication Across Channels

Since there are more than a few reasons for the fractured communications to occur it is something that needs resolution. One solution is making all conversations open and use public APIs for the tools to pull the conversations together. This may be the quickest means to get to capturing and stitching the conversation thread back together today. While viable there are many conversations in our lives that we do not want public for one reason or many.

Another solution is to try to keep your conversations in channels that we can capture for our own use (optimally this should be easily sharable with the person we had the conversation with, while still remaining private). This may be where we should be heading in the near future. Tools like Twitter have become a bridge between web and SMS, which allows us to capture SMS conversations in an interface that can be easily pointed to and stitched back together with other parts of a conversation. E-mail is relatively easy to thread, if done in a web interface and/or with some tagging to pull pieces in from across different e-mail addresses. Skype chat also allows for SMS interactions and allows for them to be captured, searched, and pulled back together. IM conversations can easily be saved out and often each item is time stamped for easy stitching. VoIP conversations are often easily recorded (we are asking permission first, right?) and can be transcribed by hand accurately or be transcribed relatively accurately via speech-to-text tools. Voice-mail can now be captured and threaded using speech-to-text services or even is pushed as an attachment into e-mail in services as (and similar to) JConnect.

Who Will Make This Effortless?

There are three types of service that are or should be building this stitching together the fractured communications across channels into one threaded stream. I see tools that are already stitching out public (or partially public) lifestreams into one flow as one player in this pre-emergent market (Facebook, Jaiku, etc.). The other public player would be telecoms (or network provider) companies providing this as a service as they currently are providing some of these services, but as their markets get lost to VoIP, e-mail, on-line community messaging, Second Life, etc., they need to provide a service that keeps them viable (regulation is not a viable solution in the long run). Lastly, for those that do not trust or want their conversation streams in others hands the personally controlled application will become a solutions, it seems that Skype could be on its way to providing this.

Is There Demand Yet?

I am regularly fielding questions along these lines from enterprise as they are trying to deal with these issues for employees who have lost or can not put their hands on vital customer conversations or essential bits of information that can make the difference in delivering what their customers expect from them. Many have been using Cisco networking solutions that have some of these capabilities, but still not providing a catch all. I am getting queries from various telecom companies as they see reflections of where they would like to be providing tools in a Come to Me Web or facilitating bits of the Personal InfoCloud. I am getting requests from many professionals that want this type of solution for their lives. I am also getting queries from many who are considering building these tools, or pieces of them.

Some of us need these solutions now. Nearly all of us will need these solutions in the very near future.

The Great Nuggets

I will (uncharacteristically) begin with the things I like with the analysis that I have not seen elsewhere. It is also appropriate to mention I love LibraryThing as it is really an innovative approach to organization of one's own things and there are many things it does that are done very well.

The first bit that is really good is the difference on LibraryThing in the use of tags based on the number of items in a person's collection.  The more items the great the probability that person will use tags.  The more items in a persons collections the greater number of tags they use.  Tim states it as,

Tags work best when they're about memory, so tagging makes the most sense when you have a lot of something to remember. On LibraryThing, members with under 50 books seldom tag, but users with 200 or more usually do. When you get right down to it, few of us need to remember 200 books on Amazon.

This is really good, because it points out that there is a trend toward people using tags when their collection of objects grows. Two of the reasons people tag (of many) is to aggregate as well as sort information.  It is building hooks to pull things together that are similar. It is a flat organizational structure.

Ironically, the other really good point is the core of why I think the analysis is really flawed:

In fairness, Amazon didn't give tagging a lot of prominence. Tags were stuck in the middle of their ever-lengthening book page—one section for adding your own tags, another for showing others' tags. They didn$#039;t push them very hard.

Oddly this bit is one of many reasons why the analysis does not hold up. It is really important to grasp that there are many different reasons why people tag. For tagging to work it needs to scale, but getting people to tag takes work through explanation and examples. Not only did tagging not have prevalence on the page, but tagging was not in all user interfaces as of 7 months ago on US Amazon site.

Apple to Oranges

The major problem with the analysis is the comparison is being made on 2 vastly different products, with 2 vastly different uses, with 2 vastly different points of focus.

First Order of Functionality

The first order of functionality for both sites/products is vastly different.  LibraryThing puts its functionality focus on people organizing their book (and other media) collections online.  Tagging is central to what LibraryThing does and its means to enable organization. It is for people to claim and organize what they own, share that with others, and find related materials.

Amazon has it first order of functionality around people buying products they want. From this point Amazon uses other tools to ensure the person is buying what they want or believes they want.  Tagging is not central to the first order of functionality for Amazon, but it can and is being used to help people find products. People are not organizing their collection on Amazon, well as of yet.

Options for Similar Input

The two sites compared have very different sets of tools and the numbers of options are vastly different.  The length of time the other options have been in existence is nearly an order of magnitude different. LibraryThing has puts its primary focus on tagging and ratings, while it offers the ability to have people write a review very few people people seem to write them (I would not dare say reviews are a failure on LibraryThing as they are helpful). With the focus on tagging and ratings and with few other options to manage the collection it is no wonder tagging took off.

Amazon has been around for much longer and has offered reviews, ratings, wish lists, shopping lists, sharing purchases, listmania, registries, customer discussions about the product, search suggestions, share with friends, wiki pages for products, etc.  The reviews and ratings are the two things that people using Amazon are most familiar with as they have been around for many years. All of these elements are social web means for people to comment, track (all reviews on on the identity page for the individual).  Tagging has not had anything close to the prominence on the Amazon product pages or the site as it does in LibraryThing, due to the other options and the years of habitual use of the other options.

Amazon, only in recent months, has given tagging a little more prominence on the product pages. Amazon is still iterating drastically what they are doing with tags as screen captures over the past six months (even going back 18 months) will show.  Their current iterations are really are moving toward utterly brilliant interactions in some areas (I have done a couple hour long impromptu presentations on the depth of what Amazon is doing with tagging and how it seems to be making a difference - this may become another post, but it will be really long even for my norm).

Availability of Tagging in the Tools

One of the claims that bothered me most was the claim that Amazon and LibraryThing has roughly had tagging for the same amount of time.  As of 7 months ago I had people showing me they did not have tagging functionality in the Amazon page interface. Tagging on Amazon is only available on the United States Amazon site, as far as I can tell. Claiming the date that when tagging first appeared on the Amazon product pages for some of us, it even seems to have disappeared for a while, is not really a solid claim to make. Again it is another really weak comparison.

Related to the last major point made, options for similar input, the adoption rates for new tools for people using Amazon has been relatively slow. But, the tools quickly make difference and become valuable. One example of tools is Amazon's Listmania.  This has been on the Amazon site, rather prominently, for many years but many people (the people whom I am talk to about Amazon tagging and their uses of it) have not noticed it. This is directly related to the other features that do know. But, for Amazon the Listmania has been a really successful tool for increased sales.

Mis-understanding of DefectiveByDesign

The analysis claimed the "DefectiveByDesign was spam by a few people. But, the 501 people currently using the tag 5809 times across 1063 products may have a different view.  The dominant use of the tag is for items that have copy protection and/or DRM. Talking to people who do use Amazon tags, this is often one of the tags people find very helpful as it is a tag for missing metadata about a product.  None of the product pages state the OS will not let you copy your own media you have bought or the CD you bought will not let you copy it so to put it on your iPod (which is where you listen to all of your music). This tag is not spam and for many people I have talked with about tagging on Amazon has been how people found out about Amazon tags and the value it has for them personally.

Tagging in Amazon is True Longtail Commerce

One of the things that Amazon is doing really well with tagging, not covered in the analysis piece at all, is helping people interested in the LongTail products connect and share recommendations.  Because of changes in the interface that Amazon has been iterating through (there are many more potential possibilities where Amazon tagging could really soar) more people are using the tagging to connect products not normally connected.  One tag that has been around for a while is "discomusic dot com top pick", which is a great hook for the longtail as it is the picks from a site and putting them in a context that helps join disparate items together. The Amazon tag page for discomusic dot com top pick is a great jumping off point for beginning to grasp the insane power that one variation of use for tags has within commerce.

Personal Planets

Why a personal planet? We have an incredible amount of information that passes before out eyes and that is generated by our simple actions.  Emily did a great job showing the breadth of feeds generated.  This seems a simple thin thread.  What if we could quickly scan that thread and annotate it to make it easier to refind.

Planets are relatively easy to build and it should be easy to share for others to build upon. Personally, I am really surprised there are not thousands of these out there already.  Now to start tinkering a wee bit this week.