In this post, I will share some of ideas and findings that kept me and my team busy lately. Part of the text below has been co-authored so I would like to associate Du Juan, Han-Leon Chia and Sunny Lim to this post. It voluntarily is focused on system design.
We address in a renovated way, that Patrick Lambe kindly mentioned earlier today, the question of on-line communities with sustainability in mind.
In order to do that we have decided to face reality and make ours the following assumptions:
• No matter trendy on-line communities of practices or interests are, no matter what their manager say or report, most of them fail.
• The world in which on-line communities evolve today is very different from the one in which they emerged about ten years ago. The social web emerged and thrives, but also serves as a standard.
• People are busy, lazy, self-centered and non-cooperative by nature.
1 – Elements to consider for design
There are two significant factors that surfaced recently and that impact the development of online communities:
• The algorithmic growth of information
• The rate of engagement is low
1.1 – The growth of information is algorithmic
“With a compound annual growth rate of almost 60%, the digital universe is growing faster and is projected to be nearly 1.8 zettabytes (1,800 exabytes) in 2011, a 10-fold increase over the next five years.” A typical example is the email system, where duplication is what concretely happens behind the scene and results very costly:
Source: “The Diverse and Exploding Digital Universe – An Updated Forecast of Worldwide Information Growth Through 2011″, IDC, March 2008, by John F. Gantz, Christopher Chute, Alex Manfrediz, Stephen Minton, David Reinsel, Wolfgang Schlichting, Anna Toncheva._
This implies that to be sustainable, an online community platform needs to:
• Deal with a constant flow of information, and not a static stock of information;
• Implement a granular approach to information delivery (Community > Member) to optimize/reduce the quantity of information and prevent infobesity. Information obesity happens when people are cognitively overwhelmed by the flow of information. The natural defense is that they stop using the system or throw data away.
• Display the right information to the relevant people, instead of blindly and passively making it accessible “somewhere”.
1.2 The rate of engagement is low (1/2)
According to recent data, most online communities abide by the “1-9-90 rule”.
90% of users are lurkers who never contribute, 9% of users contribute a little, and 1% of users account for almost all the action. This is:
• 1% of users participate a lot and account for the most contributions: they champion subjects and thrive on fast, up-to-the-minute content updates.
• 9% of users contribute from time to time, but other priorities dominate their time.
• 90% of users are lurkers who read or observe, but never contribute.
The above data partially explains why previous approaches of communities are often not as successful as anticipated: they address participants, eventually contributors, not lurkers. In fact, they unrealistically assume 100% participation. They have no features to engage, nor metrics to account for lurkers. Consequently, in the final analysis, the previous approach ignores 90% of the target population, leaving all its potential untapped.
This implies that to be sustainable an online community needs to engage lurkers. Participants and contributors are already motivated and this should be increased by the network effect created by the participation of lurkers. Engaging lurkers can be done by specific services where members can read or observe and complemented with simple and implicit features to have these people move to a higher level of engagement. This means we pay attention to lowering the cost of co-operation / collaboration as well as creating opportunities for initiating conversations. This is the required step not only to engage members, but also to initiate a positive flow and growth of knowledge for each member. This shall in turn attract new members and increase the base of lurkers.
1.3 The rate of engagement is low (2/2)
Another element that explains low engagement is the dispersion of resources or the distance of online communities to the center of attention. The more mouse clicks, the higher the abandon rate. This relates to the empirical observation of behavior made by American computer scientist Calvin Mooers in 1959 known as “Mooers Law”:
“An information retrieval system will tend not to be used whenever it is more painful and troublesome for a customer to have information than for him not to have it.
Where an information retrieval system tends not to be used, a more capable information retrieval system may tend to be used even less.”
In April, I have had the opportunity to attend a presentation of a knowledge management platform at a Singapore Government Agency, by the Deputy Directory for Knowledge Management (and IKMS member). The one element for explaining the success was that KM was embedded in work processes, i.e. in the constant center of attention of employees.
Back in Europe, I worked for Headshift with the great Tarik Lebtahi at Dassault Systemes, a € 1.335b software company, on the design and implementation of an intranet in which free forming communities are embedded into a set of features facilitating internal communication and daily life in the workplace (meeting room booking, holiday management, food & beverages, concierge, etc). The intranet rapidly became a key element of the center of attention of the employees, in addition to emails.
It is therefore possible to make the assumption that to be sustainable an online community needs ideally to be embedded in or become the center of attention and that information needs to be easily accessible and manipulable. This does not necessarily call for full integration of functions and services, but at least integration at the level of the user interface. What is important is to create an interface were users can visually grasp different information, to make connections that generate new ideas, and circulate it from one container to another to use it.
1.4 The classical model of information retrieval is partial
The classical model of information retrieval is partial and usually gives room in real life to browsing-based “berrypicking” search.
As Marcia J. Bates notes:
• Typical search queries are not static, but rather evolve
• Searchers commonly gather information in bits and pieces instead of in one grand best retrieved set
• Searchers use a wide variety of search techniques which extend beyond those commonly associated with bibliographic databases
• Searchers use a wide variety of sources other than bibliographic databases.
Source: Bates, M.J. (1989), The design of browsing and berrypicking techniques for the online search interface. _Online Review_, 13. 401-424.
This implies we have to multiply the ways of accessing content. The display of “most viewed”, “most commented”, “best rated”, “recently commented”, “recently updated” are current popular ways of multiplying access to content. Features facilitating the navigation between information and people (and vice versa) also ease the findability and information retrieval.
1.5 Additional elements to consider
The above can be related to the notions of:
• Networked Individualism
For Barry Wellman, who coined the term, the idea reflects simply the changing communication patterns of people, who no longer rely on a small number of localized communities (workplace, home, civic association, etc) for social support, but on a much larger number of networks, increasingly geographically dispersed. Thus, people are highly individualized in terms of the combination of networks they maintain; yet their individuality evolves within and through these networks. Wellman’s notion remains firmly grounded within a quantitative social network analysis.
• Social Translucence
This term that was proposed by Thomas Erickson and Wendy A. Kellogg to refer to “digital systems that make perceptually-based social cues visible to their users”.
• Mass Customization
“Mass customization is enabling a customer to decide the exact specification or personal attributes of a product or service, and have that product or service supplied to them at a price close to that for an ordinary mass produced alternative”. In an online environment mass customization helps to empower users, but at the same time not loose control and consistence of both infrastructure and services.
2 Conceptual basis for design: the Knowledge Sharing Stack
Based on the above and as per today, a successful candidate would be an user-customizable digital platform, that connects people and information on a shared topic of attention, and with a focus on the following: Exposition / Socialization / Participation / Collaboration / Customization.
This is the dynamic delivery of new content coming from various sources.
It can principally take the form of an RSS feed aggregator. As a result we can link to different internal and external sources and services.
This is an effortless way of referencing content and making it available to others.
It can take the form of social bookmarking and voting system.
• Participation / Collaboration
This enables easy and fuss-free ways of publishing via personal or collective tools.
It can take the form of a blog or a wiki.
It would be complemented by an archive / memorization system in order to store and preserve information in the form of documents or records.
• (End User) Customization
This is a simple way of organizing the display of available information as well as personalizing part of the displayed information.
It usually is an AJAX-based personal start page and it is currently an effective way of implementing mass customization. It has not to be confused with profiling, as personalization is often referred as profiling.
The stack of layers can be expressed like so:
A typical service would:
• Aggregate a group of people with shared interest(s), formalizing the community.
• Expose each individual to information coming from different sources. Instead of relying on deliberate searches for information in known sources, this service brings relevant information to them, from sources expected and serendipitous, which creates opportunities and material for discovery and conversations. The information delivery should be specific to the user, not a group of people. One way to do this is expose information for the group and let each user complement for his/her own preferences/needs.
• Expose the personal activity of each individual via the information: visiting, bookmarking, voting, commenting; so that they can contribute effortlessly and unconsciously to the community, particularly for implicit collective search.
• Offer a service enabling personal publication: so that each one can generate his/her own information (post / comments).
• Offer a service enabling collaborative publication: so that the community can build its own content on topics people collectively find relevant.
Such a service effectively addresses the 90% of lurkers. By lowering the cost of participation, and providing the user the actual desire to contribute, we give to all members opportunities, incentives and reasons to start contributing.
This initiates a dynamic and growing flow of knowledge, which some such as Nonaka describes as a spiral of knowledge, along the lines described in the following graph:
3 – Concept validation
The proposed way of building sustainable communities is actually the classical way of engagement on the Internet:
• Exposure/Exposition: Each user is exposed to a flow of information
• Implicit Participation: Each user progressively selects information s/he wants by referencing content and regularly visiting websites and communities. By doing so and given the social features of the platform, not only does the user filter information for her/himself, but s/he does that for the rest of members (implicit participation)
• Explicit Participation: S/he progressively participates by commenting and then by producing her/his own posts or participating to a collaborative writing activity (explicit participation)
To demonstrate, we have successfully tested these principles with free/cheap tools, in our own work environment. The test followed the process hereafter described:
We built a list of feeds, made of publicly available resources on teamwork-related topics (social computing, enterprise 2.0, web 2.0, knowledge sharing, semantic computing), on iGoogle. iGoogle not only provides content aggregation, but also content personalization. Team members have been accessing it individually and reviewing content. They can freely organize the display of these feeds. They have been given the ability to add more feeds they find relevant to their work. The underlying assumption is that an employee knows better what information is relevant for delivering his/her mission (than a system administrator). They reference content they find of interest for the team – using their personal delicious account, they tag it with a special tag (keyword). A delicious subscription has been created to get notifications on this specific tag. Those notifications are displayed in a widget that is located inside the homepage of our second-generation wiki-based team space, so that all members are aware of each other’s referencing (implicit collaboration).
After a month, members initiated proper publications (email-based) to share their insights on their readings. This initiated work-related conversations that helped the team move to a different level. Some of those conversations lead to the development of specific collaborative work, supported by the wiki.
This wiki is a one-stop platform where members find documentation, work in progress, latest “hot” information for the team as well as notifications of individual activity on wiki pages. One can easily surface multimedia content via hyperlinks or directly display Microsoft Office and Adobe documents within a wiki page.
This helped people to access more easily to information and save time across all operations (hidden costs of: searching documents, opening relevant application to access document content). It also increased awareness and created self-training on the go, so that we saved a lot of time and money in training or education. Finally it sharpened the skills of members and as such increased consistency and relevance, which surface in our activity of designing services for business owners.
To make this happen, we have allowed members to spend 30 minutes per day reading feeds. We also complemented the individual evaluation framework by adding 3 bookmarks per day as part of their tasks.
Since then, the team also has tested certain enterprise compliant tools available on the market, evaluated new and richer ways of monitoring activity and interactions and has designed an integrated solution to service sustainable online communities. I will not disclose the software names, but one might not be surprised if I say that none of them are in catalogue of big/traditional software corporations at the time I write this article.