The rise of “always-on” ubiquitous computing and increasing inter-connectedness through online communications are creating the necessary conditions for complexity and emergent behaviour in online social interaction. Systems thinking is becoming increasingly relevant to understanding how online social networks develop, yet much of our current online software and systems are still rooted in organisational process-driven ideas that cannot adequately explain phenomena we see around us.

Instead of imposing centralised one-size-fits-all software and then using a combination of coercion and marketing to encourage people to use it, we should be building smaller, more modular and adaptable software services around the very people who will use them, and they should be simple to use, ideally transparent to the user. If we are to exploit the potential of online communication to develop social capital in networks and organisations, then these online applications should aim to augment our social interaction and support our connections with others rather than replace them. In addition to building large, structured online environments where people come together to work, share knowledge and communicate, we need to equip users with the tools required to interact on their own terms, and allow the emergent properties of highly-connected social networks to come into play.

Current knowledge management, learning, CRM, communication and collaboration systems attempt to segment our online interaction into discrete channels that demand too much attention and adaptation from their users. Even some of the best online software and tools do not provide interoperability, nor are they sufficiently adaptable for their users to mould them to fit their own specific purposes. This is a problem that also applies to much of what is loosely termed social software now, such as Instant Messaging, social networking and weblog creation tools. Many of these are proprietary, inflexible, idiosyncratic and far too demanding of user attention.

A crucial differentiator for social software is the notion of adaptability to human behaviour, which means we will have to progress beyond existing information search, tagging and classification systems to enable computers to understand more about us, the way we think and what is relevant to us in any given context. The ideas behind the Semantic Web hold out much hope for making the glut of available information work for us rather than the other way around, and for letting computers process more of it without our intervention. Metadata, ontologies and interoperability standards all provide different perspectives on the online world around us. When these can be mixed and matched between people and devices and layered on top of each other, then we can start to share ideas more effectively. Other emerging protocols, such as Friend of a Friend (FOAF), that describe aspects of social networks and the connections between people could potentially facilitate trust-based interaction and network navigation.

However, social software is not just about the tools themselves, it is also about how they are designed and built, what they are used for, and also the process of their adoption, rollout and uptake.   The way developers work and also the way organisations run projects to integrate online communication systems into their work and culture are both still far too technology-centric.

We believe that any methodology for creating online social applications should begin by conducting stakeholder and social network mapping to identify the key people involved and how their inter-connections are organised. Ethnographic methods can help observe existing behaviour within these networks, and social network analysis can help identify roles, patterns of interaction and structural holes that might need to be bridged or links   that need to be strengthened. Developing from this base, we need analytical frameworks for looking at the nature of these relationships, the knowledge flows between them, the organisational processes people are involved in, and their information and interaction needs. Only when we have this, can we start to produce representative user personae that can guide development.

Although user research, surveys and focus groups have been widely used to try to provide some user input into the development process, social software development requires a much more intensive engagement with key stakeholders early in the process. This can help mobilise existing knowledge, experience, ideas and energy that can be brought to bear on the project in a positive way. For example, control of language means control of the means of self-representation, as religious and political leaders have understood for hundreds of years. Therefore, we advocate intensive stakeholder engagement in the process of information architecture development, content analysis and classification. If stakeholders can contribute to the language and terminology used to organise and mobilise knowledge, they are more likely to engage with the resulting system or product.

Based on this knowledge of the environment in which the project seeks to act, it is easier to come up with initial ideas for specific tasks that online software can support, whether it is about culture change and organisational development, communication and collaboration, learning and personal development, publishing or knowledge sharing. Then, the task is to use our understanding of network dynamics and social systems to deploy techniques, online tools and environments that can support these applications on a collective basis, whilst still addressing individual needs and characteristics. The resulting software must be designed primarily to enable users to connect with each other and help them organise and self-manage their work. This means relationship management tools and personal content / knowledge management tools are increasingly important, and these are two classes of software that have suffered more than most from bloatware tendencies and usability deficiencies.

The user experience of online social applications, derived from the research and analysis, should be as adaptable, simple and as transparent as possible. In design terms, the ideas behind the undesign movement provide a useful way forward, although this is not its main purpose. The lesson of successful sites such as Amazon, Ebay and Hotmail is to keep it simple and adaptable. However, the emotive and cultural aspects of the user experience we are trying to create must not be overlooked. If we are asking people to use our tools to augment their social networking, the usage of these tools must be easy, unobtrusive and appropriate to the context in which they operate. This applies equally to visual and structural design.

Information architecture is an area that promises to be transformed by our evolving thinking about social software. Clearly, we need to look at the users, content and context of any project in the light of our social network research. “Sites don’t exist in a vacuum,”as IA guru Louis Rosenfeld says , “but in a broader information ecology made up of content, tools, people, roles, interactions, policies, barriers, etc.” However, the potential for innovation in this area goes way beyond this.

There has always been debate about the usefulness of metadata and shared standards in the real world, but social software has the potential for something much more exciting: multiple sources of collaborative, emergent metadata that can go beyond syndication towards synchronisation. Ideally, users should be able to experience, share and manage personal knowledge according to their own individual perspectives. This would mean a move away from the primacy of systems to focus on making it easier for people to find, organise, and share knowledge and information. Combined with techniques such as latent semantic indexing and ontology building, this could make it much easier for us to personalise search engines according to our own point of view rather than just basic preferences.

The advent of XML , RDF and distributed Web services technology in general makes such ideas a real possibility for even relatively small projects. The main benefit of these approaches to technical development is that small, simple, modular applications and services can be developed rapidly, and with a much lower risk and cost. Development techniques that aim for modular code with common properties and methods are also a good idea – Fusebox is an example from the world of Web application development, but there are others. It is gradually becoming more appropriate to think in terms of supplying a pool of shared online tools within common methods and parameters, rather than building a single huge monolithic software product. If that is the case, what is stopping us from being able to build tools around the very people who use them, involving them in the development and evolution of their own online tools and applications?

In project management terms, it is also desirable to build around the existing networks, people and resources identified during the initial research, both for reasons of cost and also to ensure involvement and buy-in from the start of a project. Building online social networks should be about building organisational capacity and social capital both for the host organisation and the network they serve. Complete outsourcing of such a core function is no longer the automatic answer, with too many outsourced projects failing or, where they do succeed, creating little in the way of knowledge transfer and residual capacity for the host organisation. Purchasers may not fully understand the process at the beginning of a project, but if they don’t feel confident talking about the issues involved by the end of it, then the developers have not done their job properly. If social software is to live up to its name, it must be borne out of a partnership between stakeholders, purchasers, developers and users, and this should ideally involve de-mystifying and making more accessible the design and development process.