Modern organisations consider knowledge a vital factor of production and need to consider ways of managing this “knowledge”6. The distributed nature of agile development pose some particular challenges in this regard.
Traditionally organisations rely on generating large amounts of documentation to get a grip on the knowledge they are managing. These records “serve as useful artefacts for communication and traceability of design” (Nerur et al., 2005, p. 76). Since agile teams try to limit documentation as much as possible, this coping strategy cannot be applied. The issue is, however, more complicated and relates to the nature of this knowledge itself.
Tsoukas (1996) reconceptualises the entire firm as a “distributed knowledge system”7. He brings this in relation with sensemaking, making this conceptualisation particularly useful for organisations deploying agile development methodologies.
In considering organisations (firms) to be knowledge systems, Tsoukas also explicitly deviates from the neoclassical view of firms and the behaviourist conception of “human agents”. He rejects the notion that individuals are interchangeable or generically substi- tutable. Instead he acknowledges the view of humans as “agents, active co-producers of their surrounding reality” (p. 13). When rejecting the neoclassical “black-box” view of firms, he also references organisations as “interpretation systems” as an alternative option (Daft & Weick, 1984).
To understand the distributed nature of knowledge in organisations, a distinction be- tween tacit and explicit knowledge is needed8. Explicit knowledge is codified and stored in some document or artefact whereas tacit knowledge resides within individuals - often on a subconscious level9.
6
Knowledge Management as a field has developed from work done by many authors, including Nonaka, Snowden, De Jarnett, Clegg, Roos and Chase. The notion of “knowledge management” and its use in practice is, however, not universally accepted and is still a topic of debate in many industry and academic circles. Refer to Wilson (2002) for a more detailed discussion.
7
Rather than arguing the merits and demerits of this or that definition of KM the field itself, I will refer to a specific argument that is not dependent on an acceptance of KM as a field for its cogency or persuasiveness. Tsoukas fits the bill particularly well.
8
Tsoukas (1996, p. 14) refines this distinction to a much greater level of detail than is summarised here.
9
Tacit knowledge in this sense is a necessary enabler or precursor to all knowledge and should therefore not be contrasted with explicit knowledge as two ends of a spectrum. Both forms of knowledge are relevant and exist in an intertwined fashion.
6.4 — Knowledge Management Dimensions 112
Tsoukas (1996) points us to Nonaka & Takeuchi (1995, p. 61), who argue that “knowl- edge is created and expanded through social interaction between tacit knowledge and ex- plicit knowledge”. In their model, knowledge is created in five phases (Tsoukas, 1996, p. 14):
The process starts with the sharing of tacit knowledge by a group of individuals; tacit knowledge is subsequently converted into concepts which then have to be justified in terms of the organisation’s overarching mission and purpose; a justi- fied concept is then made tangible, usually through the building of an archetype; finally, new knowledge is disseminated to others within the organisation.
The location of tacit knowledge in several disparate individuals makes the management of this crucial organisational knowledge very difficult. Additionally the distributed nature of the knowledge relates not only to tacit knowledge residing in individuals, but also to the knowledge that exists between them. “Individual knowledge is possible precisely because of the social practices within which individuals engage - the two are mutually defined” (Tsoukas, 1996, p. 14).
Beyond the distributed nature of knowledge, the cognitive processes in organisations are often also distributed (Hutchins, 1996; Hutchins & Lintern, 1995). This means that the en- tire business process spans more than a single individual (or single mind). When combined with the emergent nature of organisations themselves, this has profound implications.
Building on the metaphor of an individual mind, Weick & Roberts (1993) have concep- tualised organisations as “collective minds”. This combines cognitive process, knowledge management and social aspects in organisations, noting that individuals “construct their actions (contribute) while envisaging a social system of joint actions (represent), and inter- relate that constructed action with the system that is envisaged (subordinate).” (p 363).
Tsoukas (1996, p. 17) provides a good summary of the difficulties organisations face when planning knowledge management in an emergent structure:
. . . all articulated knowledge is based on unarticulated background, a set of sub- sidiary particulars which are tacitly integrated by individuals. Those particulars reside in the social practices, our forms of life, into which we happen to partic- ipate. Before we are cognising subjects we are Daseins (beings-in-the-world).
An utterance is possible only by the speaker’s dwelling in a tacitly accepted background.
This background to which he refers (unarticulated) is created through a process of socialisation. This is recognised by agile development and is expressed in the pronounced focus it places on the social aspects within the process, team and organisation.
In short, companies operating in the agile development space face enormous challenges in addition to the normal difficulties companies are grappling with when trying to institute institutional knowledge management systems. The emergent configuration of development teams leads to an increased distribution, both in terms of product delivery (cognition) and the knowledge itself. This is compounded by the active reduction in documentation leading to less explicit, codified knowledge artefacts in existence. Traditionally knowledge manage- ment systems were only concerned with managing these artefacts and their absence forces organisations to critically rethink these assumptions. Lastly the recognised uncertainty in systems dealing with agile development also raises questions about traditional knowledge management approaches and their suitability given these conditions.