Evolutionary stakeholder discovery: requisite system sampling for co-creation

Several recent studies have published well-developed practices of co-creation, design facilitation, and stakeholder convening for advanced design collaboration. There may be many systemic design methodologies that prove effective in their consultative or engagement settings. Yet in any design process requiring consensus in participant decisionmaking, non-parametric design contexts I refer to as Design 3.0 and 4.0 (Van Patter and Jones, 2013), we face a practical and systemic problem with stakeholder representation. Unlike product or service design (Design 2.0) we cannot merely sample from a user base to inform design decisions targeting future product releases.
\n
\nIn Design 3.0 and 4.0, the “users” are the system. Real stakeholders are not merely representatives of a social system in which they hold membership, they are committed co-producers of the existence of the system of concern. As in a wicked problem, each selection of stakeholder matters, and they co-create a framing and context that remains path-dependent, that cannot be undone. Vision, context and direction setting are extremely sensitive to initial conditions, and – especially when performed well – may create a lock-in effect with confirmation of beliefs among actors that their choices represent desirable preferences for future system participants.
\n
\nIn systemic design we face the wicked problem dynamic of a changing problem frame with each selection of participants. We can see shifts between each stage of a progressive design process, sustaining an essentially artificial co-creation engagement. These methodologies initiate design co-creation from visioning and problem framing, through system concept formulation, and toward consensus on collective action. All of these activities require stakeholder insight and validation, and much less design guidance and content (as necessary in D2.0 product/service contexts).
\n
\nAny design process may be irrelevant if stakeholder selection fails to represent the requisite exogenous variety in their social system AND fails to enroll authentic commitment from those selected stakeholders. Because design disciplines are predicated on a tradition of creative problem solving, these functions are often underdeveloped. We do not select and enroll sufficiently well enough to guarantee an effective result.
\n
\nWestern culture now exists in what we might call a late-modernist knowledge society, and we have centred users and stakeholders as the source of knowledge and validation. Human-centering is itself presented as evidence of ethical practice, or at least, a necessary sensitivity to multivocalism in design process. However, the situated placement of (usually) self-selecting participants as representative “voices of the system” can slip into an efficient, unreflective process that escapes responsibility of future consequences of design decisions. We clearly would not decide a consensus for real social system participants. Yet how are we disclosing ourselves as lifeworld-sensitive designers, when we, perhaps even worse, decide who will be the system participants?
\n
\nDesign problematics in the many domains we now touch involve social complexity and the complex multiplicity of stakeholders. If we recognize that stakeholder co-creation is a context for design facilitation, we bring forth skills for different roles than product or service designers. We (systemic designers) are neither authentically domain experts or visionaries in the highly complex fields in which we serve, such as urban planning, healthcare, education, ecological community design, even AI and technology. The search for an authentic commitment, our stake in the game, must be negotiated in our experience and contributions to domains in which design is demonstrated through care, not just performance.
\n
\nThe proposal presents an iterative stakeholder sampling process developed for Dialogic Design (Christakis & Bausch, 2006) and other foresight methodologies, where the undersampling of variety leads to insufficient knowledge and gaps within critical areas of social representation. Based on two design action research cases performed with a large US research lab and Canadian foresight studies, we advance a sampling model that integrates four dimensions:
\n
\n• Worldview perspectives, based on Latour’s (2013) Modes of Existence ontological typology as a social theory of orthogonal perspectives,
\n• Diversity and demographic characteristics (including temporal preference),
\n• Foresight temporality and trend categories, and
\n• The SDD reference stakeholder model (Christakis and Bausch, 2006).
\n
\nA model for Requisite Stakeholder Variety enables robust sampling for ontological representation, variety, biases and diversity of knowledge, and exogenous representation commitment (e.g. skin in the game, Taleb & Sandis, 2015). A canonical stakeholder selection model maps selected foresight categories (e.g., STEEP) to worldview ontological domains, and further diversifies by variety attributes including age, culture, gender, and proposed horizon preference. This mapping identifies significant relationships of knowledge and trends across domains and disciplines. At minimum the stakeholder sampling model provides a checklist that exposes possible risks and blind spots in the available composition of stakeholders or experts. The model further provides a schema for identifying values conflicts between worldviews and other attributes associated with known stakeholder interests (such as strategic preferences that planners wish to include).
\n
\nThe requisite stakeholder variety model for stakeholder discovery was designed to address the necessary variety in high-stakes foresight for long-term R&D strategies and as a reference model for anticipatory policy research. We have proposed an approach called evolutionary sampling, that iteratively samples stakeholders from across sets of covarying dimensions identified within the social system being designed. This method also effectively enables planners and sponsors to reveal biases and risks and to trade-off potential leaders, dominant voices, and under-represented minority views within the social system of concern.

Fetched live from OpenAlex and de-inverted. Abstracts are not stored in this database: the inverted indexes are 8.6 GB of the frame’s 9.3 GB of text, and the host has 13 GB free.