Chapter 3: ICT4D design and evaluation 21
3.2 EVALUATION OF ICT4D 29
3.2.2 Information systems-centric evaluation 31
The work of Heeks (2002, 2008) related to ICT4D represents an information systems (IS) perspective. Heeks (2002) surveyed IS literature and identified three general categories regarding success: success, total failure and partial failure. It was easier to define success and failure of IS in industrialized countries: 20-25% total failure, 33-60% partial failure and only a small percentage in the success category. Heeks (2002) surmised that the success rates are much worse for developing countries8 (DCs). However, there is no real evidence due to a lack of literature on the topic in general and a lack of evaluation in particular. "The evidence base is not strong - and it urgently needs strengthening - but it all points in one direction: toward high rates of IS failure in developing countries" (Heeks, 2002). Heeks (2002) blamed the failures on the "design-reality gap", the lack of fit between the reality (actuality) and the intention (design) of ICT4D projects. He argued that at least the difficulties of ICT4D make the problems more easily recognised.
Heeks' (2002) model, then, was an assessment of the match or mismatch between "local actuality" and system design: the "design-actuality" gap. Of course, all stakeholders have their own subjective views of local actuality, so Heeks (2002) focussed on the designers and the
8 Developing 'countries' is a commonly used term that refers to what we call developing 'regions', according to Brewer et al. (2005). We also prefer the term 'regions' because many countries, e.g. South Africa, have both developed and developing regions.
users. He suggested a combination of 1) theoretical IS literature, 2) descriptive process on the IS system components and 3) analytical case studies. Heeks (2002) added consideration of objectives and values (culture and politics), staffing and skills, management systems and structures, and other resources (time and money). This is very similar to bridges.org's (2004a) monitoring and evaluation framework discussed below. The overall idea is to assess and rate the mismatch for each of the dimensions in the model (Heeks, 2002)
Pulling on the theory of Suchman (1987), Heeks (2002) then discussed how the mindset/context of the designer could get in the way and influence the system. An action "taken in the context of particular, concrete circumstances" (Suchman, 1987) and the designer have preconceived notions and assumptions about the user's situation. "This includes assumptions about the users' activities, skills, culture and objectives" (Heeks, 2002). In other words, design decisions can be laden with incorrect values and/or perceptions and there is much opportunity for things to go wrong. This is mostly due to having IS and ICT systems developed in much different environments from those in which they are eventually deployed. In industrialized regions, the worlds of the designer and the user are very similar. This is not true for ICT4D in a developing country (DC)9. "The domain of developing country information systems is particularly dominated by the transfer of industrialized country designs to DC actualities" (Heeks, 2002). The development and research for ICT4D systems is often done elsewhere, emanating from the realities of business, politics of aid, perceived cultural superiority of developed region products, and globalization. The most extreme form is when a system is built purely with developed region needs in mind, and then transplanted to an ICT4D context.
Even when systems are intended for ICT4D scenarios, a design-reality gap can occur because the bulk of the design is done by non-DC participants/stakeholders. Heeks (2002) also argued that this could also happen when non-DC people are not even involved because there is often a mindset in DCs that Western ideas are superior. Therefore, non-Western locals often adopt Western approaches even when they do not really fit the local actuality. Heeks (2002) also noted "Hard-Soft Gaps" where 'hard' represents rational design associated with
Western rationalism and 'soft' represents political reality often associated with third world cultural realities.
Heeks (2002) offered two suggestions to close the gaps: actuality and design improvisation. Actuality improvisation means "changing local actuality to make it closer to IS design" and design improvisation means "changing the (often 'imported') IS design to make it closer to DC user actuality". Heeks (2002) gave an example of a decision support system that had to be designed so particularly by a rational design process. "This deep design inscription will often create a large initial design-actuality gap" (Heeks, 2002). On the other hand, "actuality-supporting applications can be seen as more enabling" (Heeks, 2002). An example is a word processor because it is completely generic. "Shallow design inscriptions will be associated with a small initial design-actuality gap" (Heeks, 2002). Heeks (2002) meant that general purpose tools, design and components have more of a chance of lending themselves to both actuality and design improvisation, and that it is empowering local improvisation that provides an opportunity for success because that enables local take-up and ownership of ICT4D.
Lastly, when discussing the ability of locals to make the improvisations (namely, that they do not have the skills to do so), there is a need for intermediaries that he called a 'hybrid'. "Hybrids understand both context, organization, and work processes of their section and the role on information systems" (Heeks, 2002). Later, Heeks (2008) would define a hybrid as a person with intersecting skills of computer science, information systems and development studies when multi-disciplinary teams are not available or feasible. The next section addresses the third perspective of the 'hybrid', development studies.