This is a talk as part of the Design Informatics seminars at Edinburgh College of Art by Gian Marco Campagnolo on research he conducted with Neil Pollock and Robin Williams.
These notes are being written live so please excuse the inevitable shonky moment.
Frames this as a talk against dichotamania and the space between thinking fast and slow and thinking consciously rather than by instinct as per Gladwell’s Blink.
Looking at the dominance of 2×2 matrices and the academic critique of such matrices and yet their clear popularity and success in management practice. How do simple visualisations differ from more complex visualisations and their relative effectiveness.
Discourses on visualisation critique their simplification such as in HE rankings and visualisation techniques as fads and fashion. Another important limitation in visual numbers is that the production tends to be isolated from the context of their consumption eg, as part of oral presentations.
How differen or similar are they from other types of evaluation?
Subitizing is based on psychology experiment form 1880s where dots become perceived as a picture and the dots are ‘consumed’ sequentially rather than in parallel and hence become perceived as more confusing. These studies point to a threshold (point of discontinuity) of different perceptions of visualisation of numbers- although where that threshold is is disputed. Kauffman (1949) defines subitizing as the human skill to immediately make rapid accurate and confident judgements about small number. But this is not estimating which lacks the sense of accuracy an confidence of subitixzing. So this suggests that rapid judgements are not simple or careless and do allow rapid sense-making of a situation.
Of interest is identifying that point of where people reach that point of having “got it” – to understand a complex situation from the visualisations.
Looking at industry analysts such as Gartner who produce over 150 different quadrants for different technology markets – but these are very expensive to purchase. Such quadrants are some of the most influential pieces of business research with the power to make or break a new technology. These quadrants are perceived and used differently by different users.
Fieldwork: eg, to attend Gartner events and observe what users are doing. Event design is helpful here as networking and discussions are encouraged.
In discussions uses picture of the quadrants to prompt and focus discussions – a graph-elication research process – for 13 interviews in two days. These included users (software providers; manufacturers; finance); analysts; press agents and procurement/ sales professionals.
How do people use the quadrants and at what point do they subitize?
Quad used to make and defend quick evaluations: to communicate recommendations to senior managers. A key decision-maker is the CIO and the quadrant may be used to justify the CIO”s decision. [but are decisions informed by the quadrant or used to justify a decision using less complex analysis]. Decision-making is a long and difficult process to develop sufficient understanding od more complex analysis. The danger with quadrants is that they are perceived as easy to read and so decisions may be made before the necessary deeper understanding is reached?
Particularly useful in identifying established lead players in a market but also who are the smaller, successful niche players. But some interviewees claimed not to use the quadrants in their research although evidence seemed to be that the quadrants were used to frame analysis and sense-making.
Quadrants also render the research process and detail invisible. Gartner do not share their research process partly as the quadrants are to initiate a conversation rather than defend the research process.
Found evidence of the application of the subitizing range in how users make sense of the quadrants. Procurement professionals use the quadrants to identified appropriate vendors to invite to tenders – set expectations of who to include Request for Information and then the short list for Requests For Proposal.
Other analysts use other diagrammatic forms such as the Forrester wave. which includes tools to assist vendors to understand how they can improve their position. These visualisations tend to include only the top twenty or so vendors in a specific market. So such tools are used by vendors for competitive positioning and to be included in the ‘magic’ quadrant [so these visualisations effect strategy-making decisions].
Some evidence of regional biases in the consumption of the quadrant – but the model of the quadrant is used as a framing and reference model.
Interesting temporal aspect to these visualisations as each quadrant is only produced annually and so the evidence base may have a lag of three years or so?