Canberra.
The second speaker in this session at ANZCA 2010 is Marilyn Mitchell, whose interest is in visual communication; here, she focusses especially on the design of information 'dashboards' providing key performance indicators to managers in a visual format (one example Marilyn shows is for schools administrators, showing e.g. the percentage of buses running on time and a gauge of ethnic diversity in the student population).
Are such dashboards actually appropriate for providing an overview of such information? Are they cluttered with 'junk' information graphs? Some critics (e.g. Tufte) have suggested that the entire 'dashboard' metaphor is 'lame', others are actively promoting the idea, and at any rate these dashboards have become something of an industry standard, with many managers demanding them and an industry of dashboard providers now emerging.
Part of this is tied to the prevalence of key performance indicators (KPIs) to measure an organisation's baseline performance. At best, KPIs are used to steer an organisation towards its mission, aiming for a 'balanced scorecard'; the process requires senior management and a wider range of employees collaboratively to design the mission and vision of the organisation and determine how to measure its achievements. Organisational learning and growth are a special focus here, and in total, some 15-20 KPIs are usually identified and measured.
This balanced scorecard has been compared in the early literature to the cockpit of an aeroplane - and some designers took this metaphor very literally and began to design actual dashboard graphics to visualise an organisation's performance against its KPIs. Graphical representation, the argument goes, can help show all indicators at once, rather than isolated from one another, and can indicate when and where problems occur and how individual indicators influence one another; it can also help with the comparison of performance against competing organisations. It also allows for the replacement of unworkable KPIs with more insightful measures.
Visualisation can implement a number of different scale types: nominal (categorisation into types); ordinal (in some order); interval (even intervals); ratio (on a scale from zero). These can also be integrated with one another to form richer visualisations. What is necessary in a next step is to test how different dashboard visualisation styles affect the readability and comparability of individual performance indicators.