Dave Snowden

Describe don’t evaluate

RSS Feed

One of the things I have been emphasising for the last two days when I have been talking about SenseMaker® is the criticality of description over evaluation.   The minute we evaluate, assess, judge, interpret or whatever we start to reduce what we scan.  The more we can hold open a description the more we scan, the more possibility of seeing novel solutions or interesting features.   In some of our work where we have compared the way experts interpret narrative as against the original signification by the research subjects we have generally found that the former is instrumentalist, the latter descriptive.

One of the key issues in any complex adaptive system is to prevent premature convergence, to avoid hypotheses, to remain open to weak signals and options.  It was one of the main influences on my decision to use balanced triads in signification.  By creating three positive or three negative descriptors on a triad, you prevent the respondent from having any inkling as to the desired result and the positioning on the triad means that you get diverse and interesting descriptions.   It means that you get a richer range of responses when you ask the basic SenseMaker® action question: What can we do to create more stories like these and fewer stories like those.

Interpretation is also contextual, so any evaluative scale is limited in its utility over time.  On the other hand high abstraction, descriptive signification means that original data maintains its value long after the original purpose is forgotten.   In conventional surveys the output is the report and the raw data is abandoned.  In SenseMaker® the raw data retains its utility over extended periods.

Top