Engagement through co-creation

This is the seventh in a series of articles on the topic of Respondent Engagement and Respondent Survey Satisfaction.

When searching for new types of gamified survey tasks that create a heightened sense of focus and engagement, the inspiration for interactive methodologies can sometimes be found in other, non-research online systems.

In the 1990s, Dell Computer launched a novel type of computer ordering system in which people selected from a menu of different components and, in essence created a custom solution, perfect for their individual situation.  This process was called “configuration” and spawned many similar ordering systems that have become quite sophisticated over the years.  At this point, almost every car manufacturer has a “build your own” automobile configurator on their Websites that can be used to create custom-build specifications and the associated prices.

Illustration 1: Dell Computer Configurator System (Circa 2004)

During the process of making feature decisions, researchers noticed that a tremendous amount of ancillary data were being collected simultaneously:

  • How long did people play with a specific feature set?
  • What features were most traded when price exceeded budget?
  • What features were commonly left in the “basic mode” and never upgraded?
  • Did certain configurations occur more frequently with different types of customers?
  • What features were more price sensitive than others?

This real-time exercise is an example of co-creation, wherein participants are helping understand the purchase-decision process, simply by making choices that are right for them.  This choice oriented method can be analyzed using standard statistical processes by which commonly selected features can be clustered to indicate people with similar needs, wants and desires.  The clusters can be profiled to provide rich information about target groups of customers and what product characteristics are of most interest.

The notion of co-creation is not new.  Its discussion in the literature dates back to the 1980s, but was formalized by C. K. Prahalad and Venkat Ramaswamy in a Harvard Business Review article entitled: “Co-Opting Customer Competence” [January 2000].  They described four specific forms of co-creation which involve varying degrees of control over the choices and selection activities.  In the configurator example we’ve been discussing, the company provides the feature choices and the respondents control the selections.  The authors would designate this type of co-creation as “tinkering,” but a very valuable form of tinkering it is.

A practical output from a configurator-style co-creation process is an “advisor” function.  In Web-based systems, where a live human is rarely available for questions or suggestions, an artificial advisor can be useful in giving visitors suggestions that might meet their needs.  The advisor is actually the product of the clustering information that was produced from various common configurations based on shared need states.  Once the customer answers a few questions about their general situation, the advisor can (using a discriminate function typing tool) determine which type of product configuration they would most likely find satisfactory.  See an early version of Dell’s pioneering “Help Me Choose” advisor in the accompanying illustration.

Illustration 2: Dell Advisor Function (circa 2004)

A demonstration project that we created several years ago in conjunction with Symrise Corporation, allowed people to configure their own ideal cocktail.  By understanding the basic inputs (type of alcohol, flavorings, carbonation, garnishes, etc.) we were able to identify eight general groups of cocktails, one of which would appeal to almost everyone surveyed.  Using a typing tool we were able to distill a huge number of options down to five simple questions.  Once a new visitor to a liquor manufacturers’ website answered those questions, the advisor could propose a cocktail that would most likely appeal to them.  The visitor would then be free to tinker with the options, but most usually we found that the suggested solution was accepted as is.  The site could then generate a shopping list for the ingredients and instructions for creating the suggested cocktail.

Illustration 3: Symrise Cocktail Configurator Interface


When we have tested survey satisfaction with various forms of choice-based product development methodologies, we found that the activity which allows people to make free-form choices (configurator) is more popular than methods that require multiple ratings of pre-selected feature sets (various forms of conjoint).  While the two methods produce similar output, the engagement and satisfaction levels are higher when people are in co-creation mode versus response-only mode.

As always, let us know if you have experience with co-creation and active product configuration style projects.

Bill MacElroy is Chairman of Socratic Technologies, Inc. www.sotech.com

Leave a Reply

Your email address will not be published. Required fields are marked *