| The mood was sober and attendance flatter than previous years at CASRO’s tech conference in New York, perhaps reflecting the spirit of the times in the world of IT. Was it prudence or paranoia that two of the dozen sessions devoted to the domesday topics of averting cyber crime and disaster recovery? Yet the message was also that technology still has a vital role to play in helping research adjust to the times.
CRM needs MR as much as MR needs CRM
Though predicted to become a $46 billion industry by 2003, over 55% of CRM systems have failed, Antony Cospito (CRM Metrix) claimed. “MR needs to understand the goals of CRM and fill the knowledge gap,” he said. But, for research to help, it needed to move from being an activity frozen in time to being constantly-on, providing a ‘ticker’ of customer satisfaction tied to retail activity on decision-maker’s desktops.
Ken Elliott (SPSS) saw opportunities for the MR industry in bringing together data mining and market research, as data mining has a valuable and sometimes overlooked role to play in all three facets of CRM: operative, collaborative and analytical. “Analytical CRM is devoid of leadership at present,” he stated.
|
“CRM vendors are busy concentrating on dashboards and OLAP [online analytical processing from ‘data cubes’].” Criticising this commodity approach to data doesn’t help customers.
Ray Pettit (NOP World) was also scathing of CRM specialists and the growing number of ‘ROI experts’ for failing to deliver insight by providing inadequate, over-simple analytics that were “almost a joke”. CRM providers were selling a closed loop but, with no effective benchmarks or evaluation strategy, rarely delivering one. “There is often no time and no expertise within the organization to do the analysis,” he said, while at the top, CEO metrics too often focused on a narrow definition of ROI that missed out psychometrics or any real opportunity for consumer insight.
Trading off different online conjoint methods
A stream of papers examining Online Choice Methods (aka conjoint) created a much-needed debate on how to define best practice in online conjoint exercises.
Len Bayer of Harris Interactive observed that online choice software has transformed what was previously a face-to-face method, into something that is self-administered. The loss of human
|
interaction is compensated by plentiful, accurate paradata (time or keystroke data) about the choices being made.
At Harris Interactive, full profile conjoint - in which a smaller number of attributes are presented in greater detail, giving a higher degree of realism, have proved to be one of the most robust methods. But for Bayer, discrete choice conjoint, (which allows ‘none of the above’ choices), “is eclipsing other methods in the view of clients” by offering better simulation of the market situation, and less burdensome designs.
Ray Poynter observed that respondents from online panels were generally willing to work a bit harder on conjoint exercises than those recruited from web intercepts, and randomized designs were particularly robust, commenting: “Even if you don’t have the right design, they tend to work.”
It puts one in mind of monkeys with typewriters, who, given enough time, bananas and typewriter ribbons, will tap out the complete works of Shakespeare.
Somewhere out there is the optimised product - it’s a matter of how long you want to wait.
CASRO: www.casro.com
Tim Macer's website is at www.meaning.uk.com
|