Need for Speed- but at what cost?
In an increasingly globalized and competitive business environment, time is of the essence. Companies across different industries are increasingly seeking to gain market and consumer insight within shorter periods of time. Whilst technological developments such as Big Data or Online Panels facilitate this, it is worth asking whether the drive for speedy research results has been at the cost of quality respondents.
Online research panels have the incredible befit of providing access to thousands of respondents in just a matter of hours. However, this large pool of respondents will also include research bots, duplicate respondents, straight liners, speeders, etc. who participate with the aim of gaining incentives.
Over time, panel companies are getting better at screening out poor quality respondents. For instance, most providers typically employ algorithms that exclude respondents who answer the questionnaire in less than half the average length or similarly respondents that answer A to every question, etc.
Though automated checks reduce the number of poor quality respondents, experienced panel providers understand that this needs to be supplemented with human verification – especially when trying to identify duplicates or unengaged respondents. For instance, in a pool of 500 respondents, the chances of having two Emirati respondents aged 43 and living in Fujairah are fair enough. However, the chances of these questionnaires having been completed within just minutes from each other, with the same words misspelled, etc. are highly unlikely. A replacement strategy should also be in place in order to gain quality insight from suitable and engaged respondents- even if this might take more time.
In the ongoing debate about automation in the research industry, it is important not to underestimate the role that humans still play. With artificial intelligence unable to identify logical anomalies or other discrepancies, the cleaning process remains a vital part of the research project albeit one that is often overlooked in the constant quest for faster results. As such, it is important for companies to question what proportion of bad respondents they are actually willing to accept when prioritizing speedy results and what impact this will have on the final results.