FreshNetworks Blog

Online research is big business, back in 2005 a report from Cambiar predicted that it would account for about 20% of the research-related industry by 2008, and there is little to suggest this isn’t true. But online research can lead to many problems – with no way of checking the true identity of respondents it is up to panel providers and research firms to ensure they follow world-class standards and codes of conduct, and that they are constantly vigilant for possible cheats.

One cause for potential concern is that a relative small number of respondents fill in most online surveys globally. Research from ComScore in 2006 suggested that 20% of respondents account for 80% of online surveys – the fears are of expert respondents, fraudulent respondents and even automated software completing surveys.

The best providers of panels and online research will use techniques to combat the last of these – including test questions of the type “for quality control purposes please select ‘very good’ for this question”. These types of techniques can help combat software designed to automatically complete surveys as they will be unable to read and understand the semantics of the questions. They just randomly complete answers.  But there has been little than can be done to automatically spot and remove fraudulent respondents. Perhaps, until now.

A report in AdAge this week reveals a new system from Peanut Labs that is in beta testing. Optimus creates a digital fingerprint of each computer in a survey. It can then identify respondents who fake responses or professionals who complete as many surveys as possible for the incentives or just to falsify survey results. It works a bit like the systems used to identify and prevent fraudulent use of online banking. Even though it’s only in beta testing, the system is currently used by Microsoft, Warner Bros., and 20 of the 50 biggest research firms. So clearly agencies and clients alike think that it works.

But whether or not the system works it will still be unable to completely guarantee that respondents are who they say they are or that they respond genuinely. In some instances it is better to build a relationship with respondents, to build trust both ways – that they know who you are and you know who they are. This is particularly the case when using online for qualitative research where more emphasis is placed on quality and depth of responses. Online communities are helping to change the face of online research in this way. Respondents are open and honest about who they are – they get to know each other and the brand running the community, and so are much more likely to be honest and open when they respond to your survey.

At FreshNetworks we wrote a white paper on how online research communities are changing online research. If you want a copy then send me an email here.