Thursday, January 28, 2010

Why telephone based market research was so much easier

Prior to the emergence of the Internet as a source for market research survey respondents, most research was being conducted over the phone. A research company would often (and many still do) have a large number of interviewers on the phone and in front of a CATI system.

Research via the telephone was considered the closest approximation to the general population due to near 100% of households having a telephone. Combined with random digit dialing (RDD), researchers were generally happy with a probabilistic randomly selected sample using this methodology.

The significance of the adoption and accessibility to telephones essentially gave everyone an equal opportunity to reach a statistically random sample. At any point, any researcher could pick up a telephone and with minimal cost, randomly dial a telephone number. It's not perfect for a few reasons which I won't get into here, but it was accepted as the most valid methodology. Even today, many agencies, governments, and companies believe and/or still use it as a primary method of reaching respondents. Certainly, depending on the objectives and type of research, in many cases it is.

Everyone in the game had the same access to the entire sample frame.

Enter the Internet and this has all changed and for one reason. To achieve a random sample across the internet, or the closest thing to it, each person on the Internet would have an equal opportunity to be recruited and interviewed. It is not the case that any individual can log on, and randomly select an IP address and initiate an interview.

I believe the magnitude of this challenge is far beyond the reach and capability of research firms and further, the economics would prevent it from doing so. To sample the Internet, you essentially must own the Internet, or at least have access to everyone on it.

There are a few companies that may come close to being able to do this, though market research is probably too small of a market (and too problematic) for them to turn to it. Well, at least not yet. A few companies who may have the reach to do this would include Google, Quantcast and Akamai.

Market researchers should be concerned. The science behind their craft is being challenged, telephone research is shrinking due to fatigue, abuse, cell phones and do-not-call lists, and the Internet is beyond their grasp. Email panels - or more accurately described as convenience sample - have been part of a transition, though the question is a transition towards what?

No comments: