APPLIANCE OF SCIENCE

Gazing into the future and making predictions about the way new technology will revolutionise our lives will always be an imprecise science.

Mindful of the ‘Tomorrow’s World” presenters of the Sixties who promised we would all be living in eco-homes and driving solar powered cars by the Nineties, it would take a brave soul to stick their neck on the block and say how new technology will transform the market research industry over the next few years. However, the commercial rewards for technologies that can speed up information gathering at lower costs are driving developments forward at an increasing pace.

Predictably, the Internet has received the lion’s share of attention. The prospect of accessing millions of cyber gossips, debating everything from the merits of different brands of cola to the latest Vauxhall Almeira commercial, is an attractive one for the market research industry whose existence depends on gauging attitudes and opinions.

Despite the tantalising opportunity to tap into the thoughts and minds of the global village, the Internet as a medium for gathering facts and figures has many sceptics. They point out that its biggest failings relate to the demographic profile of Internet users and the problem of self-selecting respondents.

According to Peter Comley from Simon Godfrey Associates who has been investigating the potential of the Internet as a research tool, the typical Internet user is male, aged between 24 and 35, and earning an above-average income. “The user profile is very narrowly defined and, for the moment at least, it would not be feasible to use the Internet for representative surveys,” he says.

While most of the market research industry would agree, Research International’s David Walker reasons that all research methodologies have their restrictions and the Internet is no different.

“Providing you accept its limitations and recognise that it doesn’t have perfect features, there are several areas where the Internet has enormous potential.” He points out that there are currently over 15,000 bulletin boards which lend themselves to live and offline qualitative surveys, particularly for products aimed at this market segment.

“There are discussion groups on pretty much anything. If, for example, a drinks company wanted to test opinion on a new advertisement, they could post it to one of the alcohol discussion groups and invite responses via the e-mail. A more sophisticated technique would be to conduct “live” interviews or online discussions using Internet Relay Chat,” explains Walker.

Marketing consultancy Michael Herbert Associates recently conducted an online attitude survey through the Internet on behalf of the Electronic Telegraph (the Internet version of The Daily Telegraph) and were encouraged by their results. But, as research executive Sarah Davies points out, “the Internet is growing so rapidly it is difficult to predict how it will develop. However, we see its application in business-to-business research as being very important in the future”.

Comley goes along with this view and believes e-mail will soon be as common in the office as the fax machine is today. When that day arrives, business-to-business market research via the Internet will really take off. “As the system infrastructure improves, it will become much more widely used, especially with the possibility of video links in a few years time,” he says.

As a quantitative research medium, it is probably still too early to make meaningful comparisons between the relative merits of Internet/e-mail and postal surveys. Putting aside the limitations already described, Comley claims that pilot studies have nevertheless produced interesting findings.

“SGA has undertaken a couple of comparative surveys which have produced some spectacular results. We found that the response rates to postal and e-mail surveys were 15 per cent and 14 per cent respectively, but the major difference was in speed and cost.

“By post, we found people took up to ten days to respond, whereas by e-mail many people replied almost immediately, and the majority did so within three to four days. We also calculated that after you had taken into account the associated costs, the postal survey was seven times more expensive than the e-mail survey,” he says.

Despite its undoubted potential, industry commentators do not see the Internet as being a viable research tool for another five to ten years. It is hardly surprising that other mediums of data collection are being viewed more enthusiastically by the research companies.

Taylor Nelson AGB has worked alongside GEC Marconi in the development of speech recognition technology, which they see as having enormous potential for computer assisted telephone research. Bill Blyth, research director at Taylor Nelson, believes the technology will soon be at the stage where the computer will take charge of the telephone interview, allowing companies to make cost savings on the most expensive item – the interviewer – and improving the efficiency of the service.

“Our early trials with GEC convinced us that speech recognition had a whole range of potential applications in telephone research,” says Blyth.

More recently, development has been assisted with funds from the European Commission as part of the Reward project, which has brought together a European consortium of expert teleservice builders in a variety of market sectors as part of the Commission’s Telematics Programme. The Taylor Nelson team claim the speech recognition software is now sufficiently advanced for it to distinguish between male and female respondents by recognising the difference in speech patterns.

While it seems the technology has largely been cracked, the biggest challenge is to simulate the innate skills of the human interviewer and interpret more complex dialogue and responses.

Research manager Heather Piper explains the problem: “When we developed the Speech Recognition software, the starting point of the design was no different to our usual approach in questionnaire preparation. However, when a human interviewer asks a question, they will often pick up clues about the lifestyle of the interviewee simply from background noise and that particular respondent’s choice of words and expressions.

“In developing the programme, we have had to incorporate larger vocabularies than the speech platform had handled before, and develop strategies for coping when the answer given didn’t match well enough with anything we had anticipated,” she adds.

Earlier this year, Taylor Nelson completed a pilot study questioning respondents about holiday plans and claimed that of the 170 calls analysed, over 60 per cent completed successfully. The company believes this figure would have been higher than 80 per cent with a few modifications to the software.

Blyth believes that speech recognition will be available for widespread use within 18 months, though a bigger application for the technology may be in telemarketing.

“Speech recognition means that companies running direct response campaigns will have greater flexibility in monitoring customer reaction,” he says.

Interactive TV also offers enormous potential for real-time data capture and many companies have been working alongside the Cable TV stations in exploring this area.

NOP, in collaboration with a number of partner organisations including Tesco, the BBC, NatWest, the Post Office, IPC magazines and ITN, recently completed an interactive TV trial in the Cambridge area.

The initial pilot covered 100 families who were connected to the Cambridge Cablevision Network and provided with cable TV and video services on demand.

NOP’s main interest in the trial was to investigate the feasibility of conducting surveys through the cable system using software which had been developed to enable respondents to answer on-screen questionnaires relating to viewer satisfaction of the service.

Joe Michael, project manager at NOP is encouraged by the results and claims that over half the people who took part in the trial responded to the questionnaire.

“Obviously one has to take into account the novelty factor, but a 50 per cent response, without prompting, is very pleasing,” he says.

Interactive TV offers several benefits over existing research methods and other new technologies. As Michael explains: “You are dealing with real time and there is almost unlimited scope to incorporate prompts and interactive pre-recorded interviews that simulate face-to-face interviews. Respondents complete the questionnaire in their own time and do not have to return it by post.”

Quite apart from the potential cost savings, NOP believes that interactive TV can improve the quality of responses, is less intrusive and has a number of wider applications including political polling, advertising, direct response and sales promotion.

Nobody can say with certainty how the new technologies will have an impact on the market research industry, though it is unlikely that they will consign traditional tools to the dustbin overnight. Given that the application of the technologies has largely been demonstrated, how soon and how widely they are adopted will primarily be cost driven.

As Comley describes it: “The major benefits will come from the savings associated with not employing interviewers for face-to-face and telephone interviews.”