Automating market research can get things done faster and cheaper, but care must be taken to safeguard quality when designing the necessary software
Image above: Using webcams and machine-learning technologies, facial coding companies such asRealeyes measure emotional response to video content online to help brands target optimised content to the right audiences
Who in business doesn’t welcome innovations that help staff work faster and cheaper? Who wouldn’t opt for speed and efficiency wherever possible?
Market research used to take a long time and be costly. But the traditional method of commissioning research on a particular topic and waiting for the results is being replaced by real-time insight, delivered directly through a software platform.
Automation obviously allows market researchers to do more, faster and with less. However, it might get faster results and be more cost effective, but it does have its downsides.
Market research and automation go a long way back, says market research expert and managing director of The Future Place consultancy Ray Poynter, who has conducted a survey, sponsored by ZappiStore, on insight automation.
Use of automation in market research extends from punch cards and print rooms for paper questionnaires, to optical scanning, computer-aided telephone interviewing, the rise of internet data collection and the use of smartphones. Mr Poytner says: “Market research has a long history of beating itself up about not being innovative – something that I believe is more grounded in attitude than fact.”
Fewer roles for interviewers
Automation is impacting the whole industry, he says. In his survey, respondents cited salary payment, e-payment, e-booking and leave approval as ways in which their lives have been impacted by automation.
Within market research, automation impacts things such as digital-based data collection tools, which results in fewer roles for interviewers, and new approaches including automated facial coding, market research online management, social media listening and the use of apps for smartphone-based research.
In addition, sampling and quota management is highly automated, as well as some survey creation and data processing. Mr Poynter says: “Projects are increasingly using automation to enable participants to move from one aspect of a study to another without manual intervention.”
Repeat tasks are also being automated. For example, the fielding of Millward Brown’s Link test is now almost completely automated, saving enormous amounts of time and consequently money.
Giving analysts room to think
But it’s not just about saving time and money. In some cases, Mr Poynter says, automation has managed to increase engagement. “In real-time tracking, it has enabled the data-gathering to become seamless, and painless, for the participants,” he says. Additionally, automating market research can increase productivity in other roles as analysts are able to have more thinking time and less processing time.
Though in Mr Poynter’s research, respondents cited standardising as a benefit, some thought it a negative, as it can exclude potentially better solutions or make these solutions more expensive. Also, some automation is “black box”, which means it’s harder to access and make alterations.
Staff will be able to handle much larger quantities of data and better use processing power
“If reporting is automated, but the data entry on the frontline is manual, the system can look more useful than it really is, leading to faulty decisions being made,” he says. Moreover, automation may tend to report the simple truths, not the key ones.
A cost-reduction mentality afforded by mechanising these processes might end up spilling over into tasks that can’t yet be automated and this could lead to some good solutions being removed from the decision-making process.
Automation further opens up the potential for a dominant supplier to come and distort the market as Google has done, for example, with online advertising. Ultimately, like automation in any sphere, this can lead to less focus and result in job cuts.
However, in the future we’ll see more data being stored in an accessible format, with the ability to be easily interrogated with automated tools. Skilled staff will be able to conduct research, and spend more time on insight and less on administration and error-checking. They’ll be able to handle much larger quantities of data and better use processing power.
Automation is here to stay, whether we like it or not. For it to work well within market research, Mr Poynter says it’s crucial that the elements being automated are quality checked. “Buyers need to know what they’re buying,” he says, “If Toni and Guy started offering automated haircuts, I’d expect it to be to their standard. If it were Joe the barber, you’d want to see a few more people before trusting it – I can see it’s fast, I can see it’s cheap, but I need to have a few more processes to be able to assess it.”
Frédéric-Charles Petit, chief executive of digital market research company Toluna, says: “While some people may be concerned about what will be sacrificed as a result of speed and accessibility, we have found that these qualms can largely be addressed by the design of the tool. By making the interface easy to use and by making the data manipulation options comprehensive, survey providers can virtually eliminate any loss in quality that might result from automated tools.”
Mr Petit thinks the industry needs to get with the programme. “I think if we continued doing the same things that we’ve been doing as an industry for the last 50 years, we would not be an industry in ten years’ time,” he concludes. “You can’t advise clients on automation strategies when you’re not automating your own industry or your own practice.”