The focus group is dead. Long live the focus group

For many, the words ‘focus group’ conjures up a grey windowless room with fluorescent lighting, a two-way mirror and a self-selecting group of people willing to give up their time in exchange for a small financial reward.

For years, the industry has debated the future of this time-honoured research technique, thanks to questions over its subjectivity, bias in the selection process and time constraints, to name just a few. Yet at the same time, our profession is still heavily reliant on such research, and we continue to train our new graduates to use focus groups as a core methodology within their arsenal.

There’s no doubt that focus groups can prove to be a much-needed qualitative feedback tool; however, in today’s market, should a company ever base a new product or their brand’s marketing strategy on the opinion of 10, 100, or even 1,000 people? Instead, by using big data we can uncover similar findings, but based on robust data-sets numbering into the millions – which is why some commentators predict this as the beginning of the end for the focus group.

We believe, however, researchers shouldn’t ditch their traditional methods, but instead, develop and enhance them through combining with new data-driven techniques which offer solutions to some of the problems previously attached to them. Fundamentally, it comes down to what market research has always aspired to drive a truer connection with real human behaviour. Using data gives researchers the ability to solve the ongoing question of real v claimed behaviour, and so increase objectivity while limiting subjectivity.

One common criticism of focus groups is the bias introduced by researchers’ preconceived questions and discussion guides – they can only ask what they already know. Social data, on the other hand, gives brands access to pre-existing conversations.

By employing algorithms and data science to distil and surface naturally occurring themes and topics they can make use of the millions of genuine interactions within their categories, products, and brands taking place on platforms like Twitter. Adopting this technique enables researchers and marketers to surface unknown trends, via natural language processing (NLP) models, which can then be used to inform and define the qualitative research programme and identify questions researchers might not have known to ask.

The same is true of problems in the field of recruitment. Typically, after a briefing from the brand, a team of researchers would create a discussion guide from which they build a screener. Then a fieldwork team sources a small group of people from the general public who have opted to be included in market research and fit their criteria. At best you’ll end up with individuals who roughly match the requirements of the brand, but every stage relies on human judgment, allowing bias and subjectivity to creep in.

Social data can transform this process, by allowing brands to objectively detect and engage influencers who have materially affected the conversation across one or a number of mediums. For example, having used a data-driven approach to create a number of innovation territories for a leading tea brand, we then engaged and built a community of influencers to turn the data into product ideas, co-creating to develop an innovation pipeline for the business.

Using social data combined with predictive analytics can work incredibly well alongside traditional forms of research. The objectivity it can bring to question development and selection allows researchers to get the best out of focus groups, giving context to what the researcher sees happening in the data. By using a combination of technology with traditional methods, we believe that focus groups still have a bright future, so brands cannot only draw quantitative insight from 10 million people but also conduct valuable qualitative analysis using focus groups of 10.

Steve King is chief executive at Black Swan