When Arthur Sadoun became CEO of Publicis in June, his initial comments were, “This is a race. It’s a race to be relevant.” He’s not wasting time. His announcement during the Cannes Lions Festival of the development and prioritization of Marcel, an Artificial Intelligence (AI) powered “professional assistant”, became the story of the Festival.
The increasing presence of and recognition of data driven campaigns at Cannes reflects the recent trend towards new ways of working centered around data and AI which are set to transform the industry.
While creativity remains a powerful tool, there is no doubt that data and technology are changing the way brands interact, engage with, and communicate with people. The Market Research industry is not immune to the shift.
The growth of behavioural science, the development of new passive techniques, and the growing amounts of data available are great new sources for insights about brands and communications. With all this rich information, one can even wonder does market research still need to ask questions of people? Or, is asking questions becoming old-fashioned?
It is a relevant question, and there certainly are limits to the types of information that question-based surveys can reliably deliver:
- Daniel Kahneman’s seminal research on System 1 and System 2 thinking proves that much of what drives brand choice happens with little to no thought. Using only questions based on System 2, conscious, cognitive answers will not provide a complete understanding. Integrating neuroscience based techniques is required to provide better insights on how people engage unconsciously.
- With digital communications, data is abundant. People are tracked. Technology gives feedback about campaign delivery. Programmatic makes campaign delivery more targeted. The result is a more controlled exposure, with more granular details of campaign delivery, for each platform. This is more precise than relying on what people would remember about their surfing behaviour in a survey-based post-assessment tracking programme. Who can remember, given decreasing attention spans, even a couple of days after, where and when they saw a specific digital brand campaign?
- At the same time, the latest research-on-research confirms that questionnaires need to be shorter and more engaging (learning from gamification): people spend less time answering, or they drop out of overly long questionnaires without completing. Long questionnaires with repetitive attribute lists and many complicated questions that presume that people can answer like researchers or advertisers, are over (almost).
- When discussing ongoing tracking programmes, the challenge from clients is often ‘do we still need to run this programme or can we gather the same information from social data at a lower price?’ After all, social spaces generate a lot of data. For example, people spend more than one billion hours per day on YouTube. There are more than 1.15 billion daily active users of Facebook. More than 500 million tweets are sent per day. In New Zealand at least, volumes are much smaller and more limited in the insights that can be provided.
With all this data readily available, why are survey based approaches still relevant?
The pendulum is swinging back towards accountability.
Last year after the success of Brexit and Donald Trump, opinion polls were criticised, and some claimed that opinion polls were dead.
However, the recent French elections demonstrated the ongoing value of opinion polls. A purely social-based analytics trend predicted that Mr. Fillon would be the final candidate. Another System 1 approach claimed that Mrs. Le Pen had more chances to win.
Bienvenue Mr Macron. Ipsos’ polling accuracy during the 2017 French Election was the result of its strong survey based approach supported by an experienced team that offered an added-value interpretation of results, trends and relevant scenarios. Strong opinion polls are always enriched by qualitative research using various techniques, including social sources. These qualitative components enhance, but do not replace the quantitative survey program.
Accountability is of utmost importance in digital, where uncertainty has been the norm (despite all the available technology and data streams). Last year’s buzz after Facebook’s announcement regarding the errors with its viewability calculation demonstrates that even the biggest platforms have issues analysing and interpreting this data. As one of the strongest advocates for digital accountability, P&G’s Mark Pritchard’s one year window to clean up the industry is almost over. “The days of giving digital a pass are over, It’s time to grow up. It’s time for action.”
Furthermore, metrics provided by technology do not always tell the whole story: sometimes they are not the right metrics and often they are not predictive of real in-market brand impact. Beyond the hype there are still unanswered questions about alternative research techniques. None of the new methods on its own can provide all of the answers.
So what now?
To quote Keith Weed, Chief Marketing and Communications Officer at Unilever, “making real, lasting, impactful difference is not achieved by quick, reactive fixes. It requires systemic shifts in the way we work, thinks and create.”
Rather than arguing the merits of old versus new methods, let’s discuss:
- What your brand is trying to achieve. That should drive any sound research design.
- What each tool, approach or method delivers in terms of information, data and insight. This is good practice to drive the necessary change needed in research programmes.
- How to integrate new information, data, and new insights. In an increasingly complex world, integration is the sweet spot. Ultimately, we do not see the value in choosing to use one approach over another. Driving integration of multiple data sources is our new normal to push towards actionable insights that help marketers and media make the right strategic decisions.
- Integrating neuro tools like EEG and IRT, and including facial coding as standard alongside our validated pre-testing techniques to demonstrate the level of unconscious emotional engagement with the creative; see at which points engagement occurs and what kind of emotions are driven. This combination can guide clients to unleash powerful advertising.
- In digital creative testing, leveraging ‘Real Life’ research (i.e. live, in-site) while integrating survey and pure digital metrics such as viewability allows brands know how long their communications need to be viewed to have the necessary brand impact.
- Tracking is a very rich field to integrate multiple data sources. Social monitoring provides brands with early signals to build on or raises potential concerns to watch over time. The use of text analytics allows for insights to be uncovered from unstructured data sources in an efficient and consistent way. When needed, digital modules also complement tracking with digital tagging or retargeting techniques.
And these are just the latest examples. At Ipsos, we have a long track record of integration of new techniques, from qual and quant, structured and unstructured, passive, observed or solicited. We are agnostic in the way we chose and apply these techniques.
It is about applying each tool, approach and method to the relevant purpose to better understand brands and communications. Survey research gives meaning to observation and adds relevant contextual value for the brand.
And in the context of an ever-changing media environment, be prepared for the change!
- This article was written by Ipsos global communication insights leaders Charlie Ballard and François Guerin.
- Contact Jonathan.Dodd@Ipsos.com