MORNING DAY 2
Day 2 of the ESOMAR Congress 2013 kicked off in Room 1 with BIG IDEAS, Improving Solutions with presentations addressing new insights in the world of the market research surveys. This was followed by the second keynote of the conference, Galya Frayman Molinas the President of Turkey, Caucasus & Central Asia for The Coca-Cola Company. Over in Room 2 we started with BIG IMPACT, Transforming Lives, where the presentations focused on the more altruistic and social aspects of market research.
ROOM 1: BIG IDEAS – Improving Solutions – Elina Halonen
Multimode, Global Scale Usage: Understanding respondent scale usage across borders and devices
The second morning of congress started with a bang – or actually two. Following the heart-attack inducing drummers that suddenly burst into the room to startle the still slightly sleepy crowd, Melanie Courtright and Annie Pettit from Research Now kicked off with a weighty subject: scale usage and response styles, which refer to a person’s tendency to systematically respond to questionnaire items regardless of content. Understanding these tendencies is crucial for quantitative researchers, as our job is to find out what people in think so we need to be certain that the differences we observe are not an artefact of the scale that we’re using.
Melanie and Annie looked at response styles from three different angles: differences between data collection mode (web vs. mobile), demographics when collecting data through a web survey (age, gender, country and different types of scales) and social media scaling. To the relief of the many mobile survey advocates, there were no great differences in response styles between the two modes although using sliders in a mobile survey can result in slightly different data. Within web surveys, there are no significant differences between genders in response styles, so we can be confident that any differences observed between men and women are indeed true! On the other hand, as people age they tend to go more extreme in their use of scales (at around 55-64 yrs), which could potentially be due to becoming more confident in our opinions as we get older.
However, comparing different countries the differences become more dramatic with Brazil and Mexico leading the extreme response style charts while UK and Japan place towards the bottom with more considered use of scales. We should therefore be careful when comparing results from different countries, and potentially correct for these differences if possible. On the whole, using scales with more points is preferable to those with fewer points, and labelling the scale appears to help as well. With these differences in mind, the social media part was conducted within English-speaking countries only. On the whole, while men tend to be a bit more negative in their expressed sentiment whereas women are more positive, the differences slight enough not to cause concern for social media analysis. Nationality makes a difference here as well with Australians being the most positive.
Overall, there are significant differences in response styles but the evidence is not consistent or conclusive enough yet to determine whether mobile is better or worse as a data collection mode. And while there are considerable differences in response styles between countries, varying response options helps. Annie and Melanie’s presentation was certainly one of the best of the congress, and the kind of genuine knowledge sharing there should be more of at our industry events.
Best of Both Worlds? Can we make convenience samples representative?
Following on from response styles, Pete Doe from The Nielsen Company looked at ways of quantifying bias in an online measurement convenience sample as well as assessing the success of bias reduction methods.
To explore the topic, he compared a convenience sample from the Nielsen MegaPanel to a probability sample, and looked at their online activity audience and time spent for top 2000 brands and 200 channels. The convenience sample showed huge overestimates of online activity (up to 127%!), which provided a good opportunity to assess bias reduction approaches such as weighting where convenience sample is weighted to aggregate target levels from a representative sample, and sample matching using data fusion with weights from the representative sample.
Overall findings strongly suggested that weighting by demographics alone only remove about 33% of bias which still leaves far too much room for variation, but adding behavioural measures helps to reduce it. The key message from Pete’s presentation was that we can significantly reduce bias with weighting and sample matching, but demographics are usually not enough and we also need directly relevant control variables. However, we can have too much of a good thing: excessively controlling our data will only deliver diminishing returns and there will still be pockets of bias in our data. Overall, a bias-adjusted convenience sample is a practical research method, but as it doesn’t have an underlying statistical theory like probability sampling, we need to decide case by case whether the approach meets our research needs.
Measuring Up: Impact of mobile and segmentation on respondent behaviour
After some rather heavy duty stats, Kristen Luck from Decipher returned us to some more practical issues in mobile research and especially the impact it has on respondent behaviour. Kristin reminded us of the rising importance of mobile, but at the same time noted that not all mobile research occurs on apps due to distribution and panel issues as well as app overload. Increasingly
Using a case study from a research project with eBay with 1.5 million respondents, Kristin looked at the impact of recruitment methods on survey participation on mobile devices and the impact of mobile users on data quality as well as ways in which we can make mobile survey users’ experience better. In a panel sample, respondents already have an expectation of receiving surveys so they’re more prepared to start it on their preferred device which is usually on a PC. In contrast, when using customer lists, respondents don’t have this same expectation of having to do a survey so they are more likely to take it on their mobile device.
Furthermore, when respondents are taking surveys, they are more likely to drop out if they are doing it on their phone compared to a PC or even a tablet. However, there’s further variation between different phone types so it’s important to make sure a survey is compatible with different devices. Overall, the first page of the survey should be clean and easy to understand to encourage people to take it. Even more so than on web surveys, grid questions can be lethal to completion rates on mobile (up to 40% dropoff). Finally, it’s important to remember that populations from different devices might be different because of inherent demographic or psychographic differences, which would suggest that there is still so much more to know about mobile research.
Speaking the Language of Research: Why languages matter when conducting global research
Last in the morning session but definitely not least, Nikki Lavoie from Sky Consulting moved us into the world of qual. Nikki raised an important topic about the impact of language when conducting global qualitative research: can we conduct qualitative research in English in other countries? Some of the audience seemed incredulous that such practices might indeed exist, but Nikki assured us that this is in fact often the reality in many studies.
With only 2 in 5 Europeans claiming they are “comfortable” having a conversation (of any type) in English, limiting your study to English-speakers only is the very definition of a convenience sample. In addition to issues around the representativeness and relevance of speaking to such a sample in the first place, using a sample of non-native speakers raises many other, more serious considerations. The respondents will be more limited in their ability to express their thoughts as well as in their comprehension of nuances of questioning.
As well as potentially being less comfortable in expressing themselves in a foreign language they don’t speak fluently, emotions are also experienced and expressed differently in different languages, so as researchers we’re not getting an accurate read of the more implicit aspects. As a result, the research often takes longer, needs to be redone or provides no clear direction, which should surely be enough of a deterrent for anyone to reconsider doing their qualitative research in English.
Even using fluent speakers doesn’t solve the issue as research suggests how and why someone learns a language affects their understanding, in addition to the unknown variable of how much increased exposure to another language and culture affects one’s responses. Bilinguals often experience different selves in different languages, expressing different traits in each language (for example, more extroverted in one and less so in another). As a bilingual myself, Nikki’s presentation strongly resonated with me – even as a fluent English-speaker I feel different when speaking English compared to my native Finnish, and it seems somewhat crazy that anyone would even consider using non-native bilinguals for research. Let’s hope that after Nikki’s presentation even they would reconsider doing it again.
ROOM 2: BIG IMPACT – Transforming Lives Part 1 – Lucy Davison
We started with Dominika Maison from Mason Research in Poland who took us through “How research can help build a successful CSR Campaign”; a study about teenagers and alcohol, designed to encourage them to have a responsible attitude to drinking. I don’t think that the irony that her client for this study was a Polish Brewer owned by Heineken was lost on her. Dominika took us through a complex research programme involving qual, quant, online bulletin boards and a follow up evaluation, in which we discovered (not at all surprising to parents of teenagers) that 80% of 12 – 16 year olds had tried alcohol, 42% had been drunk and 36% had suffered a hangover. The problem was getting this group to share their experiences and talk about their alcohol use – hence the multi-methodological approach.
An abstemious anti-drinking campaign was clearly not going to work so they focused on helping the teens drink responsibly – these teens had no negative perceptions of drinking. A segmentation revealed the overlap between ‘responsible’ drinkers and ‘risky’ drinkers and the key moments when teens needed support to stay in the responsible camp. The resulting campaign involved interactive workshops including parents, to raise awareness of the negative consequences of alcohol, to instil positive drinking patterns and to give advice on how to refuse it.
By understanding the needs and values of this challenging target group Domenika was able to show how the year-long campaign changed behaviour. And, as one questioner pointed out later, to support her brewery client she could at least get the teens to switch from drinking vodka to beer!
Next up was Christian Bourque from Leger Marketing in Canada with his tale about encouraging parents to read to babies. In “Communicating on Early childhood Development” I was gratified to hear that it is as important to play with a baby’s book as read it, as my children frequently ate books when under the age of 1. Yum! His client Foundation Chignon works to prevent poverty through educational success by supporting young families and the goal of campaign was to encourage parents to stimulate young infants in order to improve early childhood development.
Christian looked at past mass communication exercises such as those for drinking and driving, wearing a seat-belt and anti-smoking campaigns for example. These all had a simple message – ‘stop doing this or you will die’. But, trying to stimulate parents is harder, the issue this project was dealing with was way more complex and without a direct causal chain. Parental self-efficacy is very high – all parents think they do well but think society doesn’t recognise the stress and time pressure they are under (tell me about it!). There is a general desire on the part of parents for their child to be happy but they will not respond to the direct risk or guilt-driven approach to change behaviour.
Christian took us through a four year integrated qual and quant programme with the same team of experts involved throughout, and on-going evaluation. Through helpthemgrow.com they delivered a very simple message showing importance and impact of engaging with your child from a very young age and were able to change the behaviour of parents in Canada.
Now, showing an adorable baby reading with his dad is pretty much guaranteed to get an audience to support you, and Christian had it in spades! Plus I am sure the typical MR audience finds the idea of reading to babies brilliant, so he was preaching to the converted. Even so it was inspiring to see how research used in this way can have such positive impact.
I am not going to discuss the questions here as my co-blogger from the session is keen to do that, but suffice it to say that there was a lively discussion and lots of questions. The most interesting for me was the question that if these social programmes can have a positive impact on behaviour why don’t we try and improve our clients’ products too – as in make them less harmful? Can MR do more to make itself truly ethical? Food for thought…
ROOM 2: BIG IMPACT – Transforming Lives Part 2 – Tom Ewing
The study was conducted on behalf of the Family And Child Trust, and looked at what politicians have christened the “squeezed middle” – low to moderate income families rather than those in poverty. This demographic is an electoral and political battleground in the UK – Prime Minister David Cameron has repeatedly talked about how families need to be at the heart of policy, so the impact of Cameron’s austerity policies on said families is a particularly timely topic.
Of course, under austerity research budgets get squeezed too. Ipsos MORI’s research design achieved a lot on very tight resources – a 13 month study of 11 families, with visits, phone calls, and diaries recording finance, well-being and time data. Material that turned out to be minor in project terms – like photos of what participants felt was important to them – were crucial in terms of building trust and encouraging openness.
For the findings, the word Perry and Perreira repeatedly used was “precarity” – a term popularised by social justice campaigners protesting the rise of temporary and ‘flexible’ labour, here applied to family lives Perreira described as “strikingly fragile”, even if the families themselves didn’t realise it. Even the best off participants in the study were only a shock or two away from cash flow difficulties, and as with labour precarity, it takes an emotional, psychological and physical toll.
The metaphor Ipsos MORI chose for this state of affairs was a game of snakes and ladders, with “4Cs” as the snakes creating precarity. First was the rising cost of living, with an ever-increasing chunk of household income being spent on essentials. This squeeze increases the importance of the second C, credit – easy access to credit being crucial for families, but also the source of a vicious spiral of debt. The third, cars, was a proxy for major unexpected costs and bills – a car breaking down can be a financial disaster. And the fourth C, childcare, helped prevent the cycle being broken – childcare was seen as neither affordable or flexible enough to help struggling families.
On the more positive side of the ledger, Ipsos MORI identified the ladders families could use to emerge from precarity – stability of income and expenditure, support inside and outside the family, a comfortable environment and the opportunity to spend time together – time, in fact, emerged as the critical third resource threatened by austerity, after money and well-being.
These were somewhat gloomy findings, but the research was widely picked up – even in the broadly pro-austerity tabloid press – and has helped to frame the conversation around the effects of the recession on British families. Like the earlier presentations in the session, It was an excellent example of research that is not only done well but helps to do good.
From the squeezed middle classes of Britain to the emergent global middle class – Ipsos MORI were followed by Fiona Blades of MESH Planning presenting on “Sustainable Research” and community work MESH had done with PepsiCo in Sao Paolo in Brazil. With a World Cup and an Olympics coming up, Brazil is something of a marketing honeypot right now – an awful lot of marketers are sniffing around it for “insights”. This could be a research boom, but as Blades said, there’s a responsibility to do research in a sustainable manner, not an exploitative one. A quick incentive won’t cut it.
What did MESH and PepsiCo create instead? Blades presented an approach which built from simple principles into something fresh and inspiring as she went on. Anything that touches on Corporate Social Responsibility is going to run into the deadening, worthy language of the area – “performance with purpose” in this case – and hit questions of credibility. But MESH’s idea of sustainable research is built around a simple principle – that, particularly in developing markets, the relationship between commercial researchers and the researched is fundamentally unequal, and that to redress this we should approach participants at the network (or community) not at the individual level.
In practise, this means building research communities within existing physical communities, and letting those communities dictate the ways in which both agency and ultimately client will help the community in a long-term, sustainable way. Each of the partners – client, agency, and community – needs to be giving and getting back from the relationship.
How is this different from Corporate Social Responsibility initiatives, which also like long-term investments in communities? The crucial difference is that it’s marketing, not research – the investments mustn’t be centred on the brand (which means they don’t need to have the photogenic feelgood-factor that can make CSR unevenly distributed). Instead of a fuzzy glow and a nice write-up in the annual report, the brand gets real, intimate, ground-level details of people’s lives through a community lens: something PepsiCo has already found useful with a deeper understanding of how the traditional family in Brazil is changing. The agency plays its part by offering training and even internships at a community level, and the community itself helps the agency by recruiting new participants.
This felt genuinely new, and inspiring. The emphasis on sustainable communities is reminiscent of MROCs – similarly long-term efforts which aim to build research relationships and shift control to participants – except the real-world benefits are more tangible, and prioritised by the community itself. A mended fence in a Sao Paolo sports park was a potent symbol of the better relationship between researcher and researched that MESH’s approach promises.
Like any new philosophy, there are issues and teething problems, some of which came out in a vigorous Q&A. For a start, sustainable research needs to deliver sustainable ROI to agency as well as client, and it requires quite an investment. Finding the right community leaders – go-betweens for community and agency – has proved tricky, particularly as their responsibility for helping allocate benefits can lead to cronyism.
Other issues raised exposed the tension on a panel split between social researchers doing their jobs and commercial researchers trying to do good: there is an understandable level of suspicion around any corporate intrusion in the social space. Standard commercial research platitudes about “brands bringing people together” aren’t great defences against hypocrisy. Every type of research brings ethical challenges, and they are very worth raising, but it’s equally worth celebrating the genuine attempts to innovate and do good seen in this passionate and fascinating session.
ROOM 1: KEYNOTE – Galya Frayman Molinas – Erika Harriford-McLaren
Dan Foreman – Highlights and Personal reflections
After 9 months in office, ESOMAR President Dan Foreman gave a great review of his personal reflections of the past year. At the beginning of his term, Dan made a promise to meet and connect with as many researchers around the world as possible to gain real insight into what is really on the minds of researchers. What he found was that we all want the some things – success for our industry, personal and professional development and the ability to work together to achieve a common goal. Seems obvious, but as Dan pointed out, it’s when people come together from over 75 countries that we realise just how connected we are regardless of borders. While there were some interesting things that happened at ESOMAR in the past year – over 200 new corporate members, 700+ new individual members (ranging in age from 23 -72), top-ranking in the GRIT report and more – it was talk about the random walk and connecting the dots backwards the really resonated. Recognising that it’s never going to be a liner path to get to where we want to be (or think we should be ), Dan reminded us that we must find value along the way with all that we do and by doing that (and not forgetting the legacies of those who came before us), we may just find that we are going to go even further and be even stronger than we ever anticipated.
Keynote – Gayla Frayman Molinas, The Coca-Cola Company
How does an iconic brand such as Coca-Cola stay relevant, loved and successful? According to keynote Galya Frayman Molinas, the rules of brand marketing are written and re-written everyday – determined by environments and localised lives. Being global brings unique challenges and raises expectations and it is because of this that Coca Cola is motivated to better understand and connect with consumers at a deeper level than just through the drinks they serve.
The world is transforming at an exponential rate. As Molinas noted, because change is constant, evolution is mandatory. Coca-Cola strives to go with the flow and connect with their consumers and does so by engaging the market, creating shared value and building and creating a network that their consumers can use to connect. So how is this done? Coca-Cola holds the belief that they have a responsibility to create value for everyones lives they touch. According to Molinas, it’s not just about shareholders anymore because there is much more to business if you want to be sustainable – for your product as well as to the consumer.
By recognising that culture is participatory – people are closer to each other than they used to be – Coca-Cola co-creates and builds with a wide variety of stakeholders in building a global culture. The borders have blurred and any brand that wants to remain relevant must find a way to insert themselves into the culture. So how does Coca actually manage to do this is? They create stories – which are without a doubt the new currency. In a cut and paste world – Molinas pointed out that these stories are what gives a company their real USP. Have the storytellers tell the narrative of your brand and become a true love-mark while also finding success. It’s about being liquid and linked and recognising that the balance must be found.
AFTERNOON DAY 2
After lunch Congress continued in Room 1 with BIG DATA, Big Deal? 2 presentations looking specifically at the use of big data, this was followed by the session BIG RETURNS, Creating Engaging Brand Experiences. Over in Room 2 the ESOMAR Young Research of the Year Award Finals were held, followed by the panel discussion – BIG ISSUE, Do We Need to Get Over Ourselves? The sessions in Room 2 ended with BIG SCREEN, But Who’s Watching, these sessions focusing on TV measurement. The conference finished in Room 1 with the penultimate keynote speaker of the conference; Oliver Percovich of the charity Skateistan.
ROOM 1: BIG DATA – Big Deal? – Tom Ewing
Steve Needel of Advanced Simulations began the session in dragon-slaying mood. Big data, he explained, was really a “small idea” inflated by our fears and hopes into something impossibly powerful. The increase in the volume, variety and velocity of data is undeniable. That it necessarily means something or must change how research is done, though, is far more arguable.
The “big data fans” Needel described are essentially cultists, repeating articles of unshakeable faith. Their cult believes in connectivity – people always connected and sharing data, and data connecting to reveal patterns unseeable at a merely human scale. To believe big data is useful, says Needel, you have to believe four assumptions. You have to believe behaviour is essentially deterministic – a particular cause will always produce a particular effect. You have to believe it will be possible to acquire and merge data on these causes and effects. You have to believe you have a workable paradigm for understanding the data you get. And finally, you have to believe you can do something about it once the pattern is understood.
Whether this is an accurate description of big data users is questionable – though it certainly seems to fit some commentators. Needel set about demolishing each of these beliefs. Behaviour is irrational and heavily influenced by moods and environmental factors, so deterministic models of behaviour break down. The ability to get the right data remains in doubt – an accurate behavioural path to purchase, for instance, would need to take into account a staggering amount of individual environmental factors. Big data sets are not magical – pattern recognition does not arise spontaneously, and nor do interpretative models. And the leap from data to action is critically neglected by the big data priesthood, with analytic first principles (“correlation is not causation”, for instance) blithely ignored.
Luckily, Needel saw a few positives. Talking about big data, he said, gets people thinking positively about the use of data in business, and once they do that it turns out that a lot of questions can be answered simply by experimenting more with your existing research data, or widely available large-scale data sources like purchase panels. It’s been a long while since these attracted anything like the buzz or capital of “big data” but the principles of research haven’t changed: develop a theory about how your products behave, test it and experiment around it.
Following Needel was a presentation that at first seemed designed to prove him right. “We are here to talk to you about something bigger than big data… disruption!” said Radhecka Roy of TNS Singapore, and she and co-presenter Sunita Venkataraman of Intel initially came across like the big data acolytes Needel had warned us about. Everything was new, disruptive, fast – after Needel’s dose of negativity, breathless stats about Instagram use or familiar examples like the Obama campaign’s data reliance seemed too lightweight. In the duo’s paper there is an example of how big data helped with a problem Intel had, but this was left out of the presentation, and its absence was felt.
But Roy and Venkataraman made some excellent points too – ones just as challenging as Needel’s. The key to their argument was speed – we live in a world where business decisions need to be made almost instantly, and research has to adapt to that. The precision or accuracy of “big data” isn’t its most important property – directional is good enough if it’s fast. Big data fans may be cultists, but Roy had a great metaphor, too, borrowed from Kantar boss Eric Salama. Researchers need to move from being farmers to chefs: ““When you typically farm tomatoes, you want every recipe to have tomatoes as a main ingredient”.
Their most controversial point, thought, laid into the comforting (for researchers!) notion that big data can never provide the “why”. “Why”, they said, is generally a psychological need – something people are curious about – rather than anything critical to business decisions. Think of your most recent decisions, they asked us, and think how often you actually needed to know the why of behaviour. This proposition drew particularly vocal reaction in the Q&A session – a sign of a nerve being touched! But for a lot of short term business decisions they surely had a point – time is the real cost of research, they reminded us, so the additional time spent getting to why won’t always be worth it.
Whatever your opinion on the “why” question, Venkatamaran and Roy offered some specific and practical tips on moving from farmers to chefs. Eliminate one backwards-looking research programme from your budget; only use short interviews; answer one question every day looking at the data you already had. Combined with an appreciation of the need to experiment and a distaste for using the past to predict the future – also challenged by the audience – it began to look as though Venkatamaran and Roy had more common ground with Needel than you’d think.
In fact, both presentations were at their strongest when they stopped trying to battle or ride the big data smoke monster – at base this was a session about how to do good research in a fast-moving world whatever the data size, and there the disagreements were far smaller. Venkatamaran and Roy believe in agile research – failing fast, embracing change, and moving from developing concepts to experimenting in the market. Needel believes in a sound theoretical base and using what you already know wisely. A combination of the two would be an excellent set of principles for any researcher, which meant this stimulating session ended on friendly terms.
ROOM 2: YOUNG RESEARCHER OF THE YEAR FINALISTS – Erika Harriford-McLaren
This year marked the 4th annual ESOMAR Young Researcher Award competition and the three 2013 finalists presented thought-provoking and innovative research concerning bi-cultural identities, migrant movements and the role of women in creating success in research and business.
The idea of a bi-cultural consumer was taken on by finalist Anne-Sophie Lenoir of the Erasmus School of Management . As diversity increases, so do advertisers attempts to reach minority populations with specific targeting strategies. However, as Lenoir’s research showed, this approach, whilst successful in some contexts, does not fully take into account the bi-cultural consumer – those second generation or further heritage cultures residing within a host land. She showed us that identity is not a uni-dimensional construct and that, in fact, inherent identity is not salient all the time. Therefore it is important to remember when targeting advertising through a heritage-cultural lens, that it may be less effective if the heritage identity is not strong anymore or lessened due to acculturation.
Nadita Mary Thomas of Kadence International followed along these lines in her presentation on the migrant dream and how it is changing, in particular in relation to identity as well as the impact of remittance. A World Bank report noted that migrant remittance in 2009 totaled USD 38 billion a year globally – nearly twice the amount of official development assistance provided by governments and international aid agencies. When it comes to remittance from migrants, Thomas informed us that there is seen to be a power in collective use of remittance funds – as it gives an opportunity to contribute towards immediate family and society. When looking at identity – Thomas found that migrants tend to identify themselves in two ways: the first being more along the line of “when in Rome” do as the Romans do and blend in and the second being a blend of both worlds and recognising that it is fine to have a balance between one’s native and adopted country identities. In the end, the recognition of the role of migrants and the importance of their contributions in home and host country need to be better integrated into our societies, because we must think big and be sure to observe the small, the invisible and the taken for granted.
Real growth is a real concern for our industry today and for Danielle Todd of TNS, the solution may very well lie in increased gender diversity and representation. Noting that the traditional business management style is male and that recent research such as that from the Davies Report in 2011 has shown that gender diversity is highly recommended for better business, Todd wonderfully outlined how increasing the gender diversity experience will create stronger growth and better profitability for those companies who take it on board versus those who still with the male-oriented status qou. Using a study conducted in her own company as an example (kudos to her and TNS for actually doing it and sharing the findings), Todd highlighted how perceptions of success and strength were connected with male identities, even in her office. In the end, there was a call for more of this research to support female diversity and professional development. Only time will tell…
ROOM 1: BIG RETURNS – Creating Engaging Brand Experiences – Jeff Henning
Do Emotions in Advertising Drive Sales?
Nick Langeveld and Rana El Kaliouby of Affectiva discussed their research with Mars into using facial coding to understand the relation between emotional ads and sales effectiveness. “Good ads solicit strong emotion, happiness, surprise, even crying a little. But does that emotion translate into a sales events?” asked Nick.
To study this, they had over 1,000 American, British, French and German respondents watch a number of web videos, from a library of 115 commercials for the categories of chocolate, gum, pet care, and instant foods. In turn, the webcam watched the respondents, tracking and coding their facial responses throughout the experience of watching each commercial.
Participants also provided self-reported data. Before and after they watched the ad, they answered questions about purchase intent and brand likeability. Commercials would change reported purchase intent but had little affect on brand likeability.
Affectiva used sales volumetric data for the four weeks before each ad aired and the four weeks during which it aired. Based on performance, they classified 36 commercials as “bad ads” (causing a decline in sales), 25 as “neutral ads” (no effect on sales), and 54 as “good ads” (causing an increase in sales).
Chocolate commercials provided the most positive emotions, while food ads – being more informative – produced the least. Viewers in the US showed the most emotion, and the most positive emotion, while viewers in the UK showed the least positive emotion.
Affectiva then built a model to predict short-term sales from facial codes, using sales volumetric data for the four weeks before each ad aired and the four weeks during which it aired. They found that self-reported data on the quality of an ad was 69% accurate in identifying the potential effect on sales, while facial coding was 75% accurate. Combining the two measures produced even greater accuracy.
Affectiva provided a testimonial from Laurent Larguinat of Mars, who said, “Combining Affectiva’s technology and Mars’ marketing expertise, we’ve significantly increased our understanding of the role of emotions in advertising. This knowledge will shorten the odds of producing advertising that improves our sales performance and perhaps even help us achieve ‘lightning in a bottle’.”
Old Meets New
Just as Affectiva found that combining self-reported data with facial coding produced better results, Carlos Jaime Velasco of Neurosketch in Columbia found that combining old and new research methods offers a better understanding of consumer behavior. “Sensory integration is a good metaphor for this: when the brain integrates sensory information, when we listen to other people and look at them, watching their lips, our brain takes that visible information and auditory information together to help us understand what people are saying. This reduces ‘noise’ to help us understand, and that is what we want to do with different research methods and different types of information to understand the complex world of consumer behavior.”
Is there a link between consumer reports, crossmodal correspondences and emotion? “Crossmodal correspondences are how we tend as humans to match information with different sensory modalities – for instance, high pitched sounds associated with sour tastes. Understanding these helps us design better brands,” said Carlos. To learn more, Starcom Mediavest and Neurosketch conducted three quite different studies.
The first was a survey of 80 people where respondents did word association exercises and shape symbolism (a sensory congruence test) for the logos and typefaces of three brands. Respondents had more negative reactions to Carrefour, a supermarket chain, and the associations produced by their logo and typeface were all over the place. Samsung, in contrast, evoked consistent results for its typeface, which independently was seen as evocative of technology; however, it had consistent opposite results for its logo. Coca Cola evoked consistent results for its logo and typeface, both of which were associated with rounded shapes and positive emotions.
The second study was facial expression analysis, categorised manually, for 25 respondents of three brands. Participants would sometimes say “I like their product” but make a face, indicating that they rejected the product.
The third study was of 30 people doing eye tracking on four packages. Too often in Columbia and Latin America, eye tracking is used “for the sake of using it…without a proper experimental design.” Starcom Mediavest and Neurosketch tested indentations in four different bottles, and found that those indentations drew the eye. “By combining eye tracking and word association, we saw what other things got the attention of people.”
To conclude, Carlos said, “If we keep measuring bits of the consumer, we will have incomplete answers. By combining traditional measures and neuroscience methods we will have something more accurate.”
How Does Your Cappuccino Feel?
Crossmodal correspondencies are especially important when evaluating something as transient and intangible as flavour. Dominique Delfaud of Mane Flavour and Fragrance in France and John Pawle of QRi Consulting in the UK discussed using synaesthesia to create “a visually interactive experience of flavour.”
Mane, since 1871 a family-owned leader in the flavor and fragrance industry, recognises four layers of flavour consciousness:
Self-reports can only get so far. “They can tell us if they are familiar with a flavour,” said Dominique, “and they are keen to let us know if they hate it or like it, but the memories linked to the flavour or emotions are really deep and not as articulated.”
To dig deeper, John implemented a “sensations methodology,” with 250 iPad interviews with participants, each of whom tasted and rated 6 flavors while being tracked by the iPad camera. “We use the principles of synesthesia, that all the senses work in harmony. The stimulation of one sensory pathway leads to experiences in a different pathways – colours as shapes, music as colours, tastes as textures, and so on.”
Respondents were asked to answer questions such as:
- “Imagine you have gone into the world of this flavour. Imagine it is all around and inside you, and imagine that you live in this world. What colors are in this world?”
- “Imagine the world of this flavour as a material. What does it feel like when the flavour touches you?”
- As well as what need-state they associated with the flavour: e.g., relaxation, energised concentration, pick-me up.
They were also asked to rate a brand using these same questions. “This presents opportunities for interesting portfolio management,” said John, revealing what flavours are evocative of the same colours, textures and need-states as the brand.
Dominique contrasted traditional self-reported data to a fossil of a pterodactyl, “flat, dry, dull, with just liking, familiarity, and poor verbatims,” while this technique provided “a multi-dimensional image of perception of flavour,” a pterodactyl alive and in flight. “This is really fascinating for flavorists and helps bring their products to life. This type of approach allows us to sneak into our consumers’ minds.”
“Too often,” she said, “brands choose flavors based on poor understanding. This gives them a much richer understanding. It is very useful information for the brand owners. When we think of the crossmodal associations, we can align the flavors with the marketing strategy.”
Emotion and Inspiration at the Van Gogh Museum
From thinking of flavours as colours, textures and shapes, the session moved on to the Van Gogh Museum in Amsterdam. Laurine van de Wiel of the museum and Saskia Brocx of TNS NIPO in the Netherlands discussed “how emotion-based visitor research can create engaging brand experiences.”
The irony of art museum research is that in the past it has been more about still life than sunflowers. “Art is emotion. It touches us, and moves us,” said Saskia. “Art bypasses rational response and goes to the heart of what moves us. Yet most art museum research is cold and quite boring data, about visitor numbers, easy to find locations, and clean toilets.”
TNS has done audience research for 80 other museums in Netherlands through the TNS Museum Monitor, a multi-client audience monitor. Each aspect of a visit builds to produce greater loyalty, from making a visitor think, to teaching them, surprising them, touching them, and inspiring them. Too often inspiration was what was missing from the art museum experience itself.
TNS used its NeedScope model to understand emotive needs, segment visitors, and evaluate current and competitive brand delivery. The methodology went “beyond functional needs to the deep seated emotive level to quantify them and identify strengths and weaknesses.” It includes a consumer brand relationship model, a psychological model, and a projective approach. “Instead of asking questions directly, we used projective techniques so that people can talk about someone else, to get to the core of what they really want to express about the museum experience.”
The research identified six different segments of visitors: Trendy Hedonist, Wilful Discoverer, Classic Connoisseur, Intellectual Specialist, Docile Admirer, and Easy-going Connector. “Should the Van Gogh museum try to offer something for everyone? Or should we create a unique brand experience?” asked Laurine. “We want to reach as many people as possible, and it feels counterintuitive to exclude audiences.”
After much debate, the museum decided to refocus from being all things to all people to excel at serving the Easy-Going Connector. “We were convinced that it was very important to set ourselves apart. If we were anything to anyone, it would be hard to deliver on expectations. We will create a unique brand that sets expectations and delivers to those.” The Easy-Going Connector wants “a shared experience with someone; something open, low key, accessible, that delivers a fun day out.” The museum is integrating this new positioning throughout its strategy.
ROOM 2: BIG ISSUE – Are We Losing the Ability to See the Wood for the Trees – Panel Discussion – Lucy Davison
Do we need to get over ourselves? …and stop worrying too much about representativeness as opposed to delivering new insights
Now, I don’t know about you but this issue doesn’t keep me awake much and I was quite happy when right at the start session chair Ben Page from IPSOS MORI suggested the answer might be “yes” and we could all just go to the bar. But what ensued proved to be a lively and interesting discussion.
Ben structured his panel from his left to his right running from those who felt that representativeness was really important (Reg Baker, consultant to ESOMAR Professional Standards Committee) to those who thought that you could get by with a bit of intuition and a lot of different data sources in order to keep up in the brave new world of research (Simon Braier from Nokia). In between these two extremes were Manish Makhijani, a client from Unilever who took a slightly more pragmatic view of the importance of representativeness than Reg; and Frederic Charles Petit of Toluna who suggested with true Gallic Cartesian style that the question was wrong in the first place and we should instead be debating whether the MR industry is prepared to create clarity from the new world of ‘messy’ global data.
The problem with these sessions is as a non-researcher I find everyone makes a lot of sense, and am usually persuaded by whoever is speaking at that moment. I therefore changed my mind 4 times. But after three minutes from each participant which Ben controlled with excellent precision, the debate opened up to the floor. Questions flow easier in the smaller room and this was no exception.
One thing I do know is that speed (or ‘always on, real-time’ research), and the impact of downward pressures on cost are both massive issues facing the industry. Do we really have time to build a model to work out how to deal with new data sources as they emerge and are adopted by consumers – and clients alike? The pragmatic argument – that we will muddle through with a bit of trial and error as we did with the move to online – seems to be both true for research and actually a relatively sensible way to go. There are of course industry standards which the leading sample providers adhere to but on a show of hands, only 50% of the room had heard of ESOMAR’s 26 questions on online sampling, which one helpful participant pointed out had now morphed into 28 questions anyway!
The comment from someone from SSI did sound familiar; that time and cost are not the only issue but also the way researchers do like to complicate things, this puts even more pressure on sample providers. But Ben was quick to avoid the session turning into a good old moan on the part of those who collect the data, about those they collect it for. Chris Dubreuil from Research Now suggested appending sample robustness with each study to convince clients of validity of data, but the parting shot from the floor came from an agency researcher who said the problem of representativeness only really arose when the client didn’t ‘like’ the results – which is when they question if it is ‘right’ or not!
The panel summed up in double quick time with Simon and Frederic suggesting the new data sources will only improve things by providing us with the ability to triangulate different sources and create greater clarity and with Manish from Unilever suggesting that we are facing the biggest change in the industry ever in the next five years with the disappearance of custom research and the emergence of a ‘continous stream of data’.
Of course as a woman I loved the fact that ESOMAR was debating the issue of representativeness of sample with a totally unrepresentative all male sample. Ho hum.
ROOM 2: BIG SCREEN – But Who’s Watching – Tom Ewing
Enter research, to put some facts in place. This session showcased two presentations on two-screening – one from China, one from the UK – but beyond the shared topic they were very different. They showcased different approaches – neuroscience in China, ethnography and behavioural experiments in Britain. More importantly, they framed the topic quite differently.
Caroline Ji for VivaKi in China took the adversarial angle – which channel was a more important part of the marketing mix, TV or Online? Robert Ellis for Cog Research in the UK focused on synergy instead – how to draw meaning from second-screening as a behaviour in itself, not a competition between two existing ones. The happy result was that two presentations that might have contained significant overlap were able to answer quite different questions.
VivaKi’s approach was rooted in neuroscience, since neuroscience cuts through the knot of conflicting metrics around two-screening to get to the source of response: the brain. They partnered with the Brain Intelligence Neuro-Consultancy, whose Ruihong Tang explained what they had measured. Using eye-tracking and EEG tools, they collected visual attention, emotional reaction, and memory encoding data from 120 people to supplement an observation study of 200 families. This would let them settle the question of which platform performs better.
The results were grim for TV – consumers, the presenters claimed, would do a lot to avoid ads and when you factored that inattention in, the online spots outperformed their TV counterparts. They also created a lot more positive emotion – in fact the same ads, shown to the same people, triggered more positive emotion online than on TV. Surely this was a glitch in the data? But no – and here’s where the specific nature of Chinese TV and advertising came into play. Chinese ad breaks are brutally long – 20 minutes – and the dominant style of advertising is very information heavy. In that context the same ad could seem a mild inconvenience on the second screen but a real chore on TV. VivaKi’s conclusion? Adjust budgets in favour of online – they suggested around 70%-30%. As with most neuroscience studies, whether you agreed was somewhat dependent on your trust in the brain science.
Cog Research, meanwhile, gave us one of the large-scale cross-modal studies which have been a feature of this ESOMAR congress. A big quant study, plus 1650 hours of ethnographic multi-camera footage capturing every nuance of attention, plus on-device data monitoring, plus follow-up interviews – oh, and throw in a behavioural science experiment on low-attention processing for good measure. Cog’s findings were beautifully captured – their presentation had the best, and best-edited, use of film I’ve seen all conference, letting Robert Ellis illustrate the broad, sensible conclusions with wit and humanity.
His basic theme is that two-screening is, at root, not necessarily a new behaviour. We have always multi-tasked with TV – eating, flirting, and fighting while the TV carries on regardless. Two-screening is a more dynamic extension of that. People love to talk about TV, and even about ads, and as Ellis said – “when something’s interesting, you look up”.
What two-screening gives us is more interaction. Ellis outlined three types of interaction – researching TV content, sharing and discussing it, and participating more fully in it. And three tiers – 1 to 1 with an individual, conversation within an existing social circle, and discussion among new communities. We’re already seeing quiz shows and even ads (like one for Sure deodorant which turns into a game once a QR code is scanned) take advantage of all this.
With that level of richness, multi-screening keeps us in the room, making us less likely to move for the ad break. But are we paying attention to the ads, with the delights of social media in front of our noses? Ellis’ point here was very important – we pay as much attention to things as we need to. Two-screening behaviour, he said, was “meerkating” – like a meerkat on guard, we quickly and unconsciously flip our attention between screen to screen, checking in this case for interest, not predators.
Here was where the behavioural science came in. Cog set up an experiment asking some people in a waiting room to watch a TV show and others to update their social media account (while the same show played in the background). In each case viewers were exposed to an ad break, but the lower-attention sample – the two-screeners – showed no real difference in ad recognition afterwards.
So multi-screening doesn’t detract from advertising – and Cog’s study has played a part in changing the narrative around it, from TV being portrayed as an old fashioned medium to multi-screening deepening and reinforcing TV. Both these presentations were strong contributions to a topic which is often full of hot air. Cog’s piece, in particular, deserves to be the definitive word on multi-screening for some time to come.
ROOM 1: Keynote – Oliver Percovich, Skateistan – Erika Harriford-McLaren
The closing keynote of the day was one of my favourites – not because he was the most famous speaker or because he was able to cleverly link what he does to research – although he has a research background. Oliver Percovich, of Skateistan, was one of my favourites because of the absolute sincerity and passion he showed for the children he helps and how he truly brought to us the reality of risk taking and the tremendous benefit that can be gained from it.
Skateistan, was founded by Percovich 6 years ago in Kabul, Afghanistan – a very unlikely place for connecting with youth and girls via a piece of wood and four wheels. But, as anyone who has ever skated knows, passion for the sport runs deep and when that passion is combined with a desire to change the world – suddenly you find that even conflict-ridden areas can provide the perfect place to start changing the world one child and one half-pipe at a time.
As we’ve all read over the years, getting an education in Afghanistan is not an easy thing and this is especially so if you are female. The unique thing about Skateistan is that it attracted girls immediately and because there was no prior skate culture in the country – it was easier For Percovich and his team to get buy in on female participation and community approval for their involvement. Percovich saw this as a golden opportunity and seized it and in doing so has made skateboarding the biggest female sport in Afghanistan. Economic and social capital from local support and abroad has allowed Skateistan to build a true academic system – with skate and educational facilities and enabled them to expand to another Afghani town as well as to Cambodia.
At the ESOMAR Congress in Atlanta last year there was a presentation about two Afghani researchers who were killed in the field. That tragedy and the Skateistan story resonated with the ESOMAR team. This Congress saw the introduction of the ESOMAR Foundation – established for support and funding of projects around the world – ESOMAR was proud to give Skateistan the first donation from the new Foundation. I hope the rest of the industry is compelled to do the same and support this unique and amazing initiative. I know I will…
Tom Ewing is Digital Culture Officer at Brainjuicer, Elina Halonen is a founding partner of the Irrational Agency, Lucy Davison is Managing Director at Keen Mustard Marketing, Jeffrey Henning, is president of Researchscape International and Erika Harriford-McLaren is Communications Manager at ESOMAR