Think Big – that's the theme of this year's ESOMAR Congress. The demand for data keeps increasing and it needs highly trained professionals to help their clients to make sense of today's 'data noise', said Silke Muenster, VP Market Research at Philip Morris and chair of the programme committee.
Tim Harford, columnist at the Financial Times and the opening keynote speaker reminded delegates that there are two ways to success. One is to look at every individual aspect of an issue and try to improve each of them marginally. The risk of failure is low and, put together, these individual improvements once aggregated can add up to a winning margin.
But no matter how many small steps are successfully taken, nothing new will be created and only something that already exists will be improved. He compared the small step approach to short term intellectual gratification, with little risk and constant success, whereas thinking big means taking risks, which must include the acceptance of failure. All very well but isn't the role of market researchers to help clients do the right thing and avoid mistakes?
Why researchers need a 'messy mind'
During a 'Market Research at the Crossroads' morning session Liz Lonergan (Nokia) and Leanne Tomasevic (Truth Consulting) called for a less structured approach to problem solving (or a 'messy mind'). They said that MR professionals should ask themselves whether greater robustness of data really provides more insight or whether another 500 respondents will really say something new. It was probably meant to be a little provocative.
Kristen Hickey (of ruby cha cha) opined that the energy of the relationship between clients and suppliers is still central to success. She called for the re-birth of renaissance men, a research polymath, who can combine technique and creative thinking and take a holistic view.
Adam Riley (DVL Smith) presented his thoughts on the future capabilities of MR executives, based on a recent survey of marketing decision makers. Professional interpretation and data gathering needs to be enhanced by taking a holistic view and providing strategic foresight, with less focus on the 'how' and more emphasis on the 'what outcome'. It was maybe telling that technical skills did not feature prominently (but maybe that's taken for granted?).
He also identified six archetypes of executives, namely the Storyteller, an engaging communicator; the Business Driver, a results-driven leader and entrepreneur; the Research Technician; the Big Issue Thinker, a strategist with foresight who uses multiple data sources; the Consultant, who acts as a thought partner to senior managers; and the Conceptual Creative.
Shoppers are big business and Adrian Sanger (Nielsen) focussed on the opportunities of digital technology to influence purchase decision and e-commerce to facilitate transaction for CPG brands. Assessing the many digital touch points out there, he said their importance varies between product categories and consumer segments.
Koen de Vos (Mastercard) and Orlando Wood (Brain Juicer) looked at the purchase journey of digital shoppers, which today resembles more a complex purchase zone rather a traditional linear purchase path. A happy experience in the purchase zone can influence repeat visits and lead to longer dwell time and higher spending. What's more, for higher purchases credit cards are used more often than debit cards.
Thaddeus Fulford-Jones (Locately US) analysed offline and online holiday shopping behaviour of his fellow citizens (Thanksgiving and Cyber Monday). He uses location (GPS) analytics combined with short 'in the moment' mobile phone surveys.
Heather Dougherty (Experian Hitwise US) and Maria Domoslawska (Research Now) combined multiple survey methods (web, mobile, and location tracking) into a complex single source study to investigate the role of mobile phones in the shopping process. Consumers are indeed using their mobiles in-store for a wide variety of purposes (including – who would have thought – to make phone calls) but rarely use a store's mobile app which they had downloaded.
Real Research Issues
Day two opened with a focus on methodological and technical issues – an area commonly under-represented at MR conferences.
Annie Pettit and Melanie Courtright (Research Now) presented their findings on respondents' usage and understanding of scales across cultures and survey mode. Their results are based on over 20,000 respondents spread across 12 countries. Differences were found depending on questionnaire design; for example, when collecting data via the mobile internet, sliders are less likely to result in mid-point answers compared to a grid layout.
Older respondents are more likely to use the extremes on the scale, as do respondents from Brazil and Mexico, whereas Japanese respondents tend to cluster around the middle of the scale. What wasn't quite clear whether these were actually real differences or due to the survey mode and the scale used.
Nielsen's Pete Doe presented possible solutions to overcome biases of convenience samples, i.e. non-random samples that can't be projected to a universe. The author compared results from Nielsen's Megapanel (convenience sample) with its probability-based people meter sample. The convenience sample significantly overestimates unique online audiences and time spent online. Weighting or using respondent level matching with a known probability sample helps to reduce the bias. A combination of demographics and media usage (i.e. online usage, number / type of sites) seems to work best.
Kristin Luck (Decipher) analysed some 120 surveys that were conducted via mobile phones. Whilst this mode is ideal for short micro surveys, problems can arise with app overload and app distribution and she sees the future for this type of survey research using the mobile internet, but also warns that response rates can be lower when using mobile, and drop-out rates can be high on mobile for longer surveys.
Big Data or more intuition?
Stephen Needel (Advance Simulations) presented a clear and cogent summary of Big Data's shortcomings. Proponents of big data, so he argued, can be very evangelical and have the vision that everything and everybody is being connected. Unlike behavioural economists who assume irrational decision making, big data supporters claim that (purchase) behaviour can always be explained and is consistent (for example, the same weekly shopping basket). What's more, too often correlation is taken for causality and simply having access to more data doesn't necessarily increase representativeness. Data crunching companies simply do not always have the expertise to interpret.
When it comes to the question of automation and machines on one hand and human analysis on the other, there is an interesting parallel from which the MR industry could learn. I had a chance to talk to Gary Kasparov, one of the Congress 2013's keynote speakers, who also faced having to play chess against a machine (his last match against a machine in 2003 ended in draw). He is not against automation but, in his opinion, the best results can only be achieved using human intuition. Of course, human intuition can also get things wrong and that is why we often hope that machines and algorithms will prevent us from making mistakes.
Other sessions covered issues such as corporate social responsibility (Dominika Maison, Maison Research) and sustainable research (Fiona Blades, MESH Planning). A further session was reserved for more specialist methods, including facial coding; word associations, implicit tests and eye tracking; and emotion-based visitor research at the newly opened Van Gogh Museum in Amsterdam.
Robert Ellis (COG Research) reported how those who use multiple screens when watching television have a more dynamic relationship with TV content by using email or Twitter to connect with other viewers. These multi-screeners are also more likely to stay in the room and see more TV commercials; for example, ad avoidance reduces from 29% (single screeners) to 19% (two screeners).
Ruihong Tang (Brain Intelligence) and Caroline Ji (VivaKi) presented their findings on the differences between traditional TV and online video viewing and the resulting consequences for advertisers. To provide unified and comparable metrics for touchpoints across different screens and devices, they opted for the use of EEG and eye-tracking technology and came up with Visual Emotional Scores (VES), which gave online video the upper hand over TV.
At Viacom (Christian Kurz) research impacts everything the company does and how it develops its brands. An on-going study of the two-billion-strong 'Millennial Generation' (those born between 1983 and 2003) across 32 countries with 20,000 respondents found natural disasters, terrorism and, especially, the state of the economy, to be the top issues that have an impact on them. They see guarantees disappearing (employment prospects, earning potential) but they are still happy, particularly when they are with family and friends (community), and generally positive and optimistic. Maintaining local traditions is becoming more important. At Viacom, the findings from the study are mainly used for programming and content development.
Erkan Balkan (PepsiCo) and Joeri Van den Bergh (InSites Consulting) surveyed Turkish Millennials to understand their attitudes and demonstrated how their research was used to re-invigorate one of PepsiCo's Turkish brands.
We finally returned to the topic of big data: Will Goodhand (TNS) gave an upbeat presentation on the vital role of the MR executive in interpreting and making sense of Big Data, but he also warned that huge efforts are required to clean the data before it can be used.
Finally, Nadines Guhlich (SoundCloud) and Alistair Hill (On Device Research) presented examples how they enriched their big data sets (gathered from site users) with survey research to collect detailed demographics and an activity diary.