ResearchLeaderOpinion excellenceSustaining the machineMind and brand relevance with the connected consumer Share this Opinion Leader
Mind and the machineIt may be immense, fast andmind-bendingly varied. Butresearchers must remember:Big Data can no more speakfor itself than the smaller sort. Share this Opinion Leader 2
Mind and the machineBig things can be intimidating. Research 90 percent of thecannot allow Big Data to be one of them.We stand on the edge of the most exciting andtransformative period in our industry’s history:90 percent of the data in the world today was data in the worldcreated in the last two years and “data taps”such as mobile, social and POS will continue topour out raw information for us to work withat an ever faster rate.However, if our response to the new era ofdata is to retreat behind number-crunching today was created in the last two yearstechnologies, then clients and indeed humanityas a whole, will be much the worse for it.It may be tempting to conclude that humanintuition must surely give way to computersand algorithms when it comes to keeping upwith Big Data. But now, more than ever, weneed to recognise the immense, unique powerof our own minds when it comes to dealingwith information – and deciding how to acton the basis of it. Share this Opinion Leader 3
Mind and the machineSo what do we mean by “Big” exactly? created and stored simply by virtue of thingsBig Data wouldn’t be half as intimidating if it happening. It’s broken free of human controlwere just a question of having more numbers – and therefore isn’t limited as to how big itto deal with. But Big Data is bigger than that. can get and how fast it comes at us. Data’sIt represents the coming together of several Velocity, the speed at which huge volumes of itdifferent themes, each of which would be fairly can be generated, is every bit as breathtakingparadigm-shifting in its own right. as its sheer size. And the speed with which it is available raises the opportunity and the demandFirst of course, is the sheer scale of the data to work with it in real-time.now being produced and stored. Walmartcurrently handles more than 1 million customer Yet perhaps the most challenging shift of alltransactions every hour, in databases estimated is that this size and speed is combined withto contain more than 2.5 petabytes. Such an an explosion in variety of data forms. Big Dataorganisation may soon have created more data comes in all shapes and sizes. Researchers areevery hour than research surveys have ever leaping on new types of data source – anddelivered. With data storage doubling every new types of source are leaping on us: fromyear, there appears no constraint on the amount mobile activity to Twitter feeds, geo-locationof information that we are dealing with. information, facial expression capture and much more. We are quickly moving fromConnected to the size of data but equally dealing in numerical scores to dealing in shapes,significant is the fact that it now generates itself. movement patterns, expressions – and humanData no longer needs to be created through a language. And such data does not come readilyquestionnaire carefully crafted by a researcher, packaged for analysis; using it must involveor painstakingly collected by a field agent; it is translating it as well. Share this Opinion Leader 4
Mind and the machineYou created it: you deal with it that we should try to solve, what we think the In his book The Signal and The Noise, USFaced with such challenges, it’s tempting to answers should look like, and what data forms election poll guru Nate Silver devotes a chapterbelieve that computational power, which has we can enlist to help provide those answers. to global warming and the fact that it wouldtaken the lead in creating this new world of And these judgements are human ones. be impossible to find any evidence of this ininformation, must also take the lead in defining the notoriously unstable climate record, werehow we deal with it. In this view of the world, In the Big Data era, the human imagination scientists not armed with a theory telling themthe researcher starts to look less like a person, continues to play an essential role in envisaging exactly what to look for – and which data tomore like supercomputer in a bunker: one what our many different data sources can be prioritise. It’s an important reminder that...where we simply have to feed in the right made to do, and in aggregating, translating and biggerquestion or combination of questions, plug it coding them to enable them to do it. To takeinto the river of Big Data – and wait for the a very simple example, Google can predict aanswer to pop out. But there are significantdangers to this approach. If Big Data ends up flu epidemic by spotting spikes in searches on cold and flu remedies. This is a tremendously ...thebecoming processed and commoditised data,then we are all in trouble. cool thing, but it only works because somebody realised that this pattern is significant – and data is, the more it needs to become that it correlates to something meaningful andDigesting really raw data useful. Similarly, micro-location data gives TNSIt’s a mistake to believe that data can ever a powerful new tool for mapping movementspeak for itself. Data always speaks with ahuman voice; it can’t say anything otherwise. around stores – but it is only powerful because we have established an understanding of what articulate.Every statistic that we deal with is the result these movements mean.of subjective judgement about the problems Share this Opinion Leader 5
Mind and the machineFrom data creators to data curatorsIn the old days (of six months ago), the rawnumbers that we sat down to analyse weren’treally raw at all; they were shaped by humanhands even before they came into existence.The art of designing a questionnaire involvesfinely balanced judgements on which questionsto ask and how to ask them. Whether to scorepreferences out of five, seven or ten can triggersome pretty serious debates with good reason –these things have a big influence on our abilityto spot patterns, make connections and providemeaningful insight. And judging how to ask of research, we must continue to apply the as linking spend and retention data to customerquestions should, of course, be closely related to same standards to independently generated experience surveys, as we already do at TNS.the challenge of what you are looking for. Big Data that we would if we had created it In the future, we will find more and more ourselves. And this will require leveraging much scenarios where the data we aggregate doesIn the Big Data era, we are no longer data hard-won experience about how data works. not include traditional surveys at all. In all ofcreators, designing the structure of information The skills that once went into the design of these contexts, it’s not just a question of beingfrom the outset; instead we are data curators, research instruments such as questionnaires will excited about what data can do. It’s equallyworking with information that has been remain crucially important in aggregating and important sometimes to step back, look at howgenerated independently. As such, we will selecting data sources, and deciding exactly complete and representative a given set of dataface many new challenges and require many how they relate to one another. For now, this is, and ask ourselves rigorous questions aboutnew skillsets. However, as we evolve the role might involve incremental improvements such what questions it is really qualified to answer. Share this Opinion Leader 6
Mind and the machineThe continuing evolution of analytics competitive relationships, owned and disputedAt TNS, we’ve already evolved from the era of territory and areas of opportunity. We thenad-hoc analysis, when researchers collected data build on this structural understanding with morewith little reference to how it would eventually action-oriented means of addressing the data:be used (and then looked through it in the “Groupings” to segment the subject matter andhope it would reveal something useful). Today “Drivers” to reveal the variables that influencethe design of the instruments for a particular relevant results, including causal connectionspiece of research is informed from the start by that can be far from immediately apparent.the challenge of how best to answer businessquestions. This approach may be structured, but it retains grounds for flexibility. It provides a checklistThe conceptual framework that we use for for where and how to look for patterns andany type of analysis reflects how the human themes. In the Big Data era, we will learn to and expecting machines to design them in thebrain naturally makes sense of information. look for different types of patterns in vastly first place. We must not fool ourselves thatThis framework consists of four different diverse forms of data, but human reason Artificial Intelligence (AI) is ready to take on theways of looking at any set of data, whether remains the key driving force in identifying them task of formulating questions and crafting theit was generated through research or arrived, and drawing purposeful connections between algorithms to answer them. After all, even thoseBig Data-style from independent sources. them. 157984631069555555345791326054 that welcome the concept of a technological“Dimensions” and “Landscape” address the singularity in which human-designed AIstructure of information; the first seeking Computational muscle can give research the 057879894455555555555204654765 surpasses that of humans themselves, don’tout common themes across a data set (the scale and speed that we will increasingly 573457913555555555555502106187 envisage it happening until at least 2045. That’skey themes defining a product category, for require in the Big Data era, but it is important a long time to wait to take real advantage ofexample), the second looking more closely at to distinguish between automating processes 141662046555555555555511579846 Big Data. 916305021055555555555120578798 676461157985555555552145734579 Share this 605487916305021061879214166204 Opinion Leader 765141676461157984637916305021 7 879465050312057879894465654214
Mind and the machineData and the human imaginationImposing structure on Big Data will throw upsome intriguing challenges – and these challenges Bigwill involve logical leaps and lateral thinkingfor which the human brain remains our bestavailable tool. What is a meaningful means ofscoring a positive tweet or Facebook rant? What Businessaspect of somebody’s location is actually relevantto the client brief – and what other sources ofinformation can be integrated or overlaid to givecontext to this information? The location of a car data solutionsby itself is meaningless. If it’s a car unable to fitinto the WalMart parking lot on Black Friday, itbecomes a whole lot more interesting.When we talk about deploying computational questions that we can ask, and the speed with dangers: that we define in advance what it mustpower in the Big Data era, we must therefore which we can answer them. Big Data can unleash look for and how it must look for it, leading tobe pretty clear about what we are asking the potential of human insight and human reason standardisation and blinkered, undifferentiatedcomputers to do. We must continue to exercise in ways never envisaged before. thinking, and that we confuse correlation withour judgement as to which information is valid causation, failing to exercise human judgementand valuable, and how its many varied forms can The greater computational power that will about which results are meaningful and whichbe coded in meaningful ways. As data curators, enable us to make the most of Big Data must are not. The challenges of the Big Data era will bethat’s our job. But by unleashing the power of be harnessed to an expanded role for the challenges for the human imagination and humantoday’s machines we can dramatically increase human mind. Depending too much on non- judgement as much as for IT infrastructure.the scope of data that we can use, the range of human processing power creates two potential We need to welcome them as such. Share this Opinion Leader 8
About Opinion LeadersOpinion Leaders are a regular series of articles from TNS consultants, based on their expertise gathered throughworking on client assignments in over 80 markets globally, with additional insights gained throughTNS proprietary studies such as Digital Life, Mobile Life and the Commitment Economy.About TNS About the authors Mark Kingsbury is Global Head of Marketing Sciences,TNS advises clients on specific growth strategies around new market entry, innovation, brand switching and leading the TNS Advanced Analytics teams around thestakeholder management, based on long-established expertise and market-leading solutions. With a presence world and driving initiatives to ensure that TNS definesin over 80 countries, TNS has more conversations with the world’s consumers than anyone else and understands and delivers best practice approaches to addressingindividual human behaviours and attitudes across every cultural, economic and political region of the world. our clients’ business issues. Mark began his careerTNS is part of Kantar, one of the world’s largest insight, information and consultancy groups. at Research International and held senior roles at Quadrangle and Ipsos before joining TNS at the start of 2012.Please visit www.tnsglobal.com for more information. Bob Burgoyne is Development Director, GlobalGet in touch Marketing Science at TNS. Bob has a backgroundIf you would like to talk to us about anything you have read in this report, please get in touch via encompassing both digital and analogue researchenquiries@tnsglobal.com or via Twitter @tns_global methodologies, answering business questions for primarily mobile sector clients in both emerging and developed markets. He now focusses on non-traditionalReferences research techniques and data sources, exploring the1. Nate Silver, The Signal and the Noise opportunities which they offer to provide better (richer,2. Daniel-Kahneman, Thinking Fast and Slow quicker, more exact) answers to our clients’ questions. Share this Opinion Leader 9
Pat Estes