MARKETING RESEARCH OF TOMORROW.....(Start again)….. – From Sacred Cows, Micro-Surveys and Expressive Research to Data Myning, Polymeasures and Futures Research

Professor Luiz MoutinhoFoundation Chair of MarketingAdam Smith Business SchoolUniversity of Glasgow,Scotland.

Time and again we see institutional blindness, compounded by dangerously misleading survey results. Consumers change their minds, faster

than market research can predict them….!

What we need is insight…., not just data!

• The Mindset of “Old Research” is fundamentally around asking questions….

• The Mindsets and Skillsets of “New Research” are more around exploring and interrogating existing datasets.

• The Golden Age of Research that is dawning is driven by new data sources that will allow us to focus on the drivers of human behaviour rather than the descriptors of that behaviour.

• The aim is to have a consumer who thinks, “I will invest my information (and time,effort and attention as well as my money) with you because you are good at using it for me, on my behalf”. It makes marketing research a service to the consumer. It requires the company (and marketing) to embrace a new role as consumer agent.

• We are only at the very beginning of this journey – from “brand building” and “customer relationship management” to “consumer agency”

• The new type of understanding is driven less by knowing about consumers and more by understanding with them!

LINGUISTICS

The team “market research” has been incredibly detrimental to the industry, as it focuses on the process and not the value of the data-driven strategy. The term “market research” has minimized the industry’s influence within the corporate decision-making structure by reinforcing the notion that market research is all process and data and that market researchers manage the process and date delivery, but not insights or strategy that flows from this process.

In a sense, “insights” is a much better term , because it concentrates the mind on the value derived from the data. But, “insights” as a term is still sub-optimal in that (a) it completely omits what will be done with these insights and (b) the strategy that will flow from them.

Old methodologies are getting new looks. Ethnographies certainly seem to be making a comeback, as do the kind of lengthy in-depth interviews needed for some of the newest collaging exercises. This trend is called “deep qual” or “qualitania”.

Deeper Insight into the Whole Consumer

While it is true that we have more data on consumer than ever before, we are still starved for a holistic understanding of the whole consumer. The only way to truly develop this full understanding fo the whole consumer is via ethnography, and this is why ethnographic tools are coming quickly back into vogue within the market research industry. Moreover, in a world where consumers are less and less likely to respond to a basic survey, ethnographic data may be an important part of the industry’s transition. Interestingly, the trends toward ethnography and hosted online communities may squeeze traditional focus groups from both sides, reducing future growth of this qualitative workhorse.

Key Mega Trends Transforming

Marketing Measurements

Observational Measurements

Observational research techniques-sometimes called digital ethnography-are not a replacement for more overt data-collection methods, like face-to-face survey, but they are an important addition when attempting to obtain natural, unprompted insights into the behaviour of consumers.

Co- Pattern Research

• Plotting Culture Networks reveals patterns and these cultures are Micro Factories of Meaning.

• Marketing Research should reveal Meaningful Patterns.• Patterns that reveal Movement of ideas through culture.• Expressed Signals as Transmedia Management.• Catalytic Insights will come from leveraging technology in

human ways.

Sacred Cows in MR that will become obsolete quicker than you can imagine…

1. In -Person Focus Group

2. Online Surveys. Micro-surveys. Modular data – fusion techniques. Geofence-driven in- the- moment mobile feedback. Indirect Measurement. Facial sentiment recognition. Mobile neurofeedback that’s the future.

3. The Quant / Qual Duality .Sequential and compartmentalised. Time compression. The words will collide. Both will occur simultaneously. Reset: Deeper, Faster and more insightful research.

4. The Rational Frame. Respondents forces to use rating scales to explain behaviour.This is not the way humans operate.Behavioural economics. Creative disruption.

The New Era of Expressive Research

Shift from Reflexive to Expressive Models. Expressive Research is Real Time Research.

Research and Subject

Reflexive Research

Circular relationship of cause and effect

confirmation

Networks

Expressive Research

Revealing the structure and mechanics of meaning

Research must reflect this shift.

Research models must get in sync and transform the structure of engagement

UNSYNCHEDReflexive Research

Consumer Savviness

Privacy Concerns

A Flood of Data

Patterned Data as

Narrative

Anonymous Clusters

Culture Driven

SYNCHEDExpressive Research

We seek to reveal the consumer journey… but marketing research methods focus on only part of this journey…

Trajectory

Storywhere

most research focuses

MOTIVES

What it is Missed

Existing hidden symbolic production of thoughts

Introspective Thoughts

• Lack of Meaningful Structured Samples keeps Cutting Edge Insights Elusive.

• Do you hear the reverb in the Data?• Are you amplifying Real Intelligence or human error?• We need to distinguish more clearly between Measures and

Insights.• We must cultivate Real Intelligence. • We need a new grammar for research. Reveal Real Patterns that

can act on in ways that have sustainable impact.

Data Myning

• If data is the new resource, expect consumers to start demanding their share of its value.

• Big data discussion. The value of customer data to businesses. Now, savvy consumers are reversing the flow: seeking to own and make the most of their lifestyle data, and turning to brands of proactively offer better solutions.

Bio-Physio – Biological Measures + Physiological measures

Lab-Mechanical Observation Neuroscience

Experience – Sampling Measures (ESM)

Self reporting of mental processes Validity-Correlation ESM Measureswith physio measures, psychological tests and Behavioural indices

Implicit / Indirect MeasuresFunctional Properties?Objective Property of the Measurement?- Not self assessment, but another behaviour. They can provide unique insight into effects of automatic processing on real-life behaviour. Mood, Emotion, Voice Prints,… Facial Electromyography

Self – Reporting MeasuresConcerns about validity of causal conclusions – systematic response distortions. Method variance, monomethod bias, and the psychometric properties(reliability +validity)of questionnaire scales. Context-design of studies, statistical analyses.

Polymeasures

Biological Measures (Biometrics) – Mechanical Observation

Facial Coding, Iris Recognition, Retinal Scan, Fingerprint recognition, Handwritten biometrics, Voice Prints, Voice Pitch Analysis, Utterance Analysis, Human-Computer Interaction, Vein Matching, Palm Vein Prints, Palm Dorsa, Scan Brain Waves, Pupillometrics, Heart Beat, Degree of Sweating, Eye-Tracking, Glasses Mounted Cameras, Telemetric Measures for Electrodermal Activity…..

Physiological Measures – Neuroscience (1)

(participant cannot /L about response)

BRAIN PHYSIOLOGY

• EEG – Electroencephalography – records electrical activity along the scalp. Voltage fluctuations from ionic current flows within neurons of the brain. Neural oscillations.

• PET – Positron emission tomography – 3D image/picture of functional processes in the body. Pairs of Gamma Rays. Tracer Concentration.

• MRI – or nuclear (NMRI) or MRT (tomography) – to visualise internal structures of the body. Uses magnetic field gradients that allow spatial information.

• fMRI – measures brain activity by deleting associated changes in blood flow. BOLD/contrasts (hemodynamic response).

• TMS – Depolarisation or hyperpolarisation in the neurons of the brain. Uses electromagnetic induction with weak electric currents. Using a rapidly changing magnetic field.

BRAIN

PHYSIOLOGY

Physiological Measures – Neuroscience (2)

tDCS – form of neurostimulation which uses constant, low current directly to areas of the brain via small electrodes. Can increase cognitive performance.

fNIRS – DOT (Functional Near – Infrared Spectroscopy) – (Diffuse Optimal Tomography). Advantages – used outside hospital environments, allows interaction with participants, less stringent in selection of participants. Disadvantages – only reaches cortex and not possible to quest all the brain simultaneously.

BRAIN

PHYSIOLOGy

New Measurements

Psychophysics – the scientific study of the relation between stimulus and sensation. The analysis of perceptual processes by studying the effect on a subject’s experience or behaviour of systematically varying the properties of the stimulus.

Complementarity

Physio data is unlikely to capture the full experience of emotion. Physio sata is continuous and can be measured without the conscious introspection required for self-reporting. The big problem with conscious introspection about emotion is that it is divorced from the way in which we tend to experience emotion in real life. Physio data is objective, can be reliably measured. Physio markers-oxytocin, tuf-alpha, heart rate variability, galvanic skin response, dopamine,…..

Key Mega Trends Transforming

Marketing Measurements

Speed of measurement

The near-real-time intelligence delivery is permeating nearly all facets of marketing measurements. Even if measurements are

not delivered instantaneously in a slick, colourful dashboard, the expectation of faster data and actionable insights is growing.

Speed is a competitive advantage.

Live Data

Real-time capabilities provide access to live data, empowering Marketing Developments with immediate feedback and reduced response time, which also create faster turnaround time.

Real-time capabilities reduce the speed to activate a survey and provide access to live data. Marketing Departments can be ready to update, change or correct their marketing activities instantly, if needed.

Live Data (Cont.)

A pre-programmed over-the –air marketing research system includes transmission facilities for delivery a regularly scheduled television signal and a low power, microwave transmitter or satellite transponder for delivering a special over-the-air television signal, including substitute programming, to cooperating households.

Marketing Research of Tomorrow

• The line between data and insight will diminish. • Integrated data mining tools that will integrate social media data,

sales data / retail data and with customer profile will be used in predictive models to forecast behaviours and trends.

• Impact of each data- based decision will be assessed through financial value, i.e., cost of analysis and revenue generated.

Using individual-level analysis, companies are increasingly able to customise their marketing activities through catalogues, direct advertising and promotions, online shopping sites, customer service, store-and customer-specific products and planograms, information kiosks and electronic shopping assistants.

To fully realise this level of integrated research, organisations will need powerful information systems to collect and analyse individual-level data, and the ability to customise marketing programmes.

Computational Performance Marketing Analysis

(the next generation of data analytics….) which integrates highly scientific, algorithmic approaches from various disciplines, including:

• Statistical machine learning• Combinatorial optimisation• Information retrieval• Experimental designs.

The Validity of Experimental Methods for Analysing Consumer Behaviour

The problems encountered when marketing research is conducted under actual market conditions are generally well recognised. These problems are primarily related to (1) the lengthy observation period that is often required, (2) the cost of data collection, (3) the many uncontrollable variables that can become unmanageable, and (4) mechanical and procedural problems of working with various marketing institutions. As a result ,the use of experimental-laboratory situations in marketing research will receive considerable emphasis in the future.

Intelligent DM-Techniques• Market Basket Analysis (MBA)• Memory-Based Reasoning (MBR)• Automatic Cluster Detection• Fuzzy Systems (FS)• Link Analysis• Decision Trees• Artificial Neural Networks (ANN)• Genetic Algorithms (GA)• Bayesian Data Mining• Indexing by Latent Semantic Analysis

Insights Management

With the geometric expansion of consumer data, the need for management of this data gas become intense. At a very granular level ,the simple fact is that corporations that choose to be data-driven have more consumer data streams (attitudinal, conversational and behavioural) than ever before. Making sense of these data streams and interpreting new data within the context of past research is incredibly important. There is a significant need for “insight management and archiving” and this market niche may prove to be highly lucrative to the firm that produces the most useable system and interface. While this need will be filled by software, it will ultimately require the spark of human cognition in order to pull macro level insights from numerous data streams. Without “insight management and archiving”, institutions will waste precious resources and time relearning that which has already been discovered.

The internet is making possible the collection of even more data. It is just a matter of time before we have “marketing-research agents” that scour the internet to do such activities as gather data from selected web sites at specified times, administer questionnaires (e.g. bizrate.com), or generate top-line reports on demand.

Without the deployment of “intelligent data-interpretation agents”, potentially useful information and insights will lay buried inside vast databases or will be lost in cyberspace. Already, intelligent agents on the Internet can help search for specific information (e.g., hotbot.com), compile information according to user preferences (e.g., pointcast.com), or correlate information in useful ways (e.g., Firefly, now part of Microsoft).

Portable Place-Based Research Tools

Behaviour in the real-world home or workplace setting is enormously difficult to simulate in a laboratory. The interaction of people with other people and objects in their spaces leads to unexpected behaviour that is difficult to anticipate with focus groups, surveys, and other standard product development and marketing inquiry methods.

House_n researchers have developed a collection of small, portable, and inexpensive wireless sensors, wearable devices, and associated algorithms and methodologies enabling MIT researchers and sponsors to conduct research in everyday places of living, work environments, and public spaces.

Portable Place-Based Research Tool Components

Wearable motion sensors. These small, comfortable, and low-cost accelerometer devices can be easily worn for days or weeks and used to collect data on what people are doing. MIT algorithms have been developed to automatically detect specific activities, such as walking, cleaning activities, moderate physical activity, and body posture. In combination with a mobile computing device (e.g. PDA or phone), the sensors can be used to detect specific activities of a person in real-time and provide or collect context-specific information.

Portable Place-Based Research Tools

House_n researchers have also developed a low-cost, portable kit of technologies and corresponding design methods for studying people in their own spaces. For example, researchers will be able to determine a subject’s baseline activity patterns in their own home, test the effect of prototyped innovation while the subject is living in the PlaceLab, and observe if the effect is sustained over time when a subject returns to his/her home.

Portable Place-Based Research Tool Components

• Context-aware experience sampling. Software has been developed for standard PocketPC to acquire data and context specific feedback from people. Sensors may be used to trigger questions and data capture. For instance, the computer can monitor heart rate and ask questions based upon variation in heart rate.

• Tape-on sensor kit. They can detect on-off, open-closed, and object movement events. MIT algorithms can then be used to study data and automatically detect certain activities in real-time. The tape-on sensors can be used with computing devices such as computers, phones, or PDAs to develop and test technology that automatically presents information based upon a person’s activities.

Portable Place-Based Research Tool Components cont.

• Image-based experience sampling. Computer vision algorithms can be used with portable cameras to collect video and audio data from an environment.

• Location beacons. Low-cost radio frequency devices permit approximate position detection within their environments and the tracking of movement over time.

Virtual reality simulations now allow marketers to test new marketing ideas quickly, inexpensively and confidentially.

It should be noted that, at the present time, virtual reality simulations are novel to consumers. Therefore, consumers initially spend more time navigating through the aisles and interacting with products in the simulation than they would in the conventional store. To overcome this limitation, researchers often ask consumers to go on multiple shopping trips through several product categories. By the third or fourth trip, most consumers are taking about the same time to shop as in the conventional store, and that is when shelf displays are manipulated.

New technologies will allow marketers to create even more realistic virtual shopping environments. For example, the CAVE is a room-sized virtual reality environment using four video projectors and stereo glasses to create a completely immersive, stereoscopic simulations for one or more consumers. The cost of this technology is currently in the hundreds of dollars, but is dropping quickly.

BIG DATA – MASSIVE DATA

•The trend today is towards more observations but even more so, to radically larger numbers of variables – voracious, automatic, systematic collection of hyper – informative detail about each observed instance.•Classical Methods are simply not designed to cope with this kind of explosive growth of dimensionality of the observation vector.•Some 1-8 trillion gigabytes of data were created worldwide in 2011 and the volume is doubling every two years.

Data Analysis•High – dimensional data analysis. Detection of types of patterns in the data. Intractability (Computational likelihoods). Underlying phenomenon.•Movement Data. Movement Behaviour. Movements of multiple discrete entities that change positions in space but preserve their integrity and identity.•Tick-by-tick data. Microarrays (to feed data in torrential streams).•Inverse regression. Sliced inverse regression. Dimension Reduction. Shows only the most important directions of the data.•Adaptive Regression Modelling – AdRM. Splines. Models of interactions and non-linearities.•Extensions of the K-Means Algorithm – Non-hierarchical Clustering.

Data Visualisation

• Perceptual Computing• Complex Data Analysis• Information Aesthetics• Visual Calltime and Information Design• Interface for browsing multi-facetted data structures• High-dimensional data

And……….. The Visual

•Visual Analytical Methods. Visual Data Mining.•Graphical Mapping and Interaction Techniques. Flow Maps.•Map-Reduce Programming Model.•Hyper-spectral Imagery.

Foresight

The next evolution in market research may be in the area of futures research or foresight in which futures markets and Delphi panels play a leading role. In addition, a greater appreciation for the impact of so-called “wild card events” (low probability, high impact events) on attitudes and behaviour, may push forward thinking organisations to create a futures/scenario planning function in order to move swiftly in an inflection point.

Futures Research – Methodologies + Methods

EXPERTS

Consensus Analysis

Casual Layered Analysis (CLA)

Cross Impact Analysis

Think Tanks

Online Group Discussions

Morphological Analysis

Brain Monitoring

Analytic Hierarchy Process (AHP)

Industry/Company X-Ray

Jury of Opinion

Backcasting

Scenario Planning/ Expert Systems/ Artificial Intelligence

Delphi

CONSUMERS/PSYCHOLOGICAL/CONTEXTUAL

Consumer Shadowing

Futures WheelIn-situ Research

Ethnographic Mimicry

Visioning/Preferred Futures

Living In Arrangements (LIA)

Ethnographic Episodes

Means-End-Chain (MEC)

Digital Ethnography

Mobile Information Technology (MIT)

Interpretative Simulations Sensory Ethnography

Digital Anthropology

Consumers’ Imprinted Research

Networked Narratives

Swarms or Collectives Research (Physics and Biology)

Are we entering the Golden Age of Marketing Research Technologies?

…..Oh, can anyone recognise the Marketing Research Industry anymore? Are they technologists or researchers? We have data collection automation, social, media listening platforms, facial coding technologies, panel platforms, routers, online communities, text analytics and sentiment analysis technologies, big data, dynamic dashboards, mobile research platforms, …… and the list does not end.

Research is Formalised Curiosity. It is poking and prying with a (Good) purpose.