AI and Automation – More Time for the Good Stuff.

AI and Automation.  These are terms that were first associated with big data, then with structured data, and now with small unstructured data (read “qualitative”).  The capabilities of AI and automation are filtering into most aspects of the qualitative process.  And the result is qualitative research that is often cheaper, faster, and – recognizing that this word is subjective – better.  But more importantly, it opens the analyst’s time for the good stuff – solving the client problem.

Here are a few ways our qualitative life is being changed by automation and AI.

• Sampling – sample providers are using API’s to integrate with qualitative platforms to enable fast, efficient, and cheap access to consumers for a variety of qualitative approaches: communities, IDI’s, online groups and emerging hybrid approaches.
• Data collection – Online platforms for qualitative and hybrid approaches  allow for quick turn qualitative research unimaginable a few years ago. Need to conduct 12 groups globally in a day? Done.  Need to test concepts with 1000 people and get qualitative feedback in 24 hours? Done.  Need to conduct “micro-communities” for a week on new positioning? Done.
• Moderation – the emergence of talent marketplaces now gives buyers the ability to become research DJ’s; to mix the right individuals with the right skillsets globally with the right technology to meet the business need in real time and all online. Qualitative research is entering the “Democratization Age”, where technology drives the human elements as much as it does the data collection process.
• Analysis – advances in AI, text analytics and data visualization tools now allow for the fast (within seconds, literally) categorization, clustering, thematic discovery, emotional analysis and even voice or facial analysis of responses as part of the tool kit. This makes the analysis of transcription, video, image and voice data as easy as running basic descriptive statistics in quantitative data.

It’s important to note that the effectiveness of all these tools are dependent on the right people, talking about the right things, in the right way.  Recruiting, study structure, and moderation are still critical elements.  AI and automation are still only as good as the thoughtful humans directing it. The advancement in these tools are drastically reducing the human time and effort it takes to gather and analyze qualitative information while improving the types of information and volume of information that can be analyzed.

This leaves the analyst with more information to work with and more time to focus on the implications of the insights to address the business issue.  Isn’t that what we really need – more time for the good stuff?

Writing a high-quality screener is critical to getting the right respondents for your qualitative research project. Since screeners are actually short questionnaires, why not apply the same skills and techniques to developing your screeners? On May 12, 2017 L&E hosted a webinar with David F. Harris, author of, The Complete Guide to Writing Questionnaires: How to Get Better Information for Better Decisions. David provided detailed examples of how to improve your screeners to get better respondents for your qualitative research.

Webinar Presentation

To download a PDF copy of David’s presentation from the webinar, click here.

Webinar Recording

About David F. Harris

David F. Harris conducts qualitative and quantitative research for companies in a variety of industries. As founder of Insight & Measurement, he also conducts training and consulting on questionnaire design and screener development. He is author of, The Complete Guide to Writing Questionnaires: How to Get Better Information for Better Decisions. He received his B.A. from Reed College, in Portland, Oregon, and his M.A. in Quantitative Psychology from the L. L. Thurstone Psychometric Laboratory at the University of North Carolina at Chapel Hill.

Prefrontal Cortex, Limbic System, System 1, Behavioral Economics, Emotional Quotient – all phrases that didn’t matter in the market research industry until recently (the definition of recently being up for debate). The point is, market researchers have learned that emotions play a critical role in consumer behavior and satisfaction.

Knowing that emotions play a role and knowing which emotions play what role are two very different things. While no one has all the answers to this problem (and all the answers may not even exist), here are some issues to consider when you take on the task of understanding emotions in a business context.

• Segmentation: It’s not news that people are different from one another. But particularly with emotional measurement, the emotions that people feel about a category, product, or brand are often related to their degree of engagement. For example, heavy users of Apple have a different and deeper emotional connection to the products and brand than moderate users of Apple. In addition to engagement, the emotional nature of the individual person can play a role in the connection to the brand. There are several ‘personality traits’ type models that can help structure the understanding that comes through qualitative research.
• Context: When delving into the emotions elicited by a category, product, or brand – context can be incredibly important depending on the nature of the product and the situations in which the product is used. Some products are personal by nature in that there is not much social interaction related to the product. Cleaning supplies are a reasonable example in that they are generally used in a private setting (Saturday morning around the house) and not part of a large social effort. In contrast, birthday cakes are almost always used in a social setting – be it large or small. The emotions elicited by either can be generally viewed as the emotions in any situation where these products are used. But let’s take the example of coffee. Coffee is sometimes used in personal settings and sometimes in social settings. The emotions elicited in each can be very different from each other with the exact same product.

• Conscious: Many of the System 1 approaches would have you believe that all emotions are nonconscious or the decisions made from these emotions are all nonconscious. While this may be true for some people, most can articulate real emotions with a modest degree of depth and accuracy. Happy vs. Sad, Scared vs. Comforted, etc. In situations where nuance is not warranted, in depth interviews (and a good tool kit from a moderator) may be sufficient to uncover the emotions and emotional drivers for a project.
• Nonconscious – Implicit vs. Biometric Measures: In those cases where nuance is required or the respondent may be either unwilling or unable to understand or articulate, implicit association and biometric feedback are the two categories of tools that offer the most insight. A distinction is made here because definitions of these two words are wavering – Implicit is any tool that delves into nonconscious emotions but does not measure some component of the body in doing so. For (at least our definition of) implicit, the tool kit is generally an implicit association test or metaphor elicitation. There are several good tools and techniques that fall in or near these definitions. Biometric feedback has made important advance in quality and cost over the past few years and is therefore being used by more and more clients. The most common of these tools are eye-tracking and facial coding. Others that are reasonably available for qualitative research include Galvanic Skin Response, EEG, fMRI, and heart rate monitoring. While these are still early in their evolution and there is much to learn, these tools can help researchers understand emotions and emotional triggers.

Emotional measurement is difficult. However, with guidance and tools, it is easier than it ever has been, and these tools make a useful addition to the qualitative researcher’s toolkit to help deliver deeper insights that deliver greater business value.

The market research industry is approaching an alarming crossroads. On one hand, DIY tools like SurveyMonkey and Google Consumer Surveys are taking off, making it possible for anyone to create their own survey for free. Research firms are left vying for work on more complex studies that require their expertise and scale, and client-side researchers are equally challenged, as budgets dry up and internal clients expect insights that are faster, cheaper and actionable.

At the same time, there is also a growing consensus that consumers cannot always accurately tell market researchers what we want to know; that consumers are highly emotional and make decisions for reasons that are often not accessible to their consciousness (and, by extension, not discernible through classic market research techniques).

Amidst this disruption, the utilization of nonconscious measurement (also called “neuromarketing” or “applied neuroscience”) for commercial research continues to be a growing trend for pursuing consumer insights.

To continue reading, get our latest eBook by clicking here (no form to fill out), Interpreting Nonconscious Data, written by Leonard Murphy, Editor-in-Chief of GreenBook.

Qualitative research is a set of methodologies whose common goal is to explore and uncover the hows and whys of consumer behavior. It’s left to quantitative research to measure the “how much”. The categories of qualitative and quantitative have been easy monikers to differentiate the concepts of exploration compared to confirmation, depth of insight compared to breadth of insight, the immeasurable contrasted to the measurable. 

But the lines between qualitative and quantitative are blurring rapidly. While there are numerous reasons for this, two of the main forces are business demands (speed, cost and acting on useful information) and the capabilities afforded by technology.

Business Demands

When corporate researchers were asked about their pain points for the 2016 Quirk’s Corporate Researcher Report – too many projects for staff or budget topped the list as always or often a pain point (60%, 50% respectively). When asked about challenges – corporate researchers noted the inability to for insights to drive action and not being able to complete projects fast enough as always or often a challenge (50%, 42% respectively).

When looking at the newer methods of research, corporate researchers view qualitative tools (as a whole) as more effective at providing actionable insights. “Online qualitative/focus groups” are viewed by 70% of corporate researchers as effective or very effective. Mobile-specific surveys are second (53%). But the next couple are also qualitative techniques – social media research (40%) and text analytics (52%).

Taking the premise that actionable insights are the number one goal for corporate researchers and the number one challenge, and that qualitative research (at least among the newer methods) generally does a better job of providing these insights, then it is likely that more energy will be directed here. At the same time, there is a methodological drag to qualitative research according to this survey. The top four drivers of methodological choice for a project are; representative sample (82% viewed as very or extremely important), proven methodology (69%), speed (58%), and response rate (52%).

While qualitative has never been about representation per se, the scale in some tools is pushing that of quantitative. Some of the newer qualitative tools do have a good way to go before they are as deemed worthy of a wide variety of use cases – but that is changing rapidly.

Technological Capabilities

Technology is changing the nature of qualitative in several different ways:

• Scale:  We currently have some qualitative information at the scale of quantitative – and a growing ability to analyze that information. The easiest examples are both based in text analytics – customer service information and social media analytics are able to provide some structure to the vast amounts of unstructured data available in their respective data sources. While there are limitations; sample management, response bias, etc., those only limit some applications. The variety of proven use cases continues to grow, expanding into consumer needs, brand management, operations, and more.
• Geography:  As the tools a respondent needs to participate in qualitative research sit in the purse, pocket, or on the kitchen table, geography is much less of an issue than it was before. Many of the corporate researchers that are using newer methods, such as online qualitative, mobile ethnography, etc. are doing so to take advantage of the access to consumers in the appropriate environment. Mobile, in particular, has the benefit of providing “in the moment” access in the way that was almost cost prohibitive using other tools – particularly low incidence or sensitive topics.
• Cost: The “somewhat” lower cost of qualitative methods is also seen as a reason for switching.  Several researchers mentioned the cost of online qualitative and communities as reasons for moving away from more traditional qualitative methods.  This was generally expressed in terms of travel cost, but occasionally as overall project cost. Though it was not mentioned, social media analytics can be very cost effective depending on the type and frequency of projects.
• Depth of Insight: Deeper information is available through communities than through other traditional methods. This is a driver of choice for a few researchers that noted the ongoing nature of a community allows a more longitudinal look at consumers and their product issues.
• Speed: Speed is one of the top two corporate researcher challenges. While technology has improved speed to insight in almost all the tools mentioned to this point – it is hit and miss.  Once a social media program or community is up and running, the speed to insight is generally improved over traditional methods.  But according to one researcher, online qualitative groups are “not a lot faster”.  However, the future holds a different story.

The Future of Qualitative Research

Recent advancements signal the future of qualitative research. Text analytics improve month by month. Several companies have had different levels of success in identifying emotional context, purchase probabilities, consumer values, and a number of other frameworks that build on our understanding of consumers and their relationship to our products. Other qualitative methodologies have automated processes for understanding the emotional connections consumers develop with products and brands through imagery, metaphors, and other projective techniques. Crowd-sourcing has been in use for several years, but new applications and technologies are able to provide deep understanding of an issue in a cost-effective and timely manner – and at the scale of quantitative research. Artificial intelligence is making its way into marketing research. These tools have been around for a couple of years, but have improved dramatically in the past year. Many of these tools measure delivery in hours or days instead of weeks or months.

In addition, we should start to see the synthesis of mobile sensors, shopping or purchase behavior, and advertising effectiveness over the next year. This puts the depth of qualitative insights in context with behaviors – which is the holy grail of market research.

The future of qualitative research addresses the key pain points of today’s corporate researcher. The lower cost provided by technology and improved user interface will reduce the impact of constrained resources. Technology is also quickly and significantly improving the time to insights. Experience will prove the application to specific use cases. Scale will reduce the concern about sample representation. And most importantly, the quality of insights provided by new qualitative methods has and will continue to improve  Many of the newer methods are combining the depth of qualitative with the breadth of quantitative – getting us to the how, the why, AND the “how much”.

This article first appeared in the 2016 Quirk’s Corporate Researcher Report.