Can AI revolutionize the quality of your data collection?

Artificial Intelligence is becoming common vernacular these days in general business as well as marketing research.  But, a quick Google search of “Artificial Intelligence and Market Research” reveals that AI has been primarily relegated to data analytics.  

That said, AI is already revolutionizing other business operations such as sales prospecting, where virtual assistants are helping human team members be far more effective than they ever could have been on their own.  For example, Conversica’s AI Sales Assistant can simultaneously hold thousands of email and text conversations with leads in a single day, with perfect follow-up and detailed performance tracking, freeing the humans to do what they do best- selling and closing deals. 

Moreover, consumers are adopting AI into their homes with increasing speed.  Think Amazon’s Echo, Google Home, and Samsung’s Bixby, to name a few.  According to Strategy Analytics, 7% of North American homes will own at least one by the end of this year, and by 2022, penetration is projected to rise to 33 percent of US households; (read more here).  Gartner says that by 2018, 30% of our interactions with technology will be through “conversations" with smart machines.

So what does this mean for marketing research?  It means that we are on the verge of a broad installed base of smart households that come with a bevvy of new opportunities to utilize this technology for improved data quality. At Diagnostic Measurement Group, we are already exploring the advantages smart voice assistant speakers provide in terms of qualitative and quantitative data collection.  

AI assistants are enabling consumers to provide both quantitative and qualitative feedback while engaged in the activity of focus.  Instead of consumers sitting in front of their computers to take surveys while recalling their past behaviors and feelings, they are engaging in research at the moment of truth.  It’s the difference between someone recalling how they felt about banal experiences like putting leftovers away into plastic containers versus walking their AI assistant through their process, finding the right sizes and shapes, describing the experience, and pointing out the challenges in situ.  And, AI assistants can ask and record quantitative, scalar questions and open-ended answers alike while consumers engage in the activity in question or complete assigned tasks.  

In partnership with Toronto-based Delvinia, we are currently experimenting with voice and conversational data collection in relation to projects fielded through the AskingAmericans data collection service. We are testing a number of potential ways to integrate voice user interfaces, like Alexa, into consumer surveys, including in-home product sampling programs.

The concept is remarkably simple:

  1. Panelists who have self-identified as an Amazon Alexa user, download Alexa Skill, and synchronize their Amazon and panel accounts.

  2. An alert is received when a survey is available.

  3. Respondents ask Alexa to open the survey and complete the questionnaire via voice.  Here’s a brief example video.

Combining this capability with DMG’s development of proprietary panels promises to enable companies to create their own AI-powered consumer panels for ongoing research that realizes huge efficiencies-of-scale in time and money with dramatically richer insights.  Best of all, with unit prices dropping into the $30 range, the smart voice assistant speaker can be an incentive in itself.

Few will dispute that AI is coming into all aspects of our lives, and the sooner Marketing Researchers embrace this technology, the better. So if you’re curious to explore more about how AI can improve the quality of your data collection, give us a call.

Want to save up to 70% on your survey research?


Getting pressure to do more research with less money?  What if a simple methodology could save your company between 50-70% on quantitative insights spending while maintaining customization for individual marketing research needs?  Well, good news, because such a methodology exists and has proven valuable for many forward-thinking companies...

Procurement departments and broader macroeconomic forces are forcing every company to scrutinize spending on marketing insights.  Marketing researchers generally agree, in confidence, that this pressure is limiting their ability to work with specialized and/or boutique suppliers who often bring a personal and hands-on expertise that the largest insights providers cannot.   The result is detrimental to the quality of the strategic guidance that marketing research departments provide. 

There is however a solution that addresses both sides of this challenge and, once properly introduced, will in many cases satisfy procurement’s savings goals and researchers’ required rigor. 

While customization options are endless, at the foundation, a proprietary panel of consumers can be cleverly designed to meet multiple objectives and panelists can be recruited to meet any required screening criteria.  This panel sample is recruited not just for one-time ad-hoc use, but for a specific time period and participation in a specified quantity of surveys.  Depending on the objectives, testing which would have required independent initiatives and spending appropriations can be accommodated through the panel, within a single budget.

In the panels we have designed and managed, we see greater than 90% panelist response rates, with costs reduced for straight survey research by up to 50% and in-home use tests reduced by as much as 70%, versus individual ad-hoc research studies. 

Of course, proprietary panels are not the panacea to end all ad-hoc research. Trade-offs are required and may include settling on testing among a more generalized sample specification, the realization that your respondents become less “naive” over time, and rigid execution on a more regular timeline, etc.  But, with a little forward planning, many of these trade-offs can be mollified with creative solutions.

two recent case studies: (client names have been blinded and specifications modified to maintain confidentiality)

1- Product Performance Panel- A “Warning System” for portfolio assessment

the challenge: A large multinational Food & Beverage client with a portfolio of four main brands and over 300 different SKUs in regular distribution knew that their years of cost-rationalization had degraded the consumer experience over time with many of their products.  But, the problem was how to efficiently and effectively determine which products were most in need of improvement.

the solution: A large panel of generalized target consumers (n=400) was recruited to participate in two different in-home product tests per month, for 12 months.  A total of over 800 products were tested, including competitors for performance benchmarking.  Through standardized reporting, the company was able to test their broad portfolio and identify those products most in need of improvement.  Best of all, the cost per product tested quantitatively was effectively shrunk to less than $5K.

2- Iterative Innovation Panel- Ensuring consumer alignment across the development process

the challenge: The cross-functional team of a Fortune 500 CPG firm, made up of Marketing, R&D (i.e., Product Development), Marketing Research, and Procurement, were fueling an innovation engine that was developing a constant flow of new product concepts, some with accompanying prototypes.  An efficient methodology was needed to quickly get consumer insights and feedback for iterative product development.

the solution: A panel of n=120 target consumers were recruited to participate in both qualitative telephone discussions and semi-monthly concept evaluations, for 3 months.  Because each survey was treated as custom, when available, prototype samples were sent to respondent’s homes for hands-on evaluation, along with the concept.  With quick-turn customized reporting, the innovation team was able to gain consumer feedback within days and received clear direction for product refinement, before final launch preparations, across a number of products.  Total costs were 50% less than what would have been spent if each test were conducted on an ad-hoc basis.

Ultimately, the design of proprietary panels is really only limited by one’s own creativity and can utilize tried and true traditional testing methodologies or the newest and most innovative data collection methodologies like AI (e.g., Amazon’s Alexa or Google Home), or even webcam-based emotion recognition, etc.

Want to learn how a proprietary panel could benefit you?  Give Diagnostic Measurement Group a call to learn more.

Logical Benchmarking

Logical Benchmarking

So you’re eagerly anticipating receiving scores from your latest marketing research study which you’ll use to guide a series of next step decisions to get your product to market.  But, have you thought about what benchmarks you’ll consider to help you UNDERSTAND whether you should “go” and move forward, or “no-go” and head back to the drawing board?

So, at the risk of stirring up an serious geek-debate among my fellow researchers...

To blind or not to blind?

To blind or not to blind?

At Diagnostic Measurement Group, we like to shout loudly, whenever the spirit moves us, “Product performance is king!”  Then, we just roll with the awkward silence and stares that follow from non-marketers.  It’s because we believe deeply that a brand can compel consumers to trial, but ultimately, it is the product experience that determines repeat purchase and supports a sustainable business.