In recent years, the market-research industry has found it increasingly difficult to find truly new customer insights to inform breakthrough innovations. Additionally, in today’s uncertain times, researchers are faced with increasing pressure to cut budgets. Forced to do more with less, many are challenged to find game-changing insights more efficiently and cost-effectively.MIT researchers Artem Timoshenko and Professor John Hauser have developed a machine-learning methodology that addresses these important industry challenges. Working with Applied Marketing Science (AMS), Timoshenko and Hauser conducted an extensive experiment comparing the output of their machine-learning model to the output of a traditional Voice of the Customer (VOC) research study. Initial experiments demonstrated that, when sufficient user-generated content (UGC) exists, machine learning can produce a complete database of customer needs comparable to the database of customer needs identified during a traditional VOC study. The machine learning method also provides several key advantages over traditional research including:
- Virtually eliminating human bias
- Cutting research costs
- Expediting research timelines
- Gathering important insights less likely to surface with traditional methods
In recent months, AMS has conducted further machine-learning experiments to explore whether the machine-learning methodology can provide benefits in quantitative research as well. Traditionally, a prioritization survey is the best way to truly assess importance and satisfaction as they relate to customer needs. However, machine learning presents an opportunity to test another method of evaluating importance and performance of needs, which may lead to a greater data dispersion, or variation among scores. This allows researchers to more easily identify the most important unmet needs. To test this, we designed an experiment using the coherence and richness of UGC to determine whether we could overcome the challenge of clustered importance/performance scores.
Learn more about our latest experiment and the findings in our recently published whitepaper, “Unearthing the Unexpected: The Power of Machine Learning for Market Research,” written by AMS Principal Carmel Dibner. The whitepaper discusses the machine-learning methodology, including applications for both qualitative and quantitative research, and presents our researchers’ latest conclusions. For any additional questions, contact Carmel Dibner at firstname.lastname@example.org or (781) 250-6325.
Click the button below to access the recently published whitepaper: