Qualitative research insights powered by AI: Matches human analysis

About Our Customer

Headquartered in the USA, our client is one of the world’s most respected FMCG companies with several successful billion dollar brands as part of their product portfolio. Over the years, they have conducted thousands of hours of in-depth consumer research to understand their customers better, the results of which are now an indispensable part of product engineering and marketing. Consumer insights gleaned from these research studies help create a differentiated product line in a highly competitive market place that helps sustain growth and increases market share.

The Challenge

The CPG market is overcrowded with thousands of brands fighting for attention and wallet share. In a tough, highly segmented market, large FMCG companies must continuously study consumers to understand the intricacies and the co-relation between product consumption and consumer behavior patterns. The most effective consumer research studies are observational in nature. Video is becoming the medium of choice to record and analyze consumer behavior. Product design changes are incorporated post-analysis of consumer behavior by R&D teams.

Consumer behavior research conducted by our customer thus far, has followed the traditional established methodology:

  • Conduct focus groups with a sample size of consumers
  • Deploy surveys and questionnaires with close and open-ended questions
  • Study hours of video footage on behavior patterns

But this approach to consumer research is fraught with challenges.

Challenge 1:  Time consuming. Manual study and analysis of recorded videos take 2x-3x of the recorded video duration, to review, analyse and generate insights for behavior and deliver the final report. Moreover, insights generated from such studies may not be easily searchable or referenced for future analysis.

Challenge 2: Manpower intensive. A large team of analysts is required to study hours of video footage. The process of review is tedious and vulnerable to human error, which could affect the quality of the analysis.

Challenge 3:  Inconsistency of analysis. Analysis of video footage studying video footage of consumer behavior and insights thus generated are vulnerable to human perception.
This results in inconsistency of insights generated, negatively impacting ROI.

Challenge 4: Timely decision-making. Since the conventional methods are time and effort-intensive, product engineering teams are frequently handicapped in taking timely decisions, which might have a negative impact on marketing and therefore, revenue growth.

Challenge 5: Resistance to technology driven analysis. Given that research into consumer behavior has primarily been the domain of human experts, accepting a technology solution to improve analysis efficiency is not easy to accomplish.

The task To overcome the challenges outlined above, we had to develop technology that achieved the following business objectives:

  • Shrinking the time taken for video analysis
  • Matching the accuracy of human analysis and eventually surpassing the same
  • Significantly reducing costs with faster time of completion and minimizing errors

The Solution

Our customer enlisted Streamingo to develop a technology solution to help them with speedy analysis of video-immersive research at scale.  FizzStream,our flagship software platform was put to work – FizzStream is a powerful, sophisticated and innovative video insights platform with inbuilt AI research technology – to deliver to expectations.

FizzStream delivered tremendous business value by:

  • Analyzing hundreds of hours of video footage simultaneously
  • Automated video tagging of objects, actions and human object interactions
  • Bringing in consistency of behavior analysis at scale with reduced business costs
  • Generation of insights from video analysis to help product engineering and analytics teams answer business questions

The Benefits

FizzStream demonstrated that technology-driven analysis of consumer research was not only capable of performing the task, but delivering value that surpasses human effort. By deploying FizzStream, our client found that they were able to achieve:

  • Speed and agility – by using technology-driven computation power, a larger number of videos could be analyzed as the time taken for analysis was reduced, leading to quicker decision-making by product engineering teams.
  • Significantly greater accuracy – with the inbuilt AI-technology, FizzStream was able to easily video tag interactions, increasing accuracy and reducing the potential for human error.
  • ROI at scale – by conducting simultaneous studies and the reusability of prior analysis helped our client derive significantly higher ROI.

Savings – by cutting down on human manpower required to study and analyze hours of  video footage, our client was able to achieve considerable cost savings–