Create Process/Tool for Analyzing Virtual Focus Group Data for a Global Consulting Firm

Project Description

We are a global consulting firm delivering $38B annual revenue with the world's largest companies. Our Rapid Customer Insights team wants your help to create new tools to analyze consumer research data. Specifically, our team conducts virtual focus group sessions for our clients where 100-500+ research participants are recruited for a 60 minute virtual discussion. The goal of the session is to better understand consumer feelings and possible actions regarding a product, market, or service.  Each session generates at least 50,000 data points which makes these difficult and time consuming to analyze and build into good storytelling slide decks for clients. 

Project Objectives

The objective of this project is to create a template, methodology, or tool (or all of the above) that can be used to better analyze future sessions in a more automated and comprehensive fashion. See attached for sample data. Our aim is to make use of “plug and play” methods for conducting standard analyses in a repeatable way. Ultimately we need to shape the analyses into a story for the client, and so we need methods and tools to more quickly and easily create consumer “personas”, identify highly divergent or convergent answers, evaluate and classify the text, and quickly spot correlations between responses that are not immediately obvious.   

As part of your proposal please answer the following questions:

• Which advanced analytics approaches could be applied to the raw data in a repeatable way to derive interesting insights that can’t already be produced with the tools we have?
• What trade-offs must be considered when choosing one algorithm, method, or tool over the other?
• How might we better handle and analyze free text fields via Natural Language Processing to understand quality, tone, and articulation of consumer responses? 
• What is your approach for turning your recommended data discovery process (using sample data) into a repeatable process for future projects? Please assume the raw data provided here will be formatted the same way for future projects. Assuming this, how could you standardize and automate the discovery phase?  
• Which technologies or methods, including third party vendor tools, could we use or build to create a repeatable process for analyzing this data?
• Using the data provided, what interesting insights can you find in these sample data sets?  
• Optional: please perform an initial analysis of the dataset we provided making use of your recommended methods or tools. Detailed submissions with proof of concept reports included will be more heavily weighted in the proposal evaluation process

Data Available for Analysis

The data from our virtual focus group sessions come in two categories: poll based responses (participants select one or more options) and open ended responses (participants respond with free form text). Participants then are able to “up vote” answers they like or agree with, which generates two additional data points per question: Popularity scores, how often the answer appeared from 0 (never)-100 (always), and Consensus scores, to what degree people agree with the statement from 0 (not at all) to 100 (complete agreement). Other data captured in each session are demographic e.g. age, location, household income, buying preferences, product use preferences, and other screening questions. We also would like to analyze the individual answers themselves using natural language processing methods to determine relative sentiment, degree of articulation, and to identify keywords. Please see attached for sample datasets for evaluation.

Project Structure

Phase 1 – Open call for proposals - Two participants will be selected to move on to Phase 2 and will receive $150 (1st) and $100 (2nd) incentive prizes. Please see attached for detailed project requirements.

Phase 2 – Proof of Concept – Each proposal will be build out / developed into a functional prototype, template, or method with displayed analytical results. Each participant will receive $1000 for their efforts, however only one will move to Phase 3 to refine their work. (Timeframe 5-7 days)

Phase 3 – Refinement – Winning participant will be given the opportunity to refine and enhance their concept, earning up to $3,000 in incentive prizes.