In market research, it can occasionally feel like the rivalry between qualitative and quantitative research is like the Red Sox vs. the Yankees. You can’t root for both, and you can’t just “like” one. You’re very passionate about your preference. But in many cases, this can be problematic. For example, using a quantitative mindset or tactics in a qualitative study (or vice versa) can lead to inaccurate conclusions.
Below are some examples of this challenge—one that can happen throughout all phases of the research process:
Clients will occasionally request that market researchers use a particular methodology for an engagement. We always explore these requests further with our clients to ensure there isn’t a disconnect between the requested methodology and the problem the client is trying to solve.
For example, a bank* might say, “The latest results from our brand tracking study indicate that customers are extremely frustrated by our call center and we have no idea why. Let’s do a survey to find out.”
Because the bank has no hypotheses about the cause of the issue, moving forward with their survey request could lead to designing a tool with (a) too many open-ended questions and (b) questions/answer options that are no more than wild guesses at the root of the problem, which may or may not jibe with how consumers actually think and feel.
Instead, qualitative research could be used to provide a foundation of preliminary knowledge about a particular problem, population, and so forth. Ultimately, that knowledge can be used to help inform the design of a tool that would be useful.
For a product development study, a software company* asks to add an open-ended question to a survey: “What would make you more likely to use this software?” or “What do you wish the software could do that it can’t do now?”
Since most of us are not engineers or product designers, this question might be difficult for most respondents to answer. Open-ended questions like these are likely to yield a lot of not-so-helpful “I don’t know”-type responses, rather than specific enhancement suggestions.
Instead of squandering valuable real estate on a question not likely to yield helpful data, a qualitative approach could allow respondents to react to ideas at a more conceptual level, bounce ideas off of each other or a moderator, or take some time to reflect on their responses. Even if the customer is not a R&D expert, they may have a great idea that just needs a bit of coaxing via input and engagement with others.
Analysis and Reporting
In reviewing the findings from an online discussion board, a client at a restaurant chain* reviews the transcripts and states, “85% of participants responded negatively to our new item, so we need to remove it from our menu.”
Since findings from qualitative studies are not necessarily statistically significant, using the same techniques (e.g., descriptive statistics and frequencies) is not ideal as it implies a level of precision in the findings that is not necessarily accurate. Further, it would not be cost-effective to recruit and conduct qualitative research with a group large enough to be projectable onto the general population.
Rather than attempting to quantify the findings in strictly numerical terms, qualitative data should be thought of as more directional in terms of overall themes and observable patterns.
At CMB, we root for both teams. We believe both produce impactful insights, and that often means using a hybrid approach. We believe the most meaningful insights come from choosing the approach or approaches best suited to the problem our client is trying to solve. However, being a Boston-based company, we can’t say that we’re nearly this unbiased when it comes to the Red Sox versus the Stankees Yankees.
*Example (not actual)
Ashley is a Project Manager at CMB. She loves both qualitative and quantitative equally and is not knowledgeable enough about sports to make any sports-related analogies more sophisticated than the Red Sox vs. the Yankees.