Counting qualitative research.
Image via: Dan Diemer
One of the biggest changes in the way I approach strategy that has taken place over the past 10 years has been the shift from relying primarily upon qualitative research to an almost exclusive reliance, today, upon quantitative research. Some of this may have to do with just simply getting better at my job, but I think a lot of it has to do with how technology has enabled formerly “soft” measures to be aggregated, counted and ranked.
A great example of this is Noah’s excellent Brandtags project. Prior to this, the default way to get at all of the different perceptions and ideas that surrounds a brand would have been to run a bunch of focus groups. But, where qualitative research uses small sample sizes and depth interviewing techniques to gather all the various meanings around a brand, Brantags uses scale and first response to deliver almost the same information.
There is, of course, one big difference. Qualitative research purists would say that the texture, nuance and context of these idea aren’t delivered through a format like Brandtags. Instead, Brandtags delivers a completely different benefit and that is the ability to be counted and ranked. Brandtags, and other things like it, make qualitative data digital, and, IMO, make it infinitely more powerful and useful in the process.
For example, both types of research could tell me that two brands have the same idea of value associated with them. However the data behind Brandtags will allow me to see which brand has the strongest association with value both in terms of raw numbers and also in relation to other ideas. Because I have numerical values associated with all of this information I can set up fairly accurate tracking to measure any changes and because these are numbers, I can mix them with a whole set of other numbers including sales and traffic data and so on.
What I lose in nuance, I am more than able to make up through being able to build much more expansive pictures of what’s really going on through being able to see the hidden relationships between different measures.
How does this fit with your experience?
Possibly related posts
Tags: planning qualitative, quantitative, research
taulpaul Says:
July 30th, 2009 at 7:10 pmWhile I rely heavy on quantitative data to make online marketing decisions, I have become more reliant on my colleague Mark Wagner's qualitative and observational data, over the past year. Bringing all three data sets together has allowed for interesting discussion, and allowed to point out irregularities that can be looked into further. This has allowed us catch several missed opportunities.
curiouslypersistent Says:
July 31st, 2009 at 1:07 pmPerhaps Brand Tags isn't the best example. It might be useful in informing “what” people are thinking, but it isn't (yet) able to give the most important piece of information – “who” is saying these things.
And of course, for the nuance, the “how” and “why” are pretty important.
I definitely agree though that other online opinion aggregators can offer real value when planning.
searchanddestroy Says:
July 31st, 2009 at 2:39 pmmore to the point about brand tags and “who” is saying what – brandtags awareness among advertising types 100%, awareness among non-advertising types 0%. that obviously is not scientifiic, but i cant find a single person outside of my work circle that (a) knows about brandtags and (b) would spend time adding associations.
i do, however, see a tremendous amount of value in the methodology and its ability to leverage scale. perhaps there is a better, more natural way to pull this data from the innernets.
adrianho Says:
July 31st, 2009 at 2:54 pmAdvertising types are probably going to be more sensitive about brand perceptions than non advertising types (although that's also not necessarily true these days anymore) However, not all qual is about getting insight from one specific group. Often what you're trying to do is to capture the spread of all possible reactions or feelings about a certain subject and gauge which are stronger/weaker etc. For this I'm not sure it matters whether I know who's answering or not, it really matters more that I have enough people answering to get the spread.
Gil Says:
July 31st, 2009 at 3:22 pmA great example of this is the data we get from tools such as Radian6, the various Google analytics tools etc. When used to analyze the posts, comments and linking habits across blogs, twitter, online communities etc. we can take away quantitative learning about scale and breadth as well as the more nuanced sentiments that come from qualitative research.
Matt Daniels Says:
August 1st, 2009 at 5:49 amThe problem with Quant is that one's answer is affected by the wording, order, and tone of questions. I'd agree that brandtags gets around this, but for a typical survey, every person is wooed by incomprehensible things that irrationally change a person's choices.
This surely happens in Qual as well, but I'd like to think its easier to diagnose the “big picture” of what someone's thinking when you looking right at them.
tomneveril Says:
August 7th, 2009 at 1:02 amThe problem with Brand Tags and most attitudinal studies: they provide precise metrics about conscious thought. But in reality, it's the SUBconscious that drives most behavior –including purchases. So, I'd like to know about new quant tools that can ethically track audience behavior; including marketing interactions, shopping, & brand usage behaviors outside of the web. (UPC scan data is an old but good example of tracking purchase behavior.) Tracking people via GPS/cell phones would be nice –but it's way too big brother-ish. In any case, we need more tools that reveal experiences and actions, not more ways to ask questions.
hartw Says:
September 4th, 2009 at 1:57 pmZJ, isn't brandtags mostly the associations of current and would-be marketers who come to the site via blogs like yours? Is there something more to it? For example, more systematic sampling, real quant measures (not just varied type sizes), and equivalence of people rating different brands?