The Danger Of Big Data
A couple of months ago, I spoke at a conference in Las Vegas. Immediately before my talk, two advertising execs, one a professed quant geek and the other a “creative,” spoke about how their agencies rely less on hunches these days and more on quantitative data to drive emotional relevance between their clients and consumers. “We can identify human emotions in massive rivers of data,” the ad men said. When I pressed them for an example during the Q&A session, they described how they had recently mined millions of clickstreams, search queries, video views, website clicks, and the like for a major mortgage lender. All in, the technology investment behind their analysis must have stacked well into six figures. And their big emotional insight? When people start shopping around for a mortgage, that’s all they can focus on until they’ve gotten it all sorted out.
I could hardly believe my ears! Any skilled ethnographer could have discovered that same insight — and then some — through a day of in-home customer visits and $150 in taxi receipts.
Customer experience professionals can now glean customer insights from social media, financial systems, emails, surveys, call centers, and digital and analog sensors. It’s amazing and wonderful, yes. But here’s the danger: Companies that become mesmerized by big data put themselves at risk of spending enormous amounts of time and money amassing new data sources — and in the process, forgetting that research methods like observation and one-on-one interviews even exist. This has the potential to create a large, and exceedingly expensive, blind spot.
Don’t get me wrong. I’m not a big data hater. However, to create a complete picture of who your customers are and what they really need, you need a combination of quantitative and qualitative research methods.
Here are two effective ways to do it. As I alluded to a moment ago, you can gauge home buyers’ attitudes, emotions, and behaviors through qualitative research with just a handful of subjects — and then validate the statistical significance of those findings through quantitative analysis. Or, moving in the opposite direction, you can identify a problem or opportunity area in the home buying experience through quantitative research — and then uncover the underlying reasons for the problem (or the specific characteristics of the desired solution) through qualitative studies.
To see how this can play out in a little more detail, consider Allegiance’s Spotlight data mining tool. It churns through reams of customer feedback and operational data to identify which aspects of the customer experience correlate with key metrics like likelihood to recommend. For example, it could tell a retailer that customers who give scores of 9 or higher for store appearance and associate knowledge are 75% more likely to be promoters. Essentially, it shows the retailer where it should focus its efforts. Once that’s been determined, the retailer still needs to figure out what exactly it should change in order to improve those two drivers. To improve perceptions of store appearance, should the company work on cleanliness, lighting, signage, or something else entirely? And what are the most effective ways for associates to share their product knowledge? The answers to these types of questions can’t be found in any survey or database. Instead, the retailer would need to observe shopper and associate behavior in stores and intercept shoppers for post-purchase interviews.
The reason we’ve developed so many different types of research tools over the years is that they all have different strengths and weaknesses. By learning about and, more importantly, using qualitative research methods, you can ensure that big data doesn’t become a danger to your company.