Editor’s note: Neil Seeman is founder and CEO of the RIWI Corporation, an online data collection company in Toronto and senior fellow at Massey College in the University of Toronto.
At the time of writing, there are roughly 500,000 hits on Google to the search string “market research” and “disruptive.” There are hundreds of company-sponsored blog posts about supposedly disruptive trends in market research, notably the hackneyed Silicon Valley slogan SoLoMo (social local mobile), online communities, big data, neuroscience or text analytics. Depending on how they are implemented, all of these trends can be disruptive − or they can be a cacophony of buzzwords meaning nothing.
Data disruption in the Web sector consists of new-generation data
In the Web sector, data companies that have exploded in acclaim or user adoption tend to be in the business of category creation – that is, they have invented or proselytized a completely novel business model for data acquisition, data delivery, data encryption or visualization. We all inherited the best example of this phenomenon from Sir Tim Berners-Lee, who invented hypertext protocol (HTTP), the foundation of all online data communication among people, organizations and countries. Finding, transferring and manipulating new, transformative data where none existed before is the essence of data innovation on the Web.
By contrast, social media data and cookie-based data are what I call one click-away data: they are not proprietary, they are at your fingertips and they have a high signal-to-noise ratio. What people post online, for example, is the aggregated opinion of highly-opinionated crowds and is limited to open-access blogs or forums (notably Twitter) that enable scraping of content more easily than do the more anodyne corporate company page data sometimes amenable for access via Facebook. These limitations befall many big data collection tools. To mitigate the challenge, it is best to focus your data innovation-seeking efforts on identifying new-generation data.
What is new-generation data?
Airbnb is a flagship example. Thanks to Airbnb, a friend’s family found a spirited recent college grad who steered the family on a homey sailboat for a luxurious vacation near Cape Cod. People know Airbnb by its so-called unique selling proposition or USP: It enables hosts to rent out spare rooms or vacant homes to strangers and it recently surpassed 10 million visits last year. But it’s also in the data acquisition sector and in my view fits into what Michael Porter and James Heppelmann have called category-creating “external information sources” – such as energy price data or pollution data – that are intensely valuable in the new technology stack for other companies’ analytics services and product databases. The novel user-review data that Airbnb captures have been applied to measure America’s most hospitable cities in terms of host quality. The predictive factors that drive that index reveal dramatic differences in host quality across U.S. cities. Zillow is the data acquisition analog of Airbnb for home sales data.
Some new modalities of data collection capture data inside a new reservoir, using a platform that had not hitherto been discovered; SurveyMonkey or Qualtrics are two such companies. Compare these technology platforms to the proliferation of text analytics. In the latter example you may be harvesting more unstructured data prone to human bias or you may be focusing your efforts in new areas (e.g., since the client says trolling the Web for text is important) but you are not finding new voices. To be sure, text analytics, 90 percent of which is unstructured, may be high-value if you use detection tools to assess what text in what part of the world signals an emerging trend to which your client needs to pay attention. For example, the competitive advantage SAS enjoys with its text analysis tools is to elegantly organize and integrate the text-based data with other SAS tools – and to shed light on the automated discovery of emergent themes of value to the client.
Innovation in the market research sector is often one click-away data
At the few market research conferences I have attended, I found that what is defined as innovative by many attendees is a data feature long in existence, and there is a herd bias inherent to the industry. That is, what passes for innovation can be one click-away data. Let’s consider big data as an example.
Just using the term big data is meaningless. The Weather Channel has large buckets of data; so do all data companies these days. All you need to do is scrape open text on Google or Twitter on any subject of your imagination. You can then sort all that text into a free semantic reasoner database to generate automatic meta tags, text summaries and related news content (e.g., OpenCalais). With tools like these, you can use off-the-shelf natural language processing and machine learning to create an ontology of seemingly interconnected words and signals. Then you’ll be able to capture and visualize all that ”Web intelligence.” You can do all of this for zero dollars – apart from your time and diligence – and you can learn how to do it all on YouTube.
But what if that resultant taxonomy of data actually created a cognition engine? That is, the data, once funneled into your proprietary algorithmic system, automatically predicted how sub-populations in every part of the world would answer bespoke future questions? That is new-generation data; the resulting artificial intelligence (AI) engine serves as a technology system to generate ongoing insights based on external, new-generation data.
What’s telling about this example is that it creates more, not fewer, commercial opportunities for insights professionals to provide meaning to the AI engine. Too often people in the MR industry, I find, are frightened by data innovation since they think it will steal business away from them. To be sure, that’s possible; all service industries, from law to financial services to higher education, are undergoing fierce external competition. But the more the MR industry itself plays in the data innovation space – inventing the very innovations that will change the industry it knows better than outside forces – the more opportunities will be created for MR. Consider the following applications.
Finding new-generation data innovation in the market research industry
1. Assess any new data innovation to determine if the data are, in fact, one click-away data. If so, there are few if any barriers to entry and the only strong business tactic to leverage one click-away data is to spend piles of money on marketing or on swallowing up other firms that do it better. On the other hand, when a company comes along with new-generation data, you may want to pay attention and partner with it. If you don’t, your competitors will.
2. Avoid the quicker-cheaper-faster race to the basement trap whenever thinking about data innovation. The industry is hollowing its value by doing this. Meanwhile, big players outside your industry are stepping like blindfolded lions on your turf, offering what they call non-predicate-based analysis (aka, non-evidence-based pattern recognition swimming in researcher bias) and they are charging global customers a fortune for it.
3. Set up an internal ideation team that has two streams. In one stream, innovations are clearly aligned with articulated company strategy. In the other, random ideas can take flight and the team can be rewarded for extinguishing an idea that does not soar after a set period of time. Hella, the parts manufacturer to BMW, used that innovation cell approach. Hella awarded employees for new business models to improve productivity. Self-organizing teams would be terminated if their ideas did not translate into success and managers were rewarded for trying. As such, managers abandoned their egos and could easily stamp out their teams’ ideas.
4. Beware the expertise trap, which seems pervasive in market research. If you see yourself as an expert in everything, from cloud computing to neuroscience, then you’ve never met a real expert. This is especially true in online data trends. We are still in the Wild West phase of the Web – which is tremendously exciting but it also means that in order to innovate disruptively with data, you need to ask a lot of naive-sounding questions, such as: “Why do people in market research insist on calling people who complete surveys ‘completes’?” “People who answer surveys are real people, aren’t they?” Find a diversity of new voices from a new source and translate those new-generation voices into insights. When that happens, everybody wins, especially the client.