The great thing about the ready availability of online market research today is the ease with which you can collect loads and loads of lovely data about your customers, am I right? Acres of bits of information about their likes, wants, needs, histories, motivations and values all detailed in digital qualitative and quantitative research output, agile and on-going. Throw your diligent social media listening output into the mix and what you’ve got is an awful lot of stuff.
But how to make sense of all that stuff? There’s the rub. Because unless you can analyse all of that data and then translate your findings into actionable insights for the business, you might as well have sat and watched back to back episodes of Gardeners’ World (not that there’s anything wrong with that)!
| Tweet This | |
| Unless you can analyse your digtal MRX data for actionable insight it's redundant |
To narrow the field, we’d better start with some terms of reference: in this article, I’ll be discussing the digital analysis of primary market research data, as opposed to for example, the performance analysis of digital marketing materials. Whilst both are key to business success, the skills sets required are unique. So what capabilities do insight departments need for the digital analysis of primary market research data?
One of the biggest issues with the analysis of online qualitative data is the sheer volume of data output that the researcher needs to be able to cut through and make sense of. It is vital therefore, that any insight department has the skills within to effectively code data, to filter it for themes and turn these into actionable insights. Powerful tools such as AtlasTI and NVivo allow for data coding and thematic discovery but the learning curve is steep, so a dedicated training course is a must as is digital aptitude.
If you simply want to present frequency of words/terms in an attractive way, it’s pretty easy to generate a Wordle (though you’ll need to do a bit of data tidying first) - most IT savvy insight professionals could master this in minutes. However, I would argue that Wordle’s are just a very high-level data summary rather than a digital analysis.
An alternative solution is to use an online research software with integral search and tagging tools. These tend to be aimed towards researchers, as opposed to digital data analysts specifically, and as such are easier to operate. Though they might provide more of a high-level analysis than AtlasTI and NVivo, they certainly go further than a Wordle summary and results are often real-time. You will still need a level of digital software understanding in your insight department to set these up but no more than would be required to conduct digital research fieldwork with the same platform.
Whilst Excel will see you a surprisingly long way with descriptive and basic statistical analysis, there are easier ways of doing a lot more with your digital quant data, not least SPSS. As a dedicated statistics software package used for logical batched and non-batched statistical analysis SPSS naturally requires a degree of bespoke know-how. It is however part of the curriculum in many Social Science and Management University courses, so the graduates in question are likely to have some level of familiarity with the digital SPSS analysis of quantitative data.
Another tool I am particularly fond of is Q (based on the R statistical analysis language). Q is designed specifically for market researchers and not only facilitates digital analysis but also allows for easy replication and templating of attractive charts and tables. If you have insight team members who have used SPSS’s custom tables add-on, they will walk though Q. Creating tables in Q is easier than in SPSS and the program also includes helpful features such as the display of significance in cross-tabs and so on.
Social media analytics tends to falls into two camps; monitoring and listening. Monitoring and listening are different and require different digital analyst skills. Monitoring is primarily concerned with the metrics surrounding social media, i.e. hits and engagements, lending itself more to the performance analysis of digital marketing materials. Social media listening however, is about looking at the sentiment contained within relevant social media comments in order to inform business strategy and for the purposes of this blog will be viewed primary market research data.
In some respects, social media listening can be like being in a busy, noisy pub with a large group; it’s easy to get over-awed by the number of conversations that you could be tuning into, not knowing what to concentrate on at any one point. There are digital tools which can support the listening process. They range in complexity but with a little time invested in training the right researcher would certainly be able to master the basics. The quality of social media listening insight is actually routed in the human ability to ask the right questions of the digital program analysing the data, and sense checking the outcomes, not their digital ability per se. There is still only so far that automated tools can go in correctly tagging and analysing sentiment in social posts. Further, it takes a very skilled insight professional to interpret the digital analysis for actionable business outcomes.
| Tweet This | |
| The quality of social media analytics insight lies in the ability to ask the right questions of the data |
Because of the need to manage huge sets of data, to query multiple data sources, and to crunch data beyond the scale that can easily be dealt with in mainstream GUI quantitative software, information handling and programming skills are big advantages in big data analysis. Those with experience of the R programming language - a particularly useful language for the statistical analysis of large data sets - are a desirable complement to any insight department responsible for big data analysis.
The tricky bit is the insight professional / big data digital analyst combo. You either need an individual who is both a programmer and insight professional - one who understands the needs of the business, the questions that need to be asked of the data, the insight goal, etc. - or a very close working relationship between your insight professional/s and your big data digital analyst.
With the right training, your insight professionals can be taught to operate many of the digital analysis programs discussed above. Such training is undoubtedly a worthwhile investment in terms of both analytical speed and data reach but it is important to note that any digital analysis program is only as good as the person using it. The most important skills required for a successful insight department are ultimately the same as in the pre-digital age: curiosity, an ability to get inside the minds of both customers and stakeholders, a thorough understanding of the business in order to ask the right questions of the data and the expertise to extract actionable insight and make useful recommendations.