The future of consumer (actual) behaviour tracking

The future of consumer (actual) behaviour tracking

If you follow any of the DigitalMR (DMR) content online, you will know that we often speak about the 3 main avenues to acquiring consumer insights:

1) Asking questions

2) “Listening” to unsolicited feedback online

3) Tracking behaviour

We also acknowledge the existence of some other, more niche sources of information, such as observation of what customers do and measuring biometrics.

This blog post is the 3rd of a series (see the previous two posted here and here), and it is about the use of behavioural data integrated with surveys and social listening, to synthesise ABSOLUTELY UNIQUE INSIGHTS! We call this the Triple S Integration = Surveys + Social + Sales. In this equation, ’Sales’ is a proxy for any kind of consumer behaviour that can be recorded and tracked.

Tracking behaviour for market research is substantially trickier than asking questions or social listening; the main reason being personal privacy! With the European Union’s GDPR (General Data Protection Regulation) officially enforced in May, this becomes an even more acute issue.

Asking questions and social listening are less of a problem. For the former we ask permission (for example upon inviting a person to join an online community), and the latter is based on information published on public websites, which comes with the possibility or even the expectation to be quoted. Having said that, some market research associations still ask their members to anonymise social media posts from consumers before reporting them to clients. We have said it on multiple occasions during the last 5 years and we will continue saying it: believe it or not, anonymising posts that were harvested from public websites violates copyright laws.

Here are some types of behaviour tracking that we can discuss individually:

  • Purchases: through retailer or brand loyalty cards, the card issuer can access information on purchases at the individual level
  • Sales: retail sales (at an aggregate level) can be known through the Nielsen/IRI reports for FMCG, and the GfK/NPD reports for Durables
  • Website visits: can be tracked at the individual level with cookies, if you are the owner of a website that is, or at an aggregate level through a panel, such as that of comScore
  • Transactional data: if you are a bank or a telecoms provider you know a lot about the transactions of your own clients
  • Search data: from Google or Bing or other search engines; some people – like the author of Everybody Lies – believe that Search Data is the only dataset we can and should believe in order to understand and predict consumer behaviour
  • App usage: on smartphones and tablets, available to the owner of the App and the operating system owner such as Google, Apple and Microsoft.

There are of course more types of behavioural data, such as voting at political elections and theatre or restaurant or other HORECA establishment visits etc., but the above are the most popular ones already in use.

Based on our experience, the types of behavioural data that can easily be integrated with surveys or online focus groups and social listening, without the need of multiple permissions and negotiations, are:

Retail sales and Google searches - at an aggregate level, as well as online visitor behaviour on a brand’s own website - at an individual level, through cookies. Getting access to transactional data from banks and telecoms companies, or purchases using loyalty cards from retailers, is quite difficult, so the only ones that can realistically use this data are indeed the owners of the data.

DigitalMR has conducted extensive R&D on data source integration with excellent results. Some people in our industry call this type of continuous integration “The Holy Grail”. The implied connotation is that useful data integration cannot be found, or it may even not exist. Well we beg to differ; we have already proven that sentiment has 81% positive correlation to retail sales and that certain KPIs measured in a brand health tracker can be replaced by high correlating KPIs from social media listening. We are open to share our research to relevant interested parties, just ping me @DigitalMR_CEO.

Share this article: