Article
Big Data: a break with the past … or not!
July 2, 2018
Transformations underway in retail and the multiplicity of data sources represent a challenge for traditional market research methods and the role of experts.
Take the example of market potential. To evaluate the potential turnover of a point of sale in a given location, the standard approach consists of drawing a line around the POS, representing the catchment zone, where most of the customer base is located, and then estimating the spending of customers at the POS, taking into account multiple criteria, such as local and online competitors, and the attraction of the retailer.
This empirical method involves a whole series of hypotheses and requires in-depth knowledge of the local context and the factors influencing consumption.
It takes a retail expert to perform such an analysis.
But it has become increasingly difficult for experts to decipher the behaviour of consumers who buy everything everywhere at any hour, in stores and online, where they live, work and play, and when they’re on the go.
It would be smarter to take a reverse approach: rather than estimating the probability that a person will go to a point of sale and make occasional or regular purchases of varying amounts, why not start by measuring observable flows in proximity to the POS and try to determine the market potential from there?
The data is exploited to analyse movements in relation to any point on a map: a shopping centre, a store (belonging to the retailer or a competitor), a selection of potential sites for your next store opening or investment, etc.
The possibilities are endless, but let’s come back to our example of potential turnover at a point of sale.
What does the raw data tell us? Nothing more than what we would glean from a map showing the to-and-fro of electronic chips.
Such analyses still require a retail expert.
Going from individual movements to flows, for example, means identifying the traits certain movements have in common, in terms of origin, destination, itinerary, frequency, time of day, etc., and assigning significance to this collection of traits. Commuter movements are one such type of flow, but there are many others, including flows passing near a competing POS. The important thing is to pick out the flows that are pertinent to a given project, which won’t be the same if you are looking to open a cocktail bar or a co-working space.
The next step consists of qualifying the flows, i.e. studying their importance for the project. This is the classic work of correlating a home address with a certain consumer profile (socio-professional category and income, for now) and the retail concept that is most likely to match.
In the future, interoperability between the Calibrate database and that of a retailer’s customers will open up incredible new avenues for analysis … creating new challenges for experts who need to plunge into the data without getting drowned!
Technology is revolutionising the work of retail experts but can’t replace their expertise. On the contrary, the masses of data available, even when using a powerful processing tool, demand an ever greater capacity to interpret them in order to inform the strategic choices of retailers and investors.
Insights in Your Inbox
Sign up for our Newsletter