Fnac draws on startup knowhow for predictive Big Data approach

By April 07, 2015

French consumer electronics and entertainment products retail chain Fnac set out to make better use of its customer data with a view to targeting its customers more accurately in its promotional campaigns and so optimising sales. The retailer has been working with a tech startup on a predictive CRM platform.

Startups are playing an ever-greater role in the exploitation of Big Data, as Mathias Herberts pointed out in a recent interview with L’Atelier. Nowadays large firms need to collaborate with tech startups in order to increase their performance and profitability, argues the Cityzen Data founder. The Fnac Group has certainly grasped this, judging by the work it has been doing recently with Paris-based tech startup tinyclues, which was showcased at the Big Data Paris event on 10-11 March. Tinyclues specialises in building SaaS (software as a service) predictive Customer Relationship Management (CRM) platforms for the use of a client firm’s marketing and CRM teams. Fnac has over 20 million customers listed in its database, generating vast amounts of specific data, an opportunity for the retail chain to optimise the targeting of its promotional campaigns by feeding customer data through a predictive CRM approach.

Probability modelling

The methodology developed by Tinyclues, which is based on machine learning algorithms, enables far more accurate customer targeting than the tool Fnac was previously using. The retail chain is now able to target promotional campaigns at the right people ‘on the fly’. Managers using the platform ‟just need to go to the Fnac product catalogue on the platform and select the product or brand codes and the sales results will be instantly displayed. You could for example find out quickly and easily that 100 packs of the Star Wars movies had been sold in the last month,” underlines tinyclues founder David Bessis, who is a former mathematics researcher at the French National Centre for Scientific Research. ‟The model extracts the population you want a report on,” explains Bessis, ‟so that you can see who you should be targeting your product at”. This method will enable Fnac to use its data in a ‘smart’ way. Previous customer targeting models were based on a somewhat simplistic retargeting approach, but tinyclues’ probability model is based on ‟very finely-tuned rules which, instead of taking a wide-ranging view of customer behaviour in terms of which areas of goods s/he is interested in, takes a much more specific view based on items s/he has actually purchased.”

Potential for 30% extra turnover

The methodology, which is highly sophisticated, ‟would be difficult for a company such as Fnac – which isn’t a technology firm – to implement,” underlines David Bessis. The startup is in fact already bringing real added-value to Fnac’s business. The method has proved highly effective with other clients, for whom ‟marketing campaigns using the tinyclues system typically deliver 30% additional revenue”, says the startup founder. Bessis, who has been working with the Fnac Group since last year, is amazed how the Big Data market has matured in that time. ‟A year or two ago, at the same Big Data event, many firms were showcasing projects that were rather ad hoc, experimental efforts,” he reveals. Today, tinyclues is working with the giants of the e-commerce world. ‟Our first major e-tailer client was a seasoned online ‘pure player’ – Paris-based Priceminister – which has a very strong innovation culture.” Bessis points out that some industries are ‟still very much in experimental mode when it comes to Big Data,” but he believes this will not be the case for very much longer. ‟The fact that firms such as Fnac and other clients of ours such as [French mail order company] 3 Suisses – companies whose customer relations go back a long way and that are perhaps justifiably more cautious vis-a-vis the use of data – are now prepared to use our solution is a sign that things are changing.”


Legal mentions © L’Atelier BNP Paribas