Case Study

Enabling 1-to-1 marketing with big data

Transforming complexity into an advantage

When four brands with different identities and market positions find themselves co-existing as part of the same Group, one of the key needs – not only for those in charge of data – is to understand how to manage and standardise the enormous amount of heterogeneous information available to the company. This is the scenario for Prénatal, Bimbo Store, King Jouet and Toys Center which since 2015 have been brought together under the control of Artsana S.p.A.

The different loyalty programmes and marketing strategies adopted by the individual brands adds yet another level of complexity to the problem. A Data Driven approach enables the Group not only to bring order to the various processes, but also to open up new and previously unexplored avenues.

Data integration

The role of Data Reply was central to the creation of a Data Lake: a single point where the data from the different channels – both digital and physical – is brought together. Sales data from physical stores and from the e-commerce channel, customer master data acquired from loyalty programmes and product data.

The implemented solution is highly flexible, thus facilitating the seamless integration with other systems or operating units. Data integration is the first step in building machine learning algorithms that enable the modelling of individual customer behaviour. The final result? The implementation of genuine 1-to-1 marketing campaigns.

What we think in relation to the consumer is almost always wrong.
The only way to know your target and to understand how they behave is by relying on data.

Marco Massara
Prénatal Retail Group Chief Digital Officer and CRM Director

Once processed, the new information made it possible to define a highly personalised customer journey programme: a win-win approach aimed at improving the customer experience and strengthening the relations between the company and its customers. The project made it possible to obtain often unknown yet crucial information, thus providing the marketing team with new tools consisting of machine learning algorithms.

Hidden child detection

Most customers sign up for a loyalty card when their first child is born. Relations with the company continue and, as time goes by, the family unit often expands. Typically however, customers will continue to use the same card, acquired years earlier, without notifying the company of the new child’s arrival. The algorithm developed by Data Reply can identify, with an accuracy of 87% and with just 12 products purchased, the existence of additional children - besides those indicated when the customer first signed up for the loyalty programme - and can estimate their age with a 3-month margin of error.

Purchase propensity score

Knowing if a customer is loyal, at risk of churning or if they have chosen not to return are some of the most valuable pieces of information for a company. Indeed, finding a way to narrow down the field of action and focusing on customers who are at risk of abandonment enables companies to save resources and to better personalise the offer. The algorithm developed by Data Reply is able to assign to each customer a value that indicates their propensity to buy, making it possible to predict their probability of return after 3 months, with a 77% degree of confidence.

Customer lifetime value & product recommender

In light of the new information obtained thanks to the data analysis carried out, it becomes possible, on the one hand, to estimate the value of the relationship between the company and a specific customer, and on the other, to identify their individual tastes. The algorithm designed to calculate the so-called “Customer Lifetime Value” makes it possible to understand, at the highest level, the profitability of individual households and, in more detail, how much each of the children contributes to this value. On the other hand, the “Product Recommender” takes into account personal characteristics such as the child’s age, as well as product characteristics, such as seasonality and price range, to identify and recommend the best set of objects in the unified catalogue. Machine learning algorithms therefore make it possible to understand which customers to focus on and to “activate them” thanks to ad hoc recommendations made on the website, personalised Direct Email Marketing campaigns and discounts on specific products.

Prenatal Retail Group

On 9 October 2017, Artsana S.p.A. acquired full control of the Prénatal Retail Group, taking over Giochi Preziosi S.p.A.’s share in the joint venture initiated in 2015 for the joint management of the Prénatal, Toys Center, Bimbo Store and, in France and Switzerland, the King Jouet chains. Through its brands, the Prénatal Retail Group offers products and services for pregnant women, mothers and children. Currently, the Group has more than 700 stores, approximately 4000 employees and a total of 24.3 million customers served.

Data Reply

Data Reply is the Reply Group company specialised in advanced Big Data and Artificial Intelligence solutions. Covering a wide range of industries and business functions, we support executive-level professionals and CEOs, enabling them to derive value from data. We build Data Platforms, define and implement Machine Learning and Artificial Intelligence models based on an efficient, replicable and scalable approach, thanks to an extensive pool of resources with advanced skills in Big Data Engineering, Data Science and Intelligent Process Automation. With a strong and continued focus on innovation, we are applying Quantum algorithms to support the optimisation of processes characterised by high computational needs.