italiano | español


Other Publications

Analytical Intelligence: the PRECISION and SIMPLICITY of the clock

Authors: Alberto Saccardi

"The invention of the clock was the highest attainment of medieval intelligence in the field of mechanics. Revolutionary in its conception, it was actually more innovative than what its inventors thought. It was the first example of a digital tool, not an analogue one: it counted a regular and repetitive sequence of distinct actions (the oscillations of a control mechanism), rather than following a continuous and regular movement like the shifting shadow of a sundial or the flow of water."

From "The Wealth and Poverty of Nations" by David S. Landes, emeritus professor of history and economics Harvard University

2001 a CRM odyssey - How to find your way around a Customer Database

Event: SEUGI (2001)
Authors: Fabio Marchetti, Filippo Avigo, Alberto Saccardi

Two brands; three networks of financial consultants; a traditional branch network; virtual channels which saw an increase of over 700% in terms of both clients and total assets in the year 2000: these are the reasons behind the decision taken by the Marketing Service of the Bipop-Carire Group to develop a dedicated customer-centric data infrastructure.

To build the Customer Database (CD) the enterprise data warehouse has been integrated with a marketing module for an all-round view of Group clients and a functional data model for Data Mining has been developed. However, not to get lost in a myriad of data and to undertake a proper CRM strategy require appropriate skills and tools to generate relevant market and customer intelligence from the CD.

In this paper we will show how we approached and resolved the essential issues in the development of the CD and how we are using it:

  • to instill trust in new clients attracted via virtual channels;
  • to drive and measure BIPOP CARIRE CRM Campaigns;
  • to produce a Profit and Loss statement for marketing activities.

Download the document (.pdf 71.9 KB)

Enterprise Miner® templates for database marketing applications

Event: SEUGI (1999)
Authors: Federico Ambrogi, Guido Cuzzocrea, Massimo Saputo

Treasure island, an enormous amount of data at your disposal and the best software environment to exploit it. Data Mining is an opportunity to gain a competitive advantage and should supply the marketing decision-makers with operative tools for the evaluation of client potential, the planning of an integrated offer and the selection of campaign targets.

Nevertheless data and technology are necessary but not sufficient. To successfully apply Data Mining solutions to Database Marketing activity you need to perform a process flow that starts from goal identification and then passes through making data available, and applying the SEMMA methodology. But this is still not enough: you need to implement the campaign in production, considering not only the new targeting rules, but traditional criteria, cross-sections and control tests.

All this comparing budgets with expected break-even per segment and averaging market penetration goals with profit maximisation ones.

Do you feel like Indiana Jones? What if you had a detailed treasure map to get you out of the jungle alive and smiling?

Download the document (.pdf 203.0 KB)

Data Model for Database Marketing Activity

Event: SEUGI (1999)
Authors: Frank Fiocca, Alberto Saccardi

What kind of data do we have collect to optimise C.R.M.? Which data do we need to understand the market and our competitors? Which logical subjects should we consider in a marketing database? How detailed and, how summarised should our data be? How should it be organised?

These are a few questions that have to be dealt with to design a marketing database for decision support. We’ll illustrate the basic guidelines to define the correct data model for a powerful database marketing activity. In particular we’ll present some cases we have encountered and resolved in different sectors.

Download the document (.pdf 194.3 KB)

How many good fishes are there in our Net? Neural Networks as a Data Analysis tool in CDE-Mondadori’s Data Warehouse

Event: SEUGI (1996)
Authors: Guido Cuzzocrea, Alberto Saccardi, Giovanni Lux, Emilia Porta, Arianna Benatti

In Data Analysis, Neural Networks (NN) are not a universal problem-solver, not a completely user friendly tool that might consider to obtain the best and quickest answer to the most complex question. In spite of that, it would be wrong to ignore the possible advantages of NN in analyzing real-world databases, when prior hypotheses are poor and linear modeling inadequate.
In classification problems such as discriminating between heavy buyers and non buyers in Direct Marketing, NN seem to be a powerful analysis tool. Are data manipulation and standard statistics therefore useless? Our experiences is that only an integrated approach produces the best results: knowledge discovery in large databases means goal identification , sample design, a reduction in data dimension, variable selection, model building (e.g. NN architecture), simulation and probabilistic error measurement.
The paper is step by step description of Neural Networks and other Data Analysis tools used to classify CDE-Mondadori’s Customers and Prospects: we will show a concrete and successful example of Data Mining with SAS System in a Data Warehouse.


Download the document (.pdf 269.8 KB)

Statistics and Democracy: prederence mapping in opinion poll analysis

Event: SEUGI (1995)
Authors: Guido Cuzzocrea, Alberto Saccardi, Valeria Severini

The Italian socio-political context has been recently characterised by important events such as the introduction of the majority electoral system and the disappearance or substantial transformation of traditional parties: these changes have highlighted the necessity of analysing the new characteristics, motivations and needs of potential voters.

Opinion polls should focus their attention not only on voting behaviour forecasting, but on voter profiles as a complex phenomenon. The analysis of results by traditional tabular layout is not suitable for the joint treatment of multivariate data matrices, emphasising singular aspects and often leading to biased interpretations. In our approach we combine the use of statistical multivariate techniques with techniques for data organisation: we will show how to give the view of the entire forest rather than that of single trees using the SAS System.

Download the document (.pdf 276.3 KB)