A section of the Observatory monitoring activity is dedicated to the prediction of the failure of Micro-Small-Medium enterprises using data from corporate websites. The primary goal of this research is to identify new indicators that can aid in predicting default, complementing the traditional indicators based on financial statements. The advantage of this approach lies in the characteristics of the web-based features that are based on open, granular and updated data.

The first experiments, as well as the subsequent analyses on large samples of companies, are described below.

Prediction of SMEs Bankruptcy at the Industry Level with Balance Sheets and Website Indicators

SMEs default forecasting models are often based on balance sheet indicators. Unfortunately, such indicators are available with a great delay and are not always effective for all industries. In this paper we try to investigate whether features built on corporate websites can be a valid alternative.
Read the paper to learn more

Websites’ data: a new asset for enhancing credit risk modeling

The paper compares two types of SMEs default forecasting models: one built with balance sheet data, the other built with corporate website data. The exploration highlights that the two sets of information recognize different types of defaulted companies.
Read the paper to learn more

Non-conventional data and default prediction: the challenge of companies’ websites

The paper illustrates how corporate websites are built and organized. This allows us to briefly indicate the process of building web indicators.
Read the paper to learn more

Predicting SME’s default: Are their websites informative?

The paper compares the performance of alternative (linear and non-linear) statistical default forecasting models based on web-based versus balance sheet-based indicators.
Read the paper to learn more