Abstract della presentazione di Giancarlo Vercellino, Research & Consulting Manager di IDC Italia, tenuta nel corso dell'executive meeting 'Big Data e Real-Time Analytics' svoltosi a Milano il 25 novembre 2015
4. European Organizations Gaining Value from Big Data
Industry-wide
insight from one-
off analysis
Optimizing a
previously
manual process
to run it at scale
Data-driven business
combining complex
models and sensor data
Danish wind-turbine provider
runs on Big Data, combining fluid
dynamic modelling with real-time
sensor data to optimize turbine
placement.
French retailer identified the
factors causing some stores to
perform far better than others,
with a causality analytics engine.
German retailer saved EUR60m
by optimizing distribution of
perishable groceries, with a retail
demand forecasting engine.
9. IDC Italia
Viale Monza 14
20127 Milano
Tel: +39 02 28457339
gvercellino@idc.com
Giancarlo Vercellino
Research & Consulting
Manager
IDC Italy
www.idc.com
Editor's Notes
European organizations are gaining value from Big Data in many ways. It’s not just about saving costs, or increasing the scope of enterprise reporting – there are many ways Big Data can impact performance. Here are some examples:
Data-driven business
Vestas
http://www-01.ibm.com/common/ssi/cgi-bin/ssialias?infotype=PM&subtype=AB&htmlfid=IMC14702USEN
Causal factors for one-off analysis
A retailer had a number of stores that performed poorly, but could not work out why this was. Retailers generally know this from experience but in this particular case the management could not understand the reason for the problem. The company loaded about 200 attributes into a causality analytics engine and found out that there were two factors affecting store performance: firstly, the area (square meters) of the store, which all retailers know about; secondly the length of shelving at child height. The successful stores had more shelving at child height. This was something the retailer had never considered – now this success factor is something the whole industry can understand. So this is not a traditional use of BI in the sense of repeated, scheduled reports – rather it is an in-depth analysis that ran once, to solve a single problem.
Automating a previously manual analytical process to run it at scale
A large German retailer saved EUR 60m over a few years by replacing a spreadsheet and expertise based system with a predictive analytics engine. For produce retailers, particularly short shelf life goods like salad, sending the right stock levels to the right stores has a huge impact on the bottom line of the organization. The decisions on where to send these goods are affected by dynamic conditions including weather, sporting events etc. so a huge amount of data needs to be included in the process in order to make an accurate prediction.
Background info on retail demand planning:
http://www.blue-yonder.com/meldedownloads/automated-decision-making-saves-otto-millions.pdf
Identify an existing problem
List possible alternatives for solving the problem
Select the most beneficial of these alternatives.
Implement the selected alternative.
Gather feedback to find out if the implemented alternative is solving the identified problem.