- Financial Services
- Other Industries
- Telecommunications / Security
- Online Analytics
- OEM Partners
This global retailer, one of the world’s largest, operates thousands of stores and also has a fast growing online business. Increasing market share and revenue in the retail market requires a better understanding of customers than the competition, and this is true with online retail as well.
Overview download as pdf
In order to continue to increase revenue, a system that would analyze online customer behavior and enable sophisticated predictive analytics was implemented. Questions such as “if we speed up page loading by one second, what is the predicted impact on sales versus the systems cost of doing so?” would be answered through this implementation.
The first phase of the project was to extract and analyze detailed web data from the company’s Omniture service. The goal was to provide the detailed data to the desktop of the business analysts. Phase 2 will enhance that data with other information from sales and customer applications. The primary challenges this retailer faced was a result of the large volume of data generated each day and captured by the Omniture web analytics service. Each day about 40 million rows of data are generated, and the goal was to have one year of history in the database on a rolling basis. This data volume presented a challenge in terms of: • Time to load data • Query response time Initially the IT group tried to load the data into the central database (a traditional row-oriented database). The resulting query performance was very slow, as is often the case when trying to do analytics against large data volumes using a database designed for online transaction processing rather than analytic applications.
In order to bypass the corporate database and load data directly to their desktops, the business analysts became both the users of the system and the implementation team. While they were very familiar with using database technology, the database needed to be self-tuning and self-managing to avoid the need for database adminstration expertise.
- Infobright Sales Engineer
Working with Infobright, the system was set up and ready to go in about two weeks. This included automating the extraction of the relevant data from Omniture and loading it into Infobright on a daily basis, and fully testing the end-to-end process. The company achieved the goals they had set: • excellent query performance • automated and daily loading of data • managing the data in an efficient way • doing analytics more easily and without DBA assistance.
Infobright is an excellent tool to manage large amounts of static data. It is great to have control of our own data, and this has become an important tool for the analyst team
- Retailer’s business analyst
The key benefit of using Infobright is improved marketing efficiency and effectiveness.
A New Approach
The Analytic Data Warehouse
Traditional data warehouse products put a tremendous burden on IT in order to create and maintain an environment that will allow users to query against large volumes of data.