Shopzilla, a popular online comparison-shopping search engine, needed a solid IT infrastructure to manage its massive volumes of information. Using Sybase® technology, Shopzilla developed a solution that delivers reports in seconds with an easy implementation and reduced administrative expenses.
- Queries that took Shopzilla up to three minutes to process in a standard RDBMS are returned in one second with Sybase IQ.
- Reduces query response times by 99 percent
- Lowers administration costs
- Improves customer satisfaction
- Easy installation
- Business services
Shopping for Customers
Shopzilla.com allows thousands of shoppers to search 30 million products from 40,000 stores. As a result, Shopzilla is an information factory, gathering millions of data points about shoppers and making that information available as a product.
“We gather demographic and psychographic information,” said Henri Asseily, chief technology officer and co-founder. “We ask people what they think of the merchant and how the merchant can improve.”
Shopzilla runs several different programs geared toward gathering data to improve merchant performance. An online retailer can receive Gold Store status by allowing Shopzilla to survey its customers directly after a sale is completed and then after the scheduled product delivery date. The company offers reports that lay out trends according to pre-selected performance metrics and conducts custom studies for clients.
Market research is also a significant part of Shopzilla’s business model. The more information it could gather and pass on to its retailers in a timely fashion, the more valuable that service would be. So while the company had a standard data warehouse in place, it could only produce monthly reports based on canned, prepackaged queries.
“We were looking for new tools so we could offer data in real time and allow merchants to query the data any way they wanted,” Asseily said.
As he embarked on the project to upgrade Shopzilla’s analytical and business intelligence infrastructure, Asseily had several criteria in mind. He wanted a system that would run in parallel with the company’s cornerstone Web operation and to simplify its infrastructure.
Asseily quickly determined that a system based on a typical relational database management system would not suffice. “A standard RDBMS would not allow us to do this. They are geared toward gathering data, not reporting it,” he said.
After surveying the market, Asseily opted for what has been dubbed the iForce Solution for Economical Data Management, a joint program from Sybase and Sun Microsystems. The iForce solution consists of the Sybase IQ running on Sun servers based on the data warehouse reference architecture Sybase and Sun had published.
“The value proposition for the reference architecture was for very large databases, but this is a good solution for the mid-market as well,” said Frank Teklitz, director of worldwide IQ programs at Sybase Business Intelligence.
The Sybase Business Intelligence division has used the iForce package to expand its marketing efforts. “This is a nice combination for resellers,” said Deborah Harrington, general manager of worldwide IQ operations at Sybase.
The first packages based on the reference architecture, she noted, allowed Sun-based systems to compete at the high end of the data warehouse market. This package allows Sun and Sybase to expand the market for DW/BI solutions. “The whole solution is designed for the reseller market,” she said.
Sybase IQ uses the same code base as Sybase’s iAnywhere database geared toward mobile and embedded applications. Tables are built column-by-column rather than row-by-row. According to Sybase statistics, this approach delivers ad hoc query performance as much as 100 times greater than a standard RDBMS. It also enables users to compress data from 30 to 60 percent, offers near-linear user and data scalability and is easier to maintain.
According to Asseily, the iForce solution has delivered on its promises. “Queries that took up to three minutes in a standard RDBMS are returned in one second with IQ,” he said.
Indeed, merchants now have access to 13 months of data made up from 15 million customer data sets that they can analyze in a wide range of ways on an ad hoc basis.
The entire project took about four months to complete. “Most of the work was in developing the front-end query tools,” Asseily said.
He added that the system was easy to install and requires very little administration.
The big payoff, however, has been in the response of Shopzilla.com’s participating merchants.
“This is a revenue line in our business,” Asseily said. “The customers say it is incredible. They are blown away by what they can do.”