Rapid and efficient big data analytics

Yellow Pages is a digital media and marketing solutions company serving 260,000 small and medium-sized businesses across Canada. In recent years, it has transitioned from a print media company to one that offers a full spectrum of digital marketing solutions and services for businesses. The organisation’s Customer and Advertiser Analytics group provides advertisers that run ads via the company’s web and mobile properties with information they need to measure return on their advertising dollars and track the success of their campaigns.

“Many of our advertisers are small and medium-sized businesses like florists, dentists and plumbers,” explained Richard Langlois, director of customer and advertiser analytics at Yellow Pages. “They aren’t always experienced in digital media and are often pressed for time, so we need to present data to them in a very intuitive and efficient manner.”

Yellow Pages chose the WebFOCUS business intelligence (BI) and analytics platform from Information Builders to support an application called YP Analytics. Yellow Pages customers use this dashboard to track interactions and measure key performance indicators focused on ROI and revenue. WebFOCUS integrates with Infobright’s analytic columnar database platform to accelerate query response times.

“Our job is to provide facts to our customers, so they can see how their digital advertising campaigns are working and what value these campaigns are delivering to their business,” Langlois said. “WebFOCUS is among multiple tools we use to help accomplish this.”

In daily data operations at Yellow Pages, approximately 52 billion rows (9 TB) of raw data are loaded and processed, containing a rolling 25 months’ worth of data. Reaching more than 20 TB of total stored data, YP’s systems must be properly packaged to enable the speed and accuracy advertisers require.

Information from Hadoop is migrated to the analytic database platform, where it is compressed to an average ratio of 35:1. Yellow Pages uses Hadoop to cluster the entire table every day to ensure fast response times. Benchmark tests have shown that response times for both canned and ad hoc queries are in the sub-second range for small and medium-sized businesses. For large advertisers with thousands of merchants, the response time may vary from two to 10 seconds.

“Because YP Analytics is a customer-facing application, performance is critical,” added Langlois. “The YP Analytics application defaults to one year of data, which represents millions and millions of events. WebFOCUS integration with other database software is changing the game for us. Such rapid response times with that volume of information are just amazing.”

Information Builders was chosen for its performance and scalability, as well as its intuitive development environment. Mobile capabilities, which enable YP customers to access and analyse information via smartphones and tablets even when they are disconnected, were also key factors.

“We needed to present our big data to our customers in a way they can easily interpret,” said Langlois. “They know they got 35,000 impressions, 800 clicks and 32 calls last month, but what does that mean? YP Analytics will show them more than just that raw activity; it will help them measure conversion rates and determine how large any related sales transactions were. In essence, it will translate digital advertising activity into dollars.”

Information Builders’ Professional Services team was on hand to help with the implementation of YP Analytics. They assisted with installation, configuration and performance tuning.

Image credit ©iStockphoto.com/Warchi