Big Data: How Kodak Gallery Sees the Big Picture

By Laurent Valadares  |  Posted 02-01-2012
By Laurent Valadares

As an online destination for storing and transforming photos for more than a decade--with more than 70 million customers and billions of images stored--Kodak Gallery has the big data in-house to prove it. In addition to our massive image data sets, we process nearly 15GB of new text data every day. Big data is, in effect, our business.

Managing all that information for customers is just one aspect of our big data challenge. It is equally important to give Kodak Gallery business analysts the tools necessary to maximize the value of that data. With this information, they can develop new products, such as our Group Albums mobile app and service, as well as track the success of various promotional campaigns. To say the least, our business environment requires advanced analytics technology that will not buckle under heavy workloads or huge data sets.

When we needed a new data warehouse, it was paramount that we deploy one that could handle the load while delivering targeted, timely and insightful reports to the business. We chose the Sybase IQ purpose-built analytics data warehouse paired with SAP Business-Objects for our business intelligence tools.

Before deploying the combination, Kodak Gallery had islands of information, which is common in enterprises that don't have a centralized data warehouse repository. This led to individuals performing their own analysis in a vacuum, often on spreadsheets. It was difficult for management to make informed decisions based on a common, centralized view of the data. Once we integrated all the data into our system, we no longer had to worry about these islands of information.

As a result, we now have a single view that presents accurate information of all key marketing activities in multiple geographies, such as:

product sales and trends over time

promotional campaign effectiveness

fulfillment program results

analyses of product launches.

For example, business analysts instantly can see how a particular marketing campaign is doing and make incremental or major changes to it in real time to improve results.

After deploying the new data warehouse, the enterprise reporting team also noticed an increase in the efficiency of the company's CRM system. This was directly due to the preprocessing of data performed inside Sybase IQ prior to sending the data to the CRM system. So, as a side benefit from the deployment, we were also able to consolidate the CRM system and the analytics to provide the data feeds from a single data warehouse source.

Finally, the shared-everything architecture doesn't need the increased CPU power, increased memory and increased storage that other platforms we evaluated require. Also, with its columnar-based database, it delivers a 10:1 compression ratio for stored data, meaning we need much less storage capacity than other systems would demand. In addition, while Kodak Gallery overall has grown in size, the enterprise reporting team has not had to expand, while continuing to provide the information the company requires.
As a result, our TCO is remarkably low.

The big data era may make some companies nervous. But we are exploiting it, leveraging it, learning from it. You might say we get the "big picture" into our business that big data gives us.

About the Author

Laurent Valadares is director of Research, Analytics and Reporting at Kodak Gallery.