Google I/O 2013: Turning BigQuery analytics into marketing strategies

Google product managers describe how digital photo company Shutterfly is a case study example of turning big data into business without massive hardware and software investments.

SAN FRANCISCO -- If there is one thing that Google has a lot of (besides money), it's data.

Many developers attending Google I/O 2013 this week are learning more and more about using that data to their benefit, whether it be through mobile commerce via Google Wallet or location APIs for Glass .

One of the more complex but logical places to start would be with BigQuery, Google's Web-based service for interactive analysis of massive data sets. The Internet giant boasts that the system can mine and analyze terabytes of data in just one click.

Using digital photo service Shutterfly as a case study, BigQuery product manager Ju-kay Kwek explained during a session for developers on Thursday morning about how businesses -- notably the marketing departments -- can turn this data around into customer and product marketing strategies.

Derek Stevenson, senior director of data strategy and analytics at Shutterfly, acknowledged some of the challenges surrounding big data, quipping that "big" is always relative.

In the simplest terms, those constraints consist of scale, performance, cost, and, well, simplicity.

 Stevenson remarked that data is often seen as a competitive advantage, but that requires owning the data and methodology, an experienced analytics team, and decisive action.

"The basic idea is that marketing often has a particular budget that they can invest in their channels," said Stevenson, adding that the common question often revolves around return on investment.

As of 2012, Shutterfly had more than 19 billion images stored with seven million customers registered. That comes with good amount of data about images, locations, cameras, and the users themselves.

"If they don't have to spend their time preparing data and aggregating and building cubes -- I've been on the IT side, and I can tell you that's not the fun part of the job," Stevenson joked.

Stevenson pointed out that the data isn't "random," meaning that insights require educated interpretation.

But then the challenge becomes scaling that analysis to cover more than 19 billion images, or 80 petabytes of data.

In order to tackle these issues, Stevenson posited that "traditional approaches" require compromise, whether its sampling or aggregation. But Stevenson lamented these approaches drop out some details.

The goal, he continued, is to maintain those details in order to interprets and inform interactions with all customers.

Stevenson admitted that IT departments often bear the brunt of the "multi-dimentional challenge" surrounding big data in sifting through all the details about users, volume, mergers and acquisitions, and more.

To fill that "gap" with BigQuery, Stevenson described that Shutterfly switched to a strategy that consists of the following five steps: data collection, storage (on BigQuery and Google's cloud), preparing the data for analysis, and then acting on the findings through email campaigns, online merchandizing, and retargeting ads.

"If they don't have to spend their time preparing data and aggregating and building cubes -- I've been on the IT side, and I can tell you that's not the fun part of the job," Stevenson joked.

Newsletters

You have been successfully signed up. To sign up for more newsletters or to manage your account, visit the Newsletter Subscription Center.
Subscription failed.
See All