Big data in the cloud a plus for SellPoints

When a retail marketing services provider found itself with a huge repository of data, it knew it spelled opportunity.
Written by Drew Turney, Contributor

Emeryville, California-based SellPoints used the cloud to create a completely new line of business based on the market insights it was generating from customers' data. Driven by performance marketing and analytics vice president Benny Blum, the company adopted a cloud computing solution to crunch their own big data product in 2012 using Hadoop as a service from provider Altiscale.

SellPoints' performance marketing and analytics vice president, Benny Blum

SellPoints delivers rich media to online and in-store point of sale systems and social media. With 251 retail partners including Wal-Mart, Costco, Best Buy, and Office Depot, and 130 million unique customers a month, that's a lot of customer behaviour.

With so much data collected from those partners, the company realised there was a business opportunity in higher-value services delivering market intelligence to its clients. The result is Omni, a Hadoop-driven platform that provides reporting and management for brands and retailers to see the impact of their marketing and sales efforts in real time.

"We wanted to create several proof of concept platforms to demonstrate the potential," says Blum. "The most successful would then be the basis for the production version that would be offered to customers."

The company created the first build on Amazon's Elastic MapReduce cloud service and RedShift, Amazon's data analytics platform. The second was built on the Altiscale Data Cloud.

Blum says several factors influenced the decision to go with Altiscale.

"Amazon's pricing structure wasn't predictable enough for the type of bursty workloads Omni would be producing," he says. "In Altiscale, user data is always available to the cluster and they provided a service level agreement of performance and reliability that we liked."

SellPoints realised that using Amazon meant they'd take on the majority of the management burden, which would mean investing more in people with Hadoop expertise. Blum said Altiscale's purpose-built Hadoop infrastructure was the only model that offered the agility for faster go-to-market with less risk. Omni was launched and the company had a new revenue stream immediately.

"Altiscale's at the core of our growth initiatives," Blum says. "They let us experiment with development at far lower risk than if we'd built our own infrastructure."

His advice for companies using similarly specialised software-as-a-service frameworks like Hadoop is to perform a careful cost analysis to understand how different offerings might affect you. More importantly, he says, you're outsourcing computing for a reason — so your people don't have to look after it.

"Make sure the service is easy for your internal staff to use."

Editorial standards