jPage: Guavus

In the HPC Hunger Game of Big Data, how to make sense of large amounts of data? The following is a blog post by WSJ about how Gurvus tackles the challenge.

Words are a likely culprit, said Anukool Lakhina, the founder and chief executive of Guavus, a data analytics start-up serving the telecommunications industry. “The vocabulary that’s being used to describe this space is tools and technologies” and that confuses customers because it doesn’t help them understand how data can be used to solve their problems, he said.

Most of the emerging technologies that are behind the excitement around data, like Hadoop and NoSQL databases, imply that if companies store their data with the right tools, they’ll be able to extract value from it later. But how can companies expect to justify a $20 million investment in storage without knowing what the return will be, Lakhina said.

Though Guavus also provides storage tools–as well as every other piece of technology that a company needs to make sense of large amounts of data, from ingestion through analytics–it puts the analytics first in its sales pitch and that’s resonating so far.

Its customers are three of the four largest mobile operators in the U.S., including Sprint, and four of the top five Internet backbone operators. Most of its deals are for multiple millions of dollars, including one customer that’s generating at least $10 million in revenue for the San Mateo, Calif. company, Lakhina said, though he wouldn’t be more specific about the company’s finances or customers.

Guavus initially helps customers translate their business problems into data problems, and then over the course of about a month, Guavus implements its analytics technology onto a customer’s existing data infrastructure to demonstrate the value of the technology without a big upfront investment. Once the value is clear, customers are willing to expand their use of Guavus, Lakhina said.

It can help customers build capabilities to analyze the potential for customer churn, bundling and pricing and to analyze network traffic and routing.

Guavas was founded in 2006, but the genesis of the company goes back more than 10 years when Lakhina was working at Sprint Labs working on a project to collect sensor data to better understand its networks. Back then, the team would fill up its entire data storage system in an hour or two, remove the hard drives and mail them—the amount of data would have overwhelmed the network back then– to a central repository to be analyzed.

“We had continuous measurement of what was happening on the network, but what was missing was the ability to make timely decisions,” Lakhina said.

It took the company two years to learn the needs of the telecommunications industry, but Lakhina said the learning curve would be shorter as it branches into new sectors.


Though it started in telecommunications, Guavus has pilot tests running at companies in utilities, transportation and manufacturing. “All these folks are seeing the data avalanche, but are often ignored,” he said.


Links:

• Guavas