Guavus raises $30M to help telcos do big data at network speed

Guavus, a San Mateo, Calif.-based company focused on analyzing network- and other sensor-based data, has raised $30 million in financing from Investor Growth Capital and QuestMark Partners, along with Artiman Ventures, Sofinnova Ventures and Intel Capital. The new investment brings Guavus’s total financing to $78 million since launching in 2006, and comes at a time when mobile providers, especially, are fighting harder than ever to woo (and keep) skeptical customers.
In a world of streaming data and lots of it, Guavus’s strategy of analyze first, store later has been paying off. When I spoke with Guavus founder and CEO Anukool Lakhina in May, he said the company was already serving three of the top four mobile and wireline providers. They’re impressed with the platform’s ability to analyze data as it streams in off the network, meaning they can make insights closer to real time rather than waiting for data to write to a disk before querying it.
In the that old model, explained Lakhina — who used to work in the high-performance computing division at Sprint (s s) Labs (and who’ll be presenting at our Structure: Data conference in March) — companies would worry about storing the data first, but the queries never catch up to rate at which the data is flowing in. By analyzing the data as it inflows in off the network, carriers can do things like tell users their data usage in near real time or detect the source of traffic spikes. In the longer term, they could figure out customers’ activity around usage, or which apps (or types of phones) are responsible for the most data consumption, and try to build customized plans or pricing tiers.
And actually, Lakhina added, although Guavus is built to capture network data, “most of our engagement takes place beyond the network.” By fusing that data with demographic and other user data, providers can do a good job segmenting customers groups around usage patterns.
Ultimately, mobile providers and anyone else generating streams of data off their networks, while also serving customers, has to come around on the idea that they need to get more focused on that data. “Our customers are using data to optimize their businesses and delight customers,” Lakhina said, but that means breaking the legacy mold of capturing data, transporting it to a central location and then finally worrying about analyzing it. Rather, they have to “collect [data] once, then analyze many, many times” at every step along the way.
Feature image courtesy of Shutterstock user PunyaFamily.