This $4 billion company is betting big on Google Cloud as it makes its algorithms smarter and its employees more productive
- Credit Karma CTO Ryan Graciano tells us that the personal finance company is betting big on Google Cloud to do all of its data processing.
- He says that the bet on Google wasn't based on saving money on servers, but rather because it makes the company's 350-person-plus engineering team more productive.
- He also says that eventually, Credit Karma will want to shut down all of its data center infrastructure, which is where most of its applications and data are still hosted.
Google Cloud is hustling hard to sign on large enterprise customers, as the company works to close the gap with the leading Amazon Web Services and second-place Microsoft Azure.
So it's a solid feather in Google's cap that Credit Karma - a well-known personal finance service with 80 million users, and said to be valued at $4 billion - is moving the entirety of its data processing services to Google's Cloud, in a project expected to complete by the end of the year.
It should be noted that Credit Karma relies on Amazon Web Services for some of its corporate applications. But for data processing, Credit Karma is already using the Google BigQuery data-crunching tool, and now plans to use it even more. Overall, Credit Karma does 8 billion predictions a day based on its data, in order to recommend credit cards and other financial products to its customers.
"It really feels like Google is a step ahead," Credit Karma CTO Ryan Graciano told Business Insider of the decision to bet on Google's data tech, specifically. "So why don't we leverage their infrastructure?"
Credit Karma was founded in 2007, when there "wasn't as much cloud," says Graciano. Amazon Web Services was in its infancy, and Microsoft and Google hadn't yet gotten into the cloud business. As such, Credit Karma's only real option was managing and maintaining its own servers in its own data centers, which it does a lot of today.
"We have a fair amount of physical infrastructure," says Graciano.
The conventional wisdom holds that much of the move towards cloud computing comes from a desire to cut costs - it's a widely-held belief that it's cheaper to host servers from the mega-efficient platforms run by Amazon and others than it is to keep your own.
That's not the case with Credit Karma, though, says Graciano. The real value, he says, is that Google Cloud makes it way faster for Credit Karma engineers to build, test, and deploy new concepts and updated models. Around half of its 700-plus person workforce are engineers of some type, so those productivity gains add up very quickly, he says.
"It's all about developer efficiency," says Graciano. "I can try to cut costs on servers, or I can make those people way more efficient."
Where before, a new service or model might take 2 years to see the light of day on Credit Karma's old systems, Graciano boasts that the turnaround time now could be 2 days.
"That's way better than 2% savings in capex," says Graciano, referring to the capital expenditures needed to maintain a data center.
He says that there was some initial pushback on the project from some of the Credit Karma engineering staff, especially those who worked most closely with the data centers and server infrastructure. As the project has progressed, though, and they've seen those productivity gains, he says they've come around.
It'll be a while before Credit Karma moves everything over to the cloud, says Graciano, just because there's so much data there - the company collects 8 to 9 terabytes of data a day, all of which has to be accounted for and migrated. Still, the eventual goal is for Credit Karma to be 100% based in the cloud. At that point, the data centers could shut down.
"We have this vision of getting to cloud as kind of this final, major step," says Graciano.
Once Credit Karma gets all of its data processing into the cloud, Graciano says that Credit Karma plans to take advantage of other Google Cloud tech, particularly around artificial intelligence. Credit Karma already does a lot of predictions; Google's AI could help them do more, faster, he says.
"I want them to keep innovating there, because I want to take advantage of those things," says Graciano.
He praises the experience of working with Google, but demurs when asked if this means that Credit Karma will ditch Amazon Web Services and go all-in on Google Cloud.
"I think it's too early to call. We want to see this project through," says Graciano.