Box's Sam Schillace on enterprise software's hard curve ahead

Schillace talks the intersection between collaboration and storage, data center scaling, one-day release cycles and where enterprise software is going.

As head of engineering at Box, Sam Schillace has a bevy of interesting problems to solve. First, he needs to figure out security, storage and scaling capacity at Box's rapid growth clip. He also needs to think about where Box as a platform and a product has to go to stay ahead of the curve.

"At our scale and growth we can't be reactive. We need to be predictive and solve for problems 18 months down the road," said Schillace.


Prior to joining Box, Schillace was engineering director of Gmail, Google Docs, Google Calendar, Google Groups, Google Reader, Blogger and Picasa. He also had a stint at Google Ventures. Schillace co-founded Writely, which was sold to Google in 2006. 

We caught up with Schillace to talk shop, software development, and where enterprise software is headed. Here are the highlights:

What are the big ideas you're wrestling with? Schillace said his first six months at Box were spent learning the players, culture, and strengths and weaknesses of his technical team. In the middle of that learning exchange, Schillace was rapping with CEO Aaron Levie about collaboration, the future of the file and what Box needs to become down the line. "We talked about the future of the file and collaborating in databases," said Schillace. "Content is the core of every business and what happens around it — collaboration, conversation, production — is the business interaction." The catch: The future of the file has been talked about for two decades.

Collaboration is a hard product nut to crack. Schillace, who had experience with both online documents and collaboration via Writely and Google Docs, said it's difficult for a collaboration app to rise above the noise. "Collaboration is fragmented," said Schillace. He isn't kidding. There are multiple vendors, but for enterprise collaboration, workers need to use the software. After a while, collaboration is weighed down by the one more tool problem. "It's hard to add value in this space (collaboration)," said Schillace. "We're looking into it to understand it better. We'll launch as an experiment and put it in front of users. Collaboration has to ease a pain point." Schillace added that collaboration in Google Docs started out as an experiment, not a formal project.

The role of the CIO is changing. Box has a customer advisory board and the big takeaway is that CIOs are no longer leading and in control of tech buying. CIOs are the ones who orchestrate. "Who's responsible for looking at the pieces and how they work together?" asked Schillace. Ultimately that fix-it person will be a CIO even if CXOs are buying their own services.

Developing features quickly. Schillace said that he moved Box releases from weekly releases to daily and even twice a day. That daily development release approach is common at consumer companies like Google and Facebook but rare in the enterprise world. "Internally, the daily release is largely a mindset. You're more willing to experiment and more willing to wait until the code is really ready," explained Schillace. "Longer dev times mean that bad code is sometimes pushed out because the next release may not be for another three months. There's a higher level of quality. It's counterintuitive."

Products benefit from iteration. Schillace estimated that he has built about 20 products in his career and "has been wrong about all of them in some substantial way." "No one predicts what users want," he added. The only way around that conundrum is to iterate and experiment.

Does the enterprise squeal over frequent releases? Schillace said the daily releases are either used internally or shared with customers who want to test the latest wares. "We have to be careful about what we build and what the chunks are," said Schillace, who added that Box is careful about touching its APIs.

On data centers, Schillace said Box is thinking through expansion international and what architecture works best. "We're getting to a scale where it's hard to move data around by pipes. In some cases, it's easier to put the data on a truck," said Schillace. Box is moving to a globally-distributed architecture because the company can't serve Europe or Asia Pacific from the West coast.

Where is enterprise software headed? Schillace noted that the current state of software feels a lot like 1997 to 1999, when desktop-first applications gave way to the Internet. Today, applications are giving way to mobile and the cloud, two categories that are likely to collapse into one. "We've hit some tipping point with mobile and cloud," he said. As a result, there are more vendors in the enterprise space than ever. "The more stuff that's brought, in the more complicated it is for vendors," said Schillace. Long-term winners need to "focus on the needs for the enterprise, but have a consumer grade user experience. You have to build to that level of quality because you don't have the CIO pushing for you. You have to fight for distribution."

Show Comments