A stalwart of the early web, Shutterfly rebuilds with the future in mind

There’s a relatively short list of Internet companies founded in the late 1990s who are still considered successful.

But Shutterfly, founded in 1999, must invariably be included alongside names like Amazon (founded 1994), EBay (1995), and Priceline (1998) are among those who not only survived the Internet bubble of the late 1990s, but has thrived since.

In an age when consumers snap photos on their phones and share them globally on Instagram and Snapchat, Shutterfly’s business of selling personal photos printed in books, on greeting cards and coffee mugs seems as quaint a success as reruns of “Home Improvement.”

Having fended off literally hundreds of photo-sharing startups in its 17 years, Shutterfly finished 2015 with revenue north of $1 billion, having grown both organically and through acquisitions from $50 million at the time of its 2006 IPO.

Now this unlikely survivor from the Internet’s early years is looking to put itself on a solid tech footing to ensure it keeps growing. Among the priorities has been a multi-year effort to modernize its tech infrastructure for new efficiencies – and new business opportunities.

That made for an interesting conversation with James Berry, Shutterfly’s VP for manufacturing and operations technology, when we shared a stage at Apigee’s Adapt or Die conference in San Francisco in September. (Google closed on its $625 million acquisition of Apigee last week.) 

“When you grow up in a world where you have to build everything in order to be successful, that gets baked into your DNA,” he said in an off-stage interview. “Shutterfly was a traditional monolithic Java application running on an Oracle database. …This was the tried-and-true method, it was how you built things in 1999 and 2000, but the industry has moved a long way since then.”

It sure has. Newer websites like Hulu and Square and Airbnb are built using Ruby on Rails, and Node.js and use modern web APIs, making them more flexible under the hood and easier to update frequently. That added flexibility also makes it easier to securely share data and services, experiment with new business models and to cut the sort of third-party deals that come naturally to most modern Web companies.

Berry recognizes that this is where Shutterfly needs to be in order to innovate and survive in today’s digital economy.

“As the rest of the world moves on and evolves, the companies who can follow suit are the ones willing to jettison things they once thought necessary but no longer are,” he said. “The ones who can’t do that struggle along and either die a slow death or flame out in a spectacular way.”

Berry admits that digital transformation with an eye toward the future can be tough, especially when there’s no immediate pain. “ When you have a billion in revenue coursing through your veins, it’s hard to get comfortable with the idea that to really unlock your business you have to run around breaking stuff.”

But not everything. Berry left Shutterfly’s commerce systems alone during the first few years of the transformation work, and devoted resources instead to re-architecting the customer experience around creating and customizing photo books and other products. 

“We started on the things that we knew we could learn from, and as we got more comfortable we took on the meatier problems,” he said. “We’ve been incrementally improving that experience, and each version is a little more modern and moves away from that old monolith,” he says.

Another key tool: APIs. Shutterfly has used software APIs for evolving its mobile commerce experience to perform analytics on its massive library of images – amounting to hundreds of petabytes — and to streamline its manufacturing and logistics operations. And while there are some external developer programs too, most of the attention goes to supporting internal teams.

For now. The modernization efforts are laying the groundwork for the day – Berry couldn’t say when it might be — that those internal APIs get exposed for external developers in a way that most Web companies tend to consider essential.

“We have plans to open them up. There’s just nothing concrete to say yet,” he said. “We’re far enough along at this point that selective systems are where we want them for internal consumption.”

And once that happens who knows how it might play out? Berry declined to speculate on specifics, but it’s not hard to imagine Shutterfly using APIs to branch out into new opportunities beyonds its current branded sites. It will be interesting to watch.

The first mission is to finish the rebuild of the core systems. “We decided early on that we wanted to rebuild them with the future in mind,” he said. Once that’s done the old monolith will be out of the way. The results are starting to show. There are now 50 different parts of Shutterfly’s core systems that can be updated individually. That’s up from zero four years ago. That’s progress, but there’s still more to do.

“We decided early on that as we’ve been rebuilding our core systems that we want to design with the future in mind,” Berry says. “We want as few constraints on our future potential as possible. We want to be where the users are.”