Betting the Firm on a Digital Infrastructure
I started in digital publishing at the onset — when you had no choice but to go out and roll your own infrastructure. There simply were no other options. Building from scratch was long and arduous, but as most anyone of a certain age will tell you, across any industry, you have a deeper understanding if you do things mechanically/analog-ally/manually/(wrongly) first — digitally second. Fast forward nearly two decades and publishers, marketers, businesses are afforded a luxury of rentable, turnkey, hosted digital solutions. Pay and play, as they say. But what starts as a luxury can become a handcuff — stunting imagination and curtailing opportunity.
At last week’s MarkLogic World, Brian Bishop, VP Platform Strategies at Springer, related his company’s path from renting to putting a stake in the ground and breaking ground on its own digital foundation. According to Springer’s 2011 Annual Report, e-revenues were just about 40 percent — and the board felt it was time to move off that 3rd party provider and be at long last the master of their own domain. MarkLogic would power the content platform — and it would be done in 11 months — before the service contract with their content vendor expired.
Bishop, who had started a “skunk works” project on MarkLogic a few years back — and had since created a handful more, knew the flexibility of the database. He was in love with MarkLogic’s “serialized search” the ability to save queries — which he used as an incubator for new revenue streams. “Whatever you can design as a query, you can sell – and in fact, you can sell the query!
Still, he had a dev team of 3 — a far cry short of what would be needed if they were to “recreate 36 man-years of code.” The
harbingers consulting team that came up with that number also laid odds that the project would fail — either not make the deadline or would be woefully lacking in features. The consulting firm’s assumption for failure was understandable. There is little information out there on how long it typically takes to rebuild a mature platform; Outsell recently did a case study on a company that did NOT use MarkLogic — and from beginning to end it was about 16m. These results hold true for most of MarkLogic other customers. Companies like McGraw-Hill Construction confirm that rebuilds take 15-18 months — usually because of the vast amount of time planning and modeling data — and with MarkLogic, they had similar results to Springer.
Despite the stacked odds, Bishop bit and took the challenge. He staffed up, cracked a few knuckles, hired a scrum master who could implement and teach them Agile and within 7 months had hit a major milestone that foretold the success of the project. The new, feature rich beta site was opened to the public last summer, and the migration was finished in November — a full month ahead of schedule. For beating those odds, I hope the payout is HUGE!