Open source & enterprise go-to-market

MySQL: Building an enterprise business on a free product

MySQL was the world’s most popular open-source database. Developers everywhere used it for free. The commercial question was how to turn that adoption into a business that large companies would pay for, year after year. I led the North American enterprise sales team that answered it.

No playbook to copy

When we started, there was no proven model for selling open source to enterprises. Red Hat was the only real example, and it sold an operating system, not a database that developers downloaded for free. We had to create our own go-to-market, our own licensing model and our own sales process.

In 2005, OEM licensing (companies embedding MySQL in their own products) made up more than 60% of MySQL’s revenue. Enterprise customers were running MySQL in production, but most of them weren’t paying for it. I proposed splitting sales into OEM and Enterprise and asked to lead Enterprise, so the company had a dedicated team going after the corporate market.

The bet: attack the incumbent where it was weakest

I pushed hard for an enterprise sales team because of a hypothesis. Oracle’s soft underbelly was the lower-end workloads, especially on the web, where MySQL was a particularly good fit and Oracle’s pricing was hard to justify. If we disrupted Oracle there, at scale, it would have to respond. One way to respond was to buy MySQL.

The economics made the case. In 2004, Investor’s Business Daily reported that a MySQL license cost about $500, against a $20,000 list price for Microsoft SQL Server and about $40,000 per processor for Oracle’s high-end database. Our CEO described MySQL as the Southwest Airlines of the database world: it got customers to the same destination without the frills.

Turn adoption into recurring revenue

We moved from one-time licenses to subscriptions: first MySQL Network, then MySQL Enterprise in tiers, and unlimited enterprise agreements for the largest accounts. I owned the commercial framework around it: discount policy, enterprise agreement tiers and a deal-approval structure that let regional leaders move fast without sending every deal to the CFO.

Media, financial services, web and government customers, including The New York Times, Dow Jones and NBC Universal, standardized on paid subscriptions.

Sell where the incumbent was too expensive

The sales team turned that hypothesis into a qualification model. MySQL won production workloads where Oracle, Sybase or Microsoft licensing was hard to justify, and we went looking for that pain directly: customers that wanted to reduce their dependence on an expensive proprietary vendor, add capacity next to an existing Oracle environment or cut data warehouse license costs. Total cost of ownership and migration services gave buyers a concrete reason to move.

Open new routes to market

I built alliances with HP, Dell and Ingram Micro, Novell and Unisys, and opened the federal market through Carahsoft. Federal required compliance work as well as sales coverage, including a DISA security review and Common Criteria certification.

The lesson: differentiation comes first

When people can use your product for free, you have to give them compelling reasons to pay.

An open-source subscription business works only when the paid product is meaningfully different from the free one. We beat our numbers, but the ceiling we kept hitting wasn’t the sales team. It was how clearly the paid product stood apart from the free one.

Sales can’t create that difference alone. I pushed product management and marketing for enterprise-only capabilities, monitoring and management tools, partner programs that required the enterprise edition, and a faster release cadence. Since then, I’ve treated working with product and marketing to build differentiation, and to tell that story to the market, as a core part of the revenue job.

The bet paid off

The hypothesis proved true. As MySQL moved into paid production workloads, Oracle responded. It bought InnoDB in 2005 and Berkeley DB in 2006, two storage engines MySQL depended on, and it nearly bought MySQL itself. Sun Microsystems bought MySQL first, for about $1 billion, in 2008. Oracle got MySQL anyway when it acquired Sun in 2010.

After the acquisition

Inside Sun, I led MySQL sales for the Americas, covering North America, Canada and Latin America. Much of that work was helping Sun’s field sales organization sell an open-source subscription: the pitch, qualification rules, account targeting from download activity, and enablement for Sun’s reps. I stayed through Oracle’s purchase of Sun.