Over the years, I’ve been asked many times about the common elements across the companies I’ve worked in. Borland, Active Software, MySQL, Zendesk, Duo Security, were all quite different. However there were several factors they had in common which might be helpful for those thinking of joining or creating a new startup company.
Each of these companies helped usher in a new technology wave that enabled growth across the industry. Borland was part of the first stage of the PC software industry and later helped usher in the Windows era as well as Client/Server computing. MySQL helped drive the acceptance of open source software for the web. Zendesk was one of the early SaaS cloud vendors. Duo security fueled the growth of B2B smartphone applications to make the internet more secure.
At the time, these bets were not at all obvious. There had been open source products before MySQL and SaaS companies before Zendesk. But these were by no means proven business models. Still, if you squinted right, you could see a future that would continue to grow. Nonetheless, there was considerable risk, and many peer companies fell by the wayside.
Part of what made the future interesting for these companies was the size of the total available market (TAM). In the early 2000s, the database industry was a stable and mature business, generating around $9 billion in revenue for Oracle, IBM, Microsoft, Sybase and a handful of others. I joined MySQL when the company was doing about $5m in ARR and there had been no new breakout database business in the prior ten years. As I did my research on MySQL it became clear that there was a large market, and that customers were deliriously happy. The same pattern held true at Zendesk and Duo Security. The final piece of the puzzle that made these companies successful was that the strategy was disruptive.
The term disruption, gets thrown around a lot by founders and investors to the point of it becoming synonymous with “new, cool stuff.” But when I speak of disruption, I am holding to the definition by Clayton Christensen in his book The Innovator’s Dilemma.
Many successful businesses have been built without fitting in to Christensen’s theory of disruption. But if you are able to tap into its power, disruption is like having the wind at your back in a marathon. It’s still a marathon, but you’ve got some staying power.
Christensen distinguishes between sustaining innovation and disruptive innovation. Sustaining innovation (such as incremental improvements in a product’s performance or price) favors the incumbent catering to their existing customers. Startups generally should avoid pursuing sustaining innovation as these improvements are easy for incumbents to adopt, even if it takes them some time.
If your new, cool, VC-backed company creates a product that is twenty or even thirty percent better (faster, cheaper, easier) than the big guys, it’s going to be extremely difficult to lure customers to a new, unproven startup. Customers who buy from large vendors often think: “we’ll just wait a year or two, and we’ll get those features, too.”
The Rules of Disruption
Christensen defines disruptive innovation as a radical change in the market that targets a new audience with different buying criteria. They aren’t seeking something “just a little better.” They are representative of a new emerging market. There are four key elements of disruption:
1. There must be a large existing market
2. There are existing incumbent vendors
3. A new entrant plays by different rules to serve an underserved market
4. The incumbents are unable or unwilling to adopt the disruptive strategy
The third item is really the trickiest. But let’s examine this further using MySQL as an example.
1. $30b market… check!
2. Large incumbents included Oracle, IBM, MSFT… check!
3. New entrant with different rules… MySQL is open source… check!
4. The incumbents won’t open source their products or offer them free… check!
Let’s drill further into MySQL’s strategy and what was going on at that time. During the so-called “database wars” of the late 1980s and early 1990s vendors like Oracle, Informix, Sybase, IBM were all focused on TPC-C benchmark performance. Each was trying to outdo the other with new features, faster disk drives, bigger processors. This was truly an era of sustaining innovation with vendors launching advanced features, more tuning options and faster benchmark scores. The result of many years of sustained improvement was that databases had become increasingly complex. Specialist DBAs were required to tune databases in order to ensure fast performance.
Finding the Underserved Market
MySQL was not the first open source database, but it was the first one that catered to web developers. In this early period of web development, the dominant tools were HTML editors and scripting languages such as PhP, Perl and Python. As web developers sought to create more dynamic sites, they looked to use a database. MySQL became the default database for web development by virtue of three factors:
It was easy to learn
It integrated well with popular web development languages
It was free when used under an open source license
So while the incumbents were fighting over a massive Enterprise market that focused on advanced features and performance, MySQL was being used by people who would never have bought Oracle, SQL Server or IBM DB2. To web developers, the existing database products were:
Too hard to learn
Required management approval
The very features that made Oracle and others attractive to Enterprise customers made them overkill for newcomers. Web developers were seeking a “good enough” solution for building dynamic web sites and applications. They didn’t need or understand most of Oracle’s features.
In effect, MySQL snuck in the side door by focusing on an under-the-radar market. The big vendors targeted major enterprise applications in finance, manufacturing, retail and ERP systems. They didn’t care about web developers. That market barely existed and certainly wasn’t large enough to attract their high-priced sales teams.
When the incumbents looked at MySQL, they dismissed it as a toy. It was a primitive database with just a fraction of the features of their products. Worse yet, MySQL users were webmasters and web developers who worked for, ugh, startups whose budgets were a fraction of those found in the Fortune 1000.
The Growth of a New Market
Over a few years, that tiny, underserved market grew by leaps and bounds, bringing a far bigger opportunity for MySQL and other web technology vendors. MySQL was part of the LAMP stack (Linux, Apache, MySQL, PhP/Perl/Python) an alternative to traditional development stacks offered by Microsoft and others. The LAMP stack became the default development approach for the Web 2.0 era.
To be clear, the LAMP stack was a marriage of convenience, more of a clever acronym than an actual software architecture. The pieces worked together, but it took more work than Microsoft’s fully integrated tools. Nonetheless, the LAMP stack provided freedom to developers that wanted to avoid vendor lock-in. With Microsoft, Oracle and others, the more technology you used, the more you had to pay in annual server license and maintenance fees. The LAMP stack enabled companies to develop a scaling strategy that eliminated the “success tax” as they grew.
Google, Facebook, Wordpress, YouTube, Wikipedia, all were built on the LAMP stack, as were thousands of startups over the next 15 years. The LAMP stack became the de facto standard for building web sites and we-based companies.
Imitation Is The Most Sincere Form of Plagiarism
Part of the reason MySQL had such unprecedented growth was that the business model was hard for the incumbent leaders to replicate. We routinely promoted MySQL’s Enterprise commercial offering as being 90% cheaper than Oracle. Not surprisingly, that wasn’t a price Oracle could match, lest it decimate its growth. MySQL’s open source business model was part and parcel of its disruptive strategy.
Oracle argued that its products were more sophisticated and more fully-featured, which was absolutely true. MySQL wasn’t trying to match Oracle features pound-for-pound. MySQL was targeting users whose needs were more modest. In that regard, both products were a good fit for their respective audiences.
As often happens with disruptive innovation, at some point, the incumbents imitate the disruptor. Oracle announced a stripped-down free product called Oracle Express, which had a subset of features and limited storage capacity. Most users looked at Oracle Express and realized that because of its limitations, it was effectively “cripple-ware” and they’d have to upgrade to the regular commercial product as their data sets grew. For developers this was a non-starter. After all, who wants to swap out databases as you grow?
Microsoft, IBM, Sybase all embarked on similar strategies announcing their own creatively named free products: SQL Server Express, DB2 Express, Sybase Express. All failed to make any mark in the industry and MySQL’s growth was unimpeded. Nonetheless, the moved helped validate the market by showing that what MySQL was doing was worth copying.
The Billion Dollar Question
MySQL continued to embrace the disruptive strategy, adopting a subscription model and integrating with new emerging programming languages and tools. While there was often discussion about limiting the features of the open source product, we knew that we could not walk away from the open source model without losing our disruptive mojo. There were still plenty of challenges, but the open source approach made us unique.
We grew the company to around $100m in recurring revenue before we were approached by Sun Microsystems with a billion dollar acquisition offer. Oracle also made overtures including an 11th hour bid to match Sun’s offer.
Working at Sun provided me further insights into the power of disruption. I got an inside view of how an early innovator like Sun could be devastated by disruptive technology, in this case the commoditization of the server industry through a combination of Windows, Linux and Intel processors. Sun, like most companies facing disruption, predictably and rationally focused on their biggest and most valuable Enterprise customers, rather than embracing new emerging technology markets. Less than two years after MySQL was acquired, Sun was acquired by Oracle.
While MySQL had a somewhat contentious relationship with Oracle, we always respected them and felt that they understood the power of disruption. Oracle has remained a remarkably good steward of MySQL.
So what can you take from all of this?
Silicon Valley founders and investors are great at hyping technologies. But it is still fairly rare for a company to fit Christensen’s definition of disruption. Nonetheless, the theory of disruption is a useful model for evaluating new technologies and understanding whether they are truly disruptive or simply “new and cool.” If you can find a technology that is opening up new markets, the disruption model can be a guide for making decisions to keep you on the right path. But be careful of falling into the trap of thinking that making something slightly better, faster, cheaper is necessarily disruptive.
I hope this post provided some clarity around how to think about disruption. In a follow-up post, I provide practical information on how maximize the impact of a disruptive strategy.