Product First, Platform Second

The debate about investing in startups that rely on cloud infrastructure has long raged within the venture community. Investors are rightfully cautious about backing companies that have a large dependency on another entity that could control a startup’s economics. That’s because platform providers have long been known to encourage startups to build on their platforms, and then once a strong ecosystem is established, to take the majority of the economics out of that ecosystem. That’s not to say that entrepreneurs can’t make a lot of money building applications on other platforms — they can. The question is whether investors can.

• For a startup to be successful building on top of a platform, there are a few considerations. If that functionality is close to the platform, the startup has to believe it can stay ahead of the platform integration expansion curve. This has been done but it is a constant treadmill — think Rational / Microsoft for development tools.
• Another option is to look at verticals, especially verticals that are overdue for disruption or that are highly fragmented. While these can be successful, they are difficult to execute — there’s a reason verticals stay fragmented and it often more to do with channel establishment than technology adoption.
• A startup can build on top of a platform where the application or applications truly are different than the core platform functionality.
• Finally, a startup can build an application that gets deployed on multiple platforms but potentially launches on one platform first. This is difficult in a world of many, many platforms (think client based applications for mobile phones before the iPhone). But in a world of two or three platforms, it is not a bad bet.

The other risk with playing too close to the platform is that over time platforms tend to integrate much of the adjacent functionality delivered by ecosystem vendors into the core platform. In many cases, while the size of a participant’s business while large by itself is relatively small in the context of the platform provider’s revenue total stream for platform related revenue, platforms by their nature add core functionality by going after the next closest piece to the platform.

Risk Mitigation
For those evaluating the risk of investing in companies building on a particular platform, rather than investing in a company that some day has the potential to be a platform in and of itself, the question is whether they can invest in the company and have it get to an interesting size so that it can have a near equal relationship with the platform. By interesting size, I mean, on a path to go public or to actually go public as a tool to have the public markets support its market position and valuation.

While Amazon may change their business model for Web Services in the future, all indications so far point to delivering a great platform for its partners by competing on operational efficiency at scale – as signaled by the recent price decrease. And Amazon is well positioned to deliver on that promise given the scale of the operation it runs.

Other platforms tend to have an innate desire to deliver not only platform functionality upon which applications can be built, but to involve themselves in distribution as well. It goes without saying that platforms, by their very nature, need applications to be successful. A platform without applications built on top of it is much like a tree falling in the forest with no one to hear it. To get these applications developed, platforms build their own applications (Office on Windows) to seed the platform, and then work to get adoption of the platform by others. They help the early ecosystem partners not only on technology but also with distribution, that is, customer acquisition.

Getting leverage on distribution cost is critical to the sucess of a venture backed company. And a platform can be a very appealing partner in getting that distribution, at the beginning. In most cases, in fact, it is simply too good an opportunity to pass up. The risk to the startup comes later when the platform provider has the option of changing the economics of the relationship. Or when the platform provider integrates external functionality that is too close to the platform.

Amazon, one of several in the platform game, is doing an excellent job signaling that it is truly a platform upon which it wants others to build. AWS appears to be about easiest access, low cost utility computing first, and distribution control a distant second, if at all. That could change, of course, but it is unlikely because it would undermine the core value proposition of AWS as a cloud platform.

That’s good news for startups and venture investors alike.

Conclusion
It is typically a bad idea to set out to build a platform company from the get-go. That’s because people buy and use products, not platforms. But those companies that are huge do ultimately become platforms. Salesforce.com was a CRM application first. Now Force.com is endeavoring to become a platform. Amazon was a book-seller first, now it’s a platform. Facebook was an application first, now it too is trying to become a platform. What is common across all these platforms and platforms-to be are that they were products first, platforms second. What will be different is the extent to which new application companies come to rely on them for distribution.

As I wrote some time ago, I’d rather have a company that ultimately defines the ecosystem and owns the customer (or user) than a company that is part of the ecosystem. I’m certainly not against investing in companies that are not ultimately platforms — many valuables ones have and will continue to be built. And by the nature of platform success, you’re a product before you’re a platform.

But those that are able to obtain product success and then make the transition to platform have proven time and again to be most valuable.