Sixteen Ventures does not generally recommend technology to our clients. The technologies used by our clients should ultimately be determined by the delivery method (on-premises, on-demand/SaaS, etc.) dictated by the market, the resources they have available, etc.
However, being in the business of software, there are some high-level recommendations we can make. Since we deal mostly with Software-as-a-Service startups these days, one of the major considerations our clients must face is whether to leverage a Platform-as-a-Service or to manage the infrastructure (or Cloud Computing resources) directly. That is not a decision that can or should be rushed into.
However, it is critical to keep in mind that choosing a platform can be a "Bet-your-business decision" and you better be on the winning side.
When choosing a platform to build, deploy, and deliver the application to your clients, the following are just a few things to very carefully consider:
- Vendor Lock-in: how open is their system? what are the barriers to entry/exit? What kind of ramp-up time will my resources need? Do they leverage open technologies? Do they have proprietary programming languages or APIs that must be used?
- Sustainability: Are they well funded? What is their burn-rate/runway? Do they have other paying customers?
- Flexibility: Will the PaaS meet your needs after version 1.0 or will you need to change? Does their roadmap match yours?
- Track record: Do they have any history to learn from? Have they learned from their past mistakes? Do they provide visibility and alerts?
- Service Level Agreements: What are their guarantees since these directly affect what guarantees you can provide your clients?
- Terms of Service: Can they kick you off if they deem you competitive?
Again, this is a “bet-your-business” decision so be diligent in your research on where to host. If you choose a company that could go away soon, be sure to avoid vendor lock-in. If you aren't worried about that, you should be since it is obvious that no business is "too big to fail" anymore!
So the catalyst of this post is that according to Techcrunch, Coghead, an early player in the Platform-as-a-Service (PaaS) industry announced on February 19, 2009 that they are shutting their doors. For the past few weeks there were rumors that they were for sale, with one possible buyer being SAP (who previously invested in the company). While it is never a good thing to see a company fail, and we wish the founding team all the best on a valiant effort, it is one of the things you as a software entrepreneur must know is possible. Nobody likes to fail, or expects to fail, but if you learn from that failure, then it was at least worth something.
What can a software entrepreneur learn from the failure of Coghead? First of all, you can see that even a company that is relatively well capitalized (around $11M over the last couple of years according to Crunchbase) from strong Venture Capital and strategic investors can still fail. Money isn't everything. You can also see that just because you leverage "cloud computing" (Coghead ran on top of Amazon Web Services), you can still burn through money quickly.
The reality is we have no real visibility into what is going on internally at Coghead so other than guessing or rumors, we can't know. So the lessons learned, until more details come out publicly, must be those that apply directly to your software startup.
What this failure does is show that choosing a platform to build and run your pure-play Web or SaaS business is a decision that should not be taken lightly. The biggest problem with Coghead is that it was proprietary, so applications built on the platform are not easily portable to other environments. You might be able to extract your data, but it is true that often the context of data is at the application level. Even if no SaaS businesses were built on the Coghead platform and it was used entirely for one-off custom or departmental applications, the fact remains that those are not easily portable.
Most PaaS companies streamline the "on-boarding" process but, for obvious reasons, make the "off-boarding" process a bit more complex. Often this is not by design or for nefarious reasons, but rather it is not something they want to encourage or it is not a priority on which to focus their often limited resources. Additionally, it is difficult since, as mentioned above, often the "context" of the data is in the application so simply extracting it is not enough.
No matter what, just remember that there are many decisions that will be made during the early phases of your software startup that will help reduce failure and set the stage for success later. Be sure to check out the slides from the Anchorage "Start Smart" seminar as it outlines some of the other things you need to be aware of. For SaaS or pure-play Web businesses, choosing the platform to deploy and deliver your application is definitely one of the earlier "bet-your-business" decisions that must be made.
Author: Lincoln Murphy (@lincolnmurphy on Twitter)