I am at heart a technologist. I still get excited hearing about new, cool software technology and solutions. All of my past experience, and now at eXeed, my time goes into figuring how to mold new technology so that it solves a real problem for a large set of potential customers. At IBM Research our group used to use the term “Research in the Marketplace”. Continuing that tradition at eXeed, we work pretty much only with startups that have unique technology, so my comments are very much focused on technology based startups.
In my previous post I tried to show why the numbers make it hard for corporations to create “startups” internally. Beyond the numbers, there are two basic truths about technology based startups that make it very difficult to cultivate them in a large corporate environment:
1. Over a short period of time technology-based startup products need to morph (sometimes repeatedly) to meet customer\market needs –experimenting in the marketplace while leveraging the same basic core technology. Very few technology startups start with the same set of product (ideas) and business model as they exit with – but in most cases the new direction still leverages the same core technology.
2. Startups start out small, with very little (if any) revenue. One thing to remember is that even an exponential curve starts out slow – and only speeds up as time goes on.
These two “facts” make it hard for a startup to be incubated in a large company – there are just too many barriers to being nimble enough to make point 1 work, and point 2 usually means that the corporation looses interest before success can be shown (especially given the ambiguity associated with point 1, and since there is usually an existing set of products that can generate more short term revenue with the same investment).
I guess the basic point is that the impedance between running an existing business and starting a new one is so great that unless the corporation is willing to experiment with new, unorthodox models, and can commit long-term high-level executive support, it just won’t work. Even then there is still the issue of getting the right people (and giving them a loose enough rein) – but I will discuss that in another post.