I want to re-define it. Chris Dixon points to this post, which defines “scaling” as:

dealing efficiently with events that occur with a predictable frequency.

This refers to the type of scale that is described in freshman economics classes. It unpacks the law of large numbers, and has important implications for the way computer scientists design systems. By that definition, a scalable business is one that can be organized in such a way that when its purpose is repeated with increasing frequency and scale, things don’t break.

Paul Graham defines a startup as:

a company designed to grow fast.

By this definition, the question of whether or not a company is scalable makes a lot of sense. It refers to a product’s natural capacity to get a lot of users very quickly. The successful start-up is the one that can make something lots of people want, and reach and serve all of those people. And “scalability” is the litmus test for that success.

But “can get to a lot of people” is the wrong litmus. Facebook, for example, makes something lots of people want, and it reaches and serves all of those people. And it got there without its core technology breaking. It is scalable, and it is “at scale’. But the Facebook experience is broken for me. It broke somewhere in between the Dunbar number (150 friends) and my current number (3,000 friends). I wanted a way to keep in touch with people I love, and to share things that are important to me with those people. But "at scale”, Facebook actually stops working. Flooding the system did not break the technology, but it broke the experience.

Scalable should be defined in different terms, then. If a startup is “Scalable”, as I see it, the optimal user experience is the “at scale” experience. Well-designed cities are scalable. When they achieve maximum density, the services rendered work great – you can walk and bike to places, you can interact with a diverse set of people, and have a wide range of activities available to you. Your immediate radius becomes a hub of intellectual and creative capacity. A dense city optimizes the experience. A dense Facebook, however, breaks the experience. Scalable, in this sense, not only means that the system can grow fast without breaking, but that a system will work well, and even better, once it’s grown.

Marketplaces tend to be more naturally scalable, by this definition. At scale, the buyers and the sellers both get happier as there are more of the other, and there is a virtuous cycle that drives growth at liquidity. Getting to scale requires delicate knob-twisting of supply and demand AND user momentum. It’s a labor of love, and it requires a clear and constant feedback loop between the product designers and the users. Entrepreneurs and investors both tend to focus much more on the momentum aspect of scalability: can it GROW. This makes sense, but is incomplete. To create lasting businesses, equal attention must be paid to the question of what happens if it grows?