There’s plenty of talk in investment circles about the startup sector, but no singular, clear definition of what a ‘startup’ actually is.
There’s plenty of talk in investment circles about the startup sector, but no singular, clear definition of what a ‘startup’ actually is.
Determining whether or not a business is a startup is something everyone involved in the startup scene faces at one time or another; and there isn’t a definitive answer.
Everyone has their own subjective opinion, and ends up at the ‘I know it when I see it’ definition, which is about the best we can do. And while I’m not going to try to provide a one-size-fits-all definition of a startup in this article, hopefully I can provide some useful framing context for the exercise.
The origin of the word ‘startup’ dates back to a 1976 Forbes article relating to investment in new companies in the “electronic data processing” field. Since then it has been closely associated with tech, with ‘startup’ and ‘tech startup’ being synonyms for a large proportion of people in the industry.
However those who are most associated with the term disagree that it is exclusively tech-oriented. Steve Blank, arguably the father of the ‘lean’ business development methodology, defines a startup as: “An organisation formed to search for a repeatable and scalable business model”.
Eric Ries, the author of Lean Startup, defines it thus: “A startup is a human institution designed to deliver a new product or service under conditions of extreme uncertainty.”
Paul Graham, founder of Y Combinator, the archetypal startup accelerator, defines it as: “A startup is a company designed to grow fast.”
Clearly, none of these definitions ties the term to tech, so in principle any business can be a startup, as long as it doesn’t have a defined business model, operates under extreme uncertainty, and is designed to grow fast.
The startup revolution is enabled by cloud technology. In the dot.com era, each website had a physical server powering it. That server may have been located in a server farm or co-location site close to an internet access point, but it was a physical asset of the organisation. Scaling a website for more customers meant buying more servers, and finding space for them, and building expensive load balancing network infrastructure to keep them from being overloaded.
All of that vanished with the advent of cloud computing. Now it costs as little as $5/month to operate a web server on the internet. If the load on that server rises, adding more servers is as simple as clicking a few buttons. Adding a load balancer is trivial. The cost and complexity of creating web infrastructure has entirely disappeared.
Critically, the cash flow problems associated with scaling a web-based business have also vanished; the bill for the infrastructure now arrives at the end of the month in which the revenue occurred, instead of three months before the revenue. Sudden, dramatic growth brings massive rewards instead of massive cash flow problems. Scaling fast became a good thing instead of a problem.
The ‘lean methodology’ is a direct result of this cost reduction. It is based on the principle of ‘build, measure, learn’, in other words: build your product, then measure customer response, and adapt your product to the response.
Because it is so cheap to build websites, it becomes easier, quicker and cheaper to perform market research experiments on your actual customers than on focus groups. Lean startups typically conduct multiple experiments simultaneously and update their systems many times per day.
This leads to launching web products much earlier in their development cycle; in fact so early in their development cycle that the organisation can do it before they’re certain of their revenue source, their customer preferences, their marketing channels, or any part of their business model.
Being able to iterate a product while the customer is using it is a key feature of a startup, and there are some organisations doing this for physical products. For example Tesla implements new features for their cars via software updates. But clearly there are limitations (it can’t replace the tyres over the internet).
Anyone actually making physical product bumps into these practical realities of production, distribution, retail channels, and supply chain management. These are all very well understood business problems that have defined solutions that are hard to optimise further, and so provide a large amount of certainty in the business.
The product must also be well researched and well designed before it ever hits the assembly line, or the business will face problems with obsolete inventory, again providing certainty. It’s hard to see how this sort of business meets the criteria for a startup.
Service-based businesses are less constrained by this.
Meanwhile, there is a growing trend in small exploratory product businesses that can be considered startup-like. These businesses outsource everything they possibly can, utilising an array of offshore freelancers and fulfilment services to provide a product to a market with minimal investment.
The result is similar to cloud computing for physical businesses; the cost of launching a new product is brought down, the cash flow problems of scaling are removed, and the business can experiment using small-batch production runs to iterate on product designs in the marketplace.
The approach of launching early, iterating fast and prioritising customer feedback to power growth is the ‘lean’ approach, and ‘lean’ is the methodology devised by startups to adapt to the business possibilities of cloud computing.
‘Lean’ businesses are typically unsure of their business model, operating in extreme uncertainty, and aiming at massive growth. With this in mind, it could be said that a startup is a business operating using the ‘lean’ methodology.