Read time: 6–8 min
Most people use the word 'startup' to mean any new small business. That is not what it means. Understanding the actual distinction — not just for the label, but for the strategy it demands — is one of the most important decisions you will make as a founder.
A startup is not a smaller version of a big company. It is a different organism, designed to do a different thing.
The most useful definition of a startup comes from Steve Blank, the Silicon Valley entrepreneur and Stanford lecturer who arguably did more than anyone to formalize how startups are built. His definition is this: a startup is a temporary organization in search of a repeatable and scalable business model.
Every word is doing work.
Temporary — because the goal is not to stay a startup forever. You are searching for something. Once you find it, you become something else: a company, a scale-up, or a failure. The moment you mistake the search for the destination, you start optimizing the wrong things.
In search of — because until you have found the repeatable, scalable model, you are still searching. You do not yet know what works. That is not a weakness. That is the accurate description of your situation. Acting like you already know, before you do, is where most early ventures go wrong.
Repeatable — because the model must work consistently, not just once. One customer does not prove a model. One great quarter does not prove a model. Repeatability is what separates a business from a lucky event.
Scalable — because the cost of growth must not grow proportionally with revenue. This is the crucial technical criterion. A business where adding ten customers requires hiring ten more staff is growing. But it is not scaling. Scaling means the denominator gets cheaper as the numerator grows.
Small and medium enterprises are measured by size: revenue, headcount, or asset value depending on the country. In Thailand, OSMEP classifies micro businesses at under 10 employees, small under 50, medium under 200. But the more important distinction is not size. It is the growth model.
An SME is typically optimizing an existing model. It knows its customers. It knows its margins and channels. The goal is operational excellence and sustainable profitability within a defined market. That is not a lesser ambition — some of the most valuable businesses in the world are structured as SMEs. A great family restaurant, a thriving consulting firm, a regional logistics company with strong unit economics — these are legitimate, valuable, and often more stable than most startups.
But the strategy, the funding model, the hiring logic, and the risk profile are completely different from a startup in search mode. The mistake founders make is sitting at the wrong table and applying the wrong playbook.
Two Tables at the Same Café
Table 1 — The startup: people furiously testing hypotheses, burning cash to find the model, pivoting when the evidence demands it.
Table 2 — The SME: a well-run operation managing known costs against known revenue, optimizing for margin and longevity.
Both are legitimate. Both require real skill. But if you sit at the wrong table and apply the wrong strategy, nothing will work — and you will not understand why.
These are not mutually exclusive. A company can be an SME by size and operating a startup model at the same time. WHOOP — the fitness wearable company — was a micro-sized company for most of its first three years. It had almost no revenue. It was definitely not a large enterprise. By OSMEP definitions, it was a micro-SME. But it was pursuing a startup model from day one: searching for a scalable, repeatable model that could grow without proportionally growing costs.
The pivot that changed everything for WHOOP was moving from selling hardware to giving the hardware away and charging a monthly subscription. Suddenly the model became exponential. Adding a new subscriber cost near-zero. Revenue could compound without headcount growing proportionally. That is when the startup model clicked into place — even though the company was still small by any size measure.
Conversely, a large company with hundreds of employees and millions in revenue can still be operating an SME model if its growth is linear — proportional to investment. Size and model are different things.
SMEs typically grow linearly. You open another location, hire another team, expand to a new product line. Revenue grows in proportion to investment. That is predictable and often highly profitable.
Startups are designed for exponential growth — the kind where adding one more unit of cost produces dramatically more than one unit of revenue. This is only possible with a business model that is structurally different: a platform that becomes more valuable as more people use it, a software product that can be replicated at near-zero marginal cost, a subscription model that compounds over time, a marketplace that grows on network effects.
The exponential growth model is what makes venture capital make sense. A VC fund invests in ten companies expecting nine to fail and one to return the fund. That math only works if the one that succeeds can grow fast enough to compensate for everything else. That requires a startup model, not an SME model.
The exponential model is what makes VC math work. But VC is a signal of startup intent — not the definition of it. You can be bootstrapped and still be a startup. You can be VC-backed and still be an SME in disguise.
No. And this is where many engineering and computer science students make a category error: assuming that because they can build technology, they should build technology into every venture, or that without technology there is no startup.
What makes a venture a startup is not technology. It is the growth model. Consider three examples:
Lemonilo — Indonesia
An Indonesian health food brand built around instant noodles made from natural ingredients. Physical product. Conventional manufacturing. But the company built a platform model around health-conscious consumers, a subscription layer, and D2C distribution that scaled independently of store count. The noodle is not the startup. The model is.
Oatly — Sweden
Oat milk. Deep food science (their oat base process is patented), but the product is a commodity beverage. What made it a startup was a deliberate distribution strategy — targeting specialty coffee shops first to build credibility, then using that credibility as leverage into retail at scale. A physical product with an exponential distribution architecture.
WHOOP — United States
A fitness wearable. The watch itself is not the startup. The subscription model, the data platform, and the continuous health intelligence layer built on top of the hardware are. The hardware became a delivery mechanism for the subscription. The unit economics only work because the subscription compounds. Strip the subscription model and you have a well-designed gadget business — an SME, not a startup.
For anyone with a technical background: your skills are a real advantage when they let you build something that could not be built otherwise — a proprietary process, a unique data architecture, an automation layer that fundamentally changes unit economics. They are less of an advantage when the technical complexity becomes the product instead of the enabler.
The question to ask yourself: is the technology the moat, or is the model the moat? The answer shapes everything about what you build and how.
✦ A Note for GVP Students
When you categorize your venture for Block D: be honest, not aspirational. Most ventures built in nine weeks are closer to SME territory — and that is completely fine. The question is whether you are designing a model that could eventually be repeatable and scalable, or one that will always require your direct involvement to operate.
Both are valid paths. Only one is a startup. Know which one you are building — and why.
.png)