What's a tech company?
Is a company a tech company?
The first thing we need to understand is that it’s not a binary classification where some companies are tech companies and some aren’t. It’s a continuum, a shade of grey. Two companies can both be tech companies but one more than the other. Just as two people can be tall, but one taller.
I’ll share a bunch of criteria below. Every tech company need not meet every criterion. Some of these indicators are probabilistic. For example, one indicator is that it’s VC-funded. It doesn’t mean that every VC-funded company is a tech company, nor that every non-VC-funded company is a non-tech company. It’s like saying men are taller than women: it’s true on average, not for every single case. So, when I say “A company is a tech company if…”, you should interpret it as “A company is more likely to be a tech company if…”
With that context, a company is a tech company…
… if they’re selling software, like Figma. Bits, not atoms1. As counter-examples, Flipkart sells physical products, and Spotify sells music.
… if you can sign up in minutes and start using it. This means self-serve, like creating a Notion account. A counter-example is WeWork asking me to talk to a salesperson.
… if a global audience uses it, like Notion. The same product / service needs to be offered globally, unlike washing machines that have different features in different countries. And people the world over need to access it using the same website. E-commerce companies fail this criterion because Indians access them at whatever.in and Americans at whatever.com.
… if they’re offering users a way to achieve their goals which wasn’t offered before. Example: Uber letting you call a cab via an app, rather than standing on the road or calling on phone. Other examples are Figma and the iPhone circa 2008.
… if they have a low marginal cost2. There are two kinds of costs in any business: fixed costs, which don’t scale much with the number of users, like R&D. And marginal costs, which scale a lot with each other. In a tech company, the balance between fixed and marginal costs tilts more towards fixed. The cost to add an additional user is very low. For example, building Figma required tens of millions of dollars of investment, but once the product has been built, it costs Figma little when you sign up. On the other hand, Toyota incurs a significant cost when you buy a car.
… if it offers a free trial or a forever free plan (freemium). A barbershop doesn’t offer the first haircut free.
… if the product / service is standardised rather than customised. Any product / service that’s sold falls in two categories: standardised, where millions of people use essentially the same product, or customised, which is built bespoke for the buyer. Tech companies offer standardised products/services3.
… if it’s based on deep tech, research that goes for on years, like the original iPhone.
… if the cost of the product or service is half or less of that of existing players. Notice the word “sustainably” — giving it away below cost by burning investor money doesn’t count. Nor does sacrificing quality or support. True sustainable cost reduction is possible only via technological advancement. For example, SpaceX’s Falcon Heavy is priced at $2K vs $5K+ for alternatives like ISRO and Soyuz, both per kg.
… if it has attracted venture capital. This is a stronger indicator in the negative: non-tech companies typically don’t attract VC.
… if it has a significant chance of complete failure, like SpinLaunch. A counter-example is setting up a new barbershop: you’re going to cut at least some people’s hair.
… if it’s growing fast at scale, sustainably. Again, we’re not talking about giving a product away below cost, or spending more in CAC than LTV. But genuine fast growth beyond a certain scale is an indicator of a tech company. Non-tech companies grow slower4.
… if they’re starting from the tech and adopting a first-principles approach by asking, “What can the technology do?” and building up from that foundation. Cryptocurrencies are the perfect example. By contrast, non-tech companies start with a business model and then ask, “How can we make the same business model run better with tech?” like State Bank of India using software to work better.
… if what they’re doing is illegal initially, but gets legalised after a decade. For example, Uber. As another example, VoIP was originally illegal in India, but legal now.
… if it’s a newly formed company. Companies become rigid and set in their ways as they age and eventually can’t do new things.
… if a lot of the company’s value derives from IP rather than other factors like brand or real estate.
Remember that all these are indicators, not black and white criteria. I’m not saying every newly formed company is a tech company — you can set up a new massage parlour tomorrow — but that the chance is lower for a 40-year-old company to be a tech company.
Just offering content like a band selling songs doesn’t qualify.
Excluding CAC.
Low marginal cost means high net margin.
Like everything else, this is a shade of grey, not black and white. When we say standardised, an example is a Macbook. Now, Apple does offer some customisations, but it’s still a standardised product: I can’t ask for a 17-inch screen, or a square screen. I can’t ask for four USB ports. I can’t buy a Macbook with a big screen but the more power-efficient and cheaper processor that comes in smaller screen Macbooks. I can’t ask for a microphone jack. Or a USB-A port. Considering all this, the Macbook can be classified as standardised. Or more standardised than customised. To be precise, like everything else in this blog post, standardised and customised are a continuum, not black and white. So when I say that tech companies offer a standardised product, you should interpret it as “tech companies offer a product that’s closer to the standardised end of the spectrum than to the customised end”.
Non-tech companies can grow quickly at a low scale, like Rameshwaram restaurant in Indirangar, which isn’t serving a million customers.