The WeWork drama of the past weeks is a great reminder of how perceptions of a start-up opportunity can change very dramatically in a short period of time. What was one of the highest flying start-ups with a private market valuation of $47B was brought down to Earth when trying to IPO. (Important to note that building a $10B – $15B start-up, the rumored valuation if the company went public today, is still an incredible achievement – even if it’s only about a third of the market cap a few months ago).

I think WeWork provides a key lesson about company building. The major point that often gets forgotten is that no business model is the same.

If you run a software start-up, gross margins are high, scaling is relatively easy, and capital requirements are relatively low. But if you run a tech-enabled service business, gross margins tend to be low, scaling is hard (you’re moving atoms instead of bytes), and capital requirements can be high (especially if you take on leases like WeWork or inventory like OpenDoor). Vertically integrated commerce or hardware companies fall into the same category of low-margin and hard-to-scale businesses.

I think that you can still build amazing companies in all of these categories, but you need to adjust your playbook. Amazon is a great example in many ways:

• Given the low gross margins in retail, Jeff Bezos focused from the beginning on building an incredibly frugal company culture. Even many years into the company, most VP’s still took the train to the airport when travelling. Frugality isn’t something you can switch on later; it needs to be there from the beginning.
• Early on, Amazon focused on amazing execution. While it’s true that no start-up can win if they fail on execution, software mistakes can usually be addressed pretty quickly, but mistakes involving “offline delivery” are much harder to fix.
• Amazon dealt with the capital-intense nature of their business by leveraging their negative cash flow cycle (they got money from customers before paying out suppliers). Later, they launched their marketplace business so that they didn’t have to hold inventory for every product by outsourcing this to suppliers.

Great start-ups can be built in many categories and with many different business models. But copying the playbook for software start-ups when building a start-up that doesn’t have the same attributes often leads to failure. The one-size-fits-all approach to company building doesn’t work…even if everyone is a technology company these days.