[nostrademons]

It's relevant to startups, though, because picking a vacant market where the complements are already commoditized is a winning strategy for a startup. It was like that for most of the tech giants of today. For Amazon, books were already a commodity, shipping was a commodity, and online checkout became a commodity shortly after they started. For Google, websites and hyperlinks were a commodity. For Apple, chips were a commodity, TVs were a commodity, floppy disks were a commodity, and programmers who knew BASIC were a commodity. For Uber, drivers were a commodity, cars were a commodity, and cell phones became a commodity.

The main difference between startups and big companies is the resources available to them. Startups do not have the ability to meddle in markets other than their own product, because they simply don't have enough people. They have to put all their efforts into their main product, because they don't have much effort to spare. But the most successful startups are the ones where the rest of the ecosystem already exists and they're just putting the keystone in place to change how society functions.

[wunderflix]

The way I understand complement like it’s defined in the article is: something you buy from different players to use with your product. Cars/gas, Hardware/SW.

It’s not something you need to produce your product in the first place, like steel for car manufacturers. Thus, I don’t think your examples qualify tbh.

[nostrademons]

Complements are essentially the "ecosystem" around your product. In a traditional microeconomics text, they're the goods that are commonly purchased alongside your product. So PCs, monitors, keyboards, and mice are all complements.

Re-reading my comment, I see some of the confusion. If you think of each of these companies as a reseller of books/websites/computers/rides, then those goods are not their complements; they're their suppliers. If you think of them as a marketplace that provides a service to find books/websites/[okay Apple makes no sense here]/rides and then takes a cut for their value-add, then they are complements: the driver and the service to find the driver are separate goods that must be booked together. There are substantial legal battles around which is a better model for the tech industry: Elizabeth Warren's campaign to prevent tech companies from also owning services that compete on their platform assumes they are marketplaces, the employee/1099 classification lawsuit assumes that they are resellers, the Australian law to force Google to pay publishers assumes they are resellers.

I could probably have picked clearer examples, eg. cheap TVs and floppy disks really are complements for 1975 Apple but chips are suppliers, while Amazon's own business plan implies that it's a retailer rather than a marketplace. Google is still a pretty good example, though: Google does not sell websites (or didn't in the 1990s, at least), but the need for a search engine arose out of there being many websites around.

[wbl]

Books are not a commodity. They are one of the most differentiated products out there, with little possible substitution among titles.

[nostrademons]

"Books" as a mass noun is a commodity, "books" as a plural noun is differentiated. (Somewhat by definition: anything that you think of as a mass noun is a commodity.) Amazon's goal was to get you to think of "books" as a mass noun, and among the general public, they largely succeeded. Book aficionados still think in terms of books as individual, differentiated titles, but these people are much more likely to go to small independent bookstores (where they will get individual, differentiated attention) than Amazon. The general public thinks "If I happen to need a book, I'll find it on Amazon".