As everything from tooth brushes to tractors grows more connected, the phenomenon is spreading to other industries.
The latest iterations gather mounds of data, analyse them and serve up the results, thus enabling all kinds of digital offerings, from predicting failures to giving advice.
One example of such a data platform is Nest, a subsidiary of Alphabet, Google’s holding company.
It sells devices, such as wireless thermostats and smoke detectors, which double as vehicles to collect information, allowing the company to offer tailored energy and security services.
This trend alarms non-tech firms.
They fret that if the likes of Apple and Google come to control such data platforms, these new intermediaries will seize a big chunk of their profits.
This fear has sparked a rush to build or buy such systems.
Last year car-makers Audi, BMW and Daimler acquired Here, a digital-mapping firm.
General Electric, an American conglomerate, is betting big on Predix, which helps customers improve how they run locomotives, jet engines and other gear.
Apollo Hospitals in India is creating a marketplace for health-care services.
Many more established companies are sure to follow suit, as are thousands of startups.
But before they tread this path, they should consider a few caveats.
First, most products and services are not substantial enough to make a good platform.
And even if they are, it is not always a good idea to turn them into one, says Ms Gawer, who is co-authoring a book to debunk myths about the concept.
The late Steve Jobs, for instance, long resisted opening Apple’s app store to others for fear of losing control.
Second, network effects often fizzle.
All sides of an online marketplace have to be nurtured in parallel to avoid imbalances, such as having far more sellers than buyers.
During the dotcom bubble most business-to-business marketplaces failed because their pursuit of growth led to such lopsidedness.
Even firms that had a head start, such as MySpace and Nokia, a social network and a mobile-phone maker respectively, didn’t manage to turn them- selves into fully fledged platforms.
Most successful ones are the product of specific circumstances and even chance, reckons Peter Evans of CGE.
Amazon, for example, took off in part because its customers did not have to pay sales tax if they were outside the firm’s home state, Washington.
Third, it is not always easy to make money from platforms.
Misjudge how much to charge each group of customers, and the flywheel can come to a juddering halt.
What is more, for a platform to make good money, switching to a rival has to be costly, argues Andrei Hagiu of Harvard Business School.
This risk even hangs over Uber, the fast-expanding taxi-hailing service: using a competitor is easy for both passengers and drivers.
Platforms with a leg in the physical world are likely to take more time to emerge than the purely digitalsort.
And firms have some alternatives: those with outstanding products and a strong brand can try to forgo others’ platforms, hoping that they succeed on their own.
But ultimately most firms will have no choice but to do business on somebody else’s digital property, and to agitate for better terms if the owner gets too greedy.
Call it the class struggle of platform capitalism.