What would you think about a company that owned a movie studio, a Zinc mine, an aluminum plant, a record studio, a book publisher, a mattress maker, a string of sugar plantations and a brand of clothes? Any self-respecting MBA will quickly label that as a conglomerate – and recoil from the implied poorly aligned incentives and shear overhead of such a behemoth. We bemoan the immense potential for destruction of shareholder value inherent in such a structure. Slightly older readers may even recognize this particular company as Gulf + Western, the poster child for the 1960’s and 70’s dead-end of capitalism, and literally the case study used in business schools for why conglomerates are Wrong. This model was once hailed as a superior model, and was briefly a darling of the Street. In the 1980’s most of these conglomerates were unwound, sparking the rise of the Private Equity industry. So then what should we call a company that runs a massive advertising network, the largest provider of video content in the world, a consumer electronics OEM, a designer of semiconductors, and the operator of both local and long-distance telecom networks? What about a company that is the leading provider of IT services, a large producer of original video content, one of the world’s largest air and ground freight networks, a leading retailer, and a major wholesaler of all sorts of goods? Those last two are, of course, Google and Amazon. How different are these models from the conglomerates of the past?

Of course there are massive differences. Gulf + Western and its brethren were a hodgepodge of unrelated businesses that were magically supposed to provide better returns because of their diversity. By contrast, the large Internet companies have solid ties among their disparate businesses. Google is looking to increase Internet access and improve the speed of its operations on the assumption that more people spending more time on the Internet will drive more traffic to Google’s core business. Amazon is tied together by AWS and Prime, lock in for corporate consumer customers. There is at least some strategic logic to their sprawl, and they both have enough discipline to prune underperforming businesses. Google, for their part, even breaks out its various ‘moonshot’ businesses, some of which are very far from the core business, on the assumption that they will someday spin off the successful ones and shut down the others.

Ultimately, the main difference between the model today and that of 50 years ago rests largely on that assumption. The Street still penalizes both companies for this sprawl, with Google X constantly receiving analyst criticism. However, that assumption rests on some very thin ice. The original conglomerates all started out as much narrower businesses, in a more logical set of industries. Then somewhere along the line the discipline went out the window, in part egged on by public market investors of the time.

It is worth thinking through how the older conglomerates went so wrong and got unwound. They were built on the idea that corporate managers would be best able to make tough decisions on which businesses to allocate capital to. Eventually, the Street decided that the markets would do a better job and forced their break up. But that only came after a decade of wild acquisition sprees and considerable corporate largesse. The projected returns did not materialize, but it took a decade of junk bonds and hostile takeovers to undo it all and force change.

Coordinating activity across large operations turns out to be incredibly difficulty, all the more so as the nature of their operations grow more disparate. Amazon has developed a deep set of culture tools to accommodate this range of activities, but ask anyone to explain how all the AWS services work together, or try to price a bundle of services across AWS, and the limitations become apparent immediately. Or, ask anyone with a hardware background what it is like working inside a software company for a taste of how different even these closely related businesses operate.

And then there is China. The big Internet companies there are even more diverse with Alibaba tied to banks, mutual funds, payment rails, delivery networks, retail and wholesale e-commerce, agricultural operations, and hundreds of venture investments. But we think this actually highlights the point. China operates under a very different set of economic rules. The scarce commodity there for a long time was management ability, and today it is… let’s call it Political Economics. The rules those companies operate under are very different from the rules and norms in the US, and so looking to those companies as a model is fraught.

The big Internet companies today seem to be much more wary of the worst excesses of the past, but in both Google’s and Amazon’s cases, the founders are still involved in big decisions. What happens when they pass from the scene?