Tech pundits have a new favorite target.

Now that some of the tech bubble hysteria has subsided, and now that the economic sky has indeed not fallen, a new punching bag has emerged: The on-demand economy that isn't quite on-demand enough.

A slew of opinion pieces have appeared over the last few weeks, including articles from Bloomberg, Quartz, TechCrunch, and the New York Times, among others. Other on demand companies aren't Uber-y enough, according to the Times. And TechCrunch's Josh Constine even declared that we are in the middle of an "on-demand apocalypse"--ostensibly, because one company, Spoonrocket, shut down in the hyper-competitive food delivery industry.

Unfortunately, this sort of hot take isn't just wrong, it also obscures what's really going on. Even worse, many of these on-demand companies are repeating the exact same mistake that the largest failures of the dot-com era made. Just not in the way these articles would have you think.

This isn't to say that investing in venture-backed startups hasn't gotten overheated. In some sectors, it almost certainly has. But there are also many companies today that have built lasting businesses with real customers and real revenue.

Now, I'm going to tell you a secret: Almost all of these successful companies have something in common. In fact, it's something they also have in common with the few companies that survived the dot-com crash, like Amazon and eBay. Like those two, today's most successful startups are all platform businesses.

Rather than making and selling products or services, as traditional companies do, these platform companies simply connect people so they can exchange products, services, or information. All of the most successful companies in the on-demand economy fit into this category, like Uber, Lyft, Handy, Postmates, Instacart, and others.

Unfortunately, the on-demand companies that don't fit this model are repeating an old mistake. Spoonrocket, the company called out by TechCrunch, is one such failure. Like Spoonrocket, many of the biggest dot-com busts were really just traditional businesses with a website (or today, an app) slapped on top.

All of the biggest failures, such as, made this mistake. They assumed that if they got big enough fast enough, they would become profitable. Because, Internet.

These businesses were banking on the vague idea of first-mover advantage supported by network effects that would make the more valuable as they grew--even though they had no discernible network effects if you looked closely.

The problem was that just putting a website on top of a traditional business didn't change the fundamental economics of that business. To do that, you needed to use an entirely new business model. That business model was the platform. (I cover this shift from traditional to platform businesses in much more detail in my forthcoming book, Modern Monopolies: What It Takes to Dominate the 21st-Century Economy.)

Today, many companies rushing to get into the on-demand economy are repeating's mistake. They see the success of companies like Uber and Handy, and think that in order to do "on-demand" successfully, you simply need to deliver a service quickly through an app.

However, these companies are only superficially copying the end result without fully understanding the underlying business model that makes it work. Like, they're taking a traditional business model and assuming that a new Internet-connected interface will change everything. It doesn't.

The advantage of Uber isn't the app; it's the favorable unit economics underlying its network-based platform business model. Unlike a traditional taxi company, Uber grows through its external network rather than by purchasing and owning resources. As a result, Uber's business model allows for very low marginal cost growth at scale.

Now, this platform business model isn't a magic wand that you can wave at any business and turn it into a winner. Uber had a distinct advantage in attacking, as the Times put it, a "customer unfriendly protectionist racket that artificially inflated prices and cared little about customer service."

Where won't the "Uber for X" model work?

Well, some industries lack durable network effects, as was the case with the once red-hot Groupon. However, once Groupon's merchants realized that the customers they got through the platform didn't bring repeat business, their attraction quickly faded and Groupon's business tanked.

What looked like a network effect was actually just a price effect--consumers were chasing cheap deals, and once those went away, so too did Groupon's users. In cases like this, sometimes the unit economics of a network-based business model simply won't work.

So yes, a few of the on-demand platforms may fail. But this isn't the on-demand apocalypse. Many of these companies will succeed. And conflating different types of business models into one, monolithic on-demand economy obscures what's really going on.

Company's that are superficially trying to copy Uber or Airbnb without understanding the underlying business model and how it works in their industry are bound to fail. Or, at best, they will become niche businesses with regional reach.

The biggest threat to most of today's successful on-demand platforms isn't unit economics or too much competition. With dozens of class-action lawsuits pending over the classification of 1099 workers, it's regulation--not unsustainable business models--that may be their downfall.