Playing To Win

Platform Business Taxonomy

The Spectrum and its Implications for Strategy

Roger Martin
8 min readMar 13, 2023


Source: Roger L Martin, 2023

I can’t help it. When I get lots of interesting feedback on a piece, I want to do more. This time it was feedback (Dan’s being primary) suggesting that even as I warned about the extremes of same/different thinking, I fell prey to it by categorizing platforms something distinctly different. Diving into that excellent question is the subject of my 14th Year III Playing to Win Practitioner Insights (PTW/PI) piece Platform Business Taxonomy: The Spectrum and its Implications for Strategy. You can find the previous 124 PTW/PI here.

A Different Taxonomy

A different way to think about platform businesses is that it isn’t an on-off switch (i.e., platform or not-a-platform) but rather a continuous spectrum from what we might characterize as a ‘normal’ business, like most of the S&P500, to one that is seen to epitomize a platform business, like eBay or Airbnb.

In laying out such a spectrum, I see two dimensions as important to cost structure and value creation, which are the two building blocks of strategy.

The first dimension is the extent to which their business involves making something in addition to operating a platform itself. There is plenty written about the characteristics a platform must have to be a successful platform (like trust, transparency, etc.), so I am not going to recite those characteristics. But I want to explore the variety of additional things that companies do along the spectrum — at least from what I can see.

The second dimension is physical versus digital. Platforms started entirely in the physical world and now find their most valuable and, often, intrusive forms in the digital world. So, I want to explore that dimension.

And the above graphic arrays a sample of the world’s prominent platforms on that spectrum against those two dimensions. On to the taxonomy…

Just Connect/Physical

Many of the world’s most historically influential platforms began life here. Their sole purpose was to connect two customer groups. They didn’t attempt to create anything other than their platform. And it was a physical platform. Examples of this include charge cards, stock exchanges and telephone systems. Each featured a fundamentally physical system. A charge card was a physical thing you carried in your wallet and handed to the merchant, stock exchanges were buildings inhabited by human traders, and telephone systems had a huge physical infrastructure, not only the company switches and lines, but the customer phone-sets as well. Arguably, the person-to-person telephone system wasn’t initially a classic two-sided market, but when the 800 number arrived, it most certainly became one.

In this type, the platform doesn’t buy goods for resale or produce anything other than the platform over which the two sides connect. And the platform tolls one side or both for use of the platform.

Obviously, all the legacy physical platforms are heading to the digital arena because of the superior cost environment in the digital domain — both cost of platform operation and cost to users of transacting. I don’t think there will be many inhabitants of this box in the future because it is hard to see how they can leverage their physical presence to create value that will overcome their cost and cost-of-use disadvantages. The likely exception is the telephone system because of its modern second purpose as the physical backbone of digital services.

Just Connect/Digital

This is the domain of some of the Internet platform originals. Their sole purpose was to connect two customer groups. They didn’t attempt to create anything other than their digital platform. Early players included eBay and YouTube. More recent examples include Airbnb, Uber, and Amazon Marketplace.

As with their physical cousins, the platform doesn’t buy for resale or produce anything additional to the platform over which the two sides connect. And again, the platform tolls one side or both for use of the platform.

My bet is that strategically, these platforms will wane over time because their value is so thin. I am always more confident of thicker forms of advantage. Unless they connect it to further production of value (which Amazon is doing with fast delivery — a further ‘product’ enhancement for the users) the advantage will remain thin. Clearly these players have both first mover and network economics advantages, but I still wouldn’t bet on the Just Connect/Digital players long-term because connection on a platform isn’t a sufficiently robust source of long-term advantage.

Connect Plus Insights/Physical

Here again, the task is to connect two markets without producing anything in addition to the platform, but in this case, the platform adds insights. The best example is American Express. Because the Amex card isn’t issued by countless independent banks, as is the case for Visa and Mastercard, Amex has full access to purchaser insights that it can sell to merchants. I think that this advantage has helped Amex prosper but like the previous physical players, it is moving towards digital as fast as it can to be able to compete with the digital players.

Connect Plus Insights/Digital

This is equivalent to the above except that the chosen domain is digital. This is the home of established giants (Facebook, LinkedIn) and successful insurgents (TikTok, Stripe). They provide valuable insights about one side of their two-sided market (e.g., Facebook users) to the other side of their market (e.g., Facebook advertisers).

I think this is a strategic high ground. Insights turbocharge the platform’s value to the ‘money side’ of the two-sided market. In due course, this box will kill the box to its left unless those to the left successfully migrate to this box.

Procure & Sell/Physical

This is arguably the biggest platform business category in history: retail. Here the platform connects sellers and buyers by creating a physical space that attracts buyers with the prospect of finding the sellers’ products nicely arrayed on their shelves. It doesn’t produce anything on its own — just the platform. But it buys the items that it puts on the platform (I.e., the store) and resells them to the shoppers. It has been a monumentally successful platform that has created many fortunes for the owner/operators of these platforms — including the world’s richest family: the Waltons.

Of course, retail is under extreme duress from its digital counterpart. Strategically, I think its only hope is to do both physical and digital and in the physical manifestation, actually produce something. And that has already happened with in-store bakeries, prepared food, in-store entertainment, etc. The advantage of procure-and-sell alone in a physical environment is just too narrow to survive.

Procure & Sell/Digital

This category is, of course, the existential threat to the previous category. Here the platform connects sellers and buyers by creating a digital space that attracts buyers with the prospect of finding the sellers’ products nicely arrayed in its infinite aisle. It doesn’t produce anything on its own — just the platform.

I would argue that Amazon started in this space but has migrated to the next space to the right by producing something of real value-added to one side of the two-sided market, and that is differentiated delivery. Now, rather than getting your items more slowly on-line than if you went to the physical store, you get them as quick or quicker. In contrast to the first two categories on the spectrum in which physical players need to migrate to digital to prosper, Amazon has migrated the opposite direction to more physically intensive to increase its competitiveness.

Make Key Piece/Physical

Next is the category of platforms that take it upon themselves to produce the critical piece that enables them to be a uniquely valuable platform. The classic example is newspapers (but the same holds for magazines and linear TV). Newspapers need to produce content that readers find attractive enough to read so that they can sell those readers to advertisers. If their key piece isn’t interesting for readers, they have nothing to sell to advertisers, and with that, they cease to be a platform.

As with all the physical categories, this category is being forced to race as fast as possible toward digital to survive because the cost of creating the product on which the platform depends is far too high to compete in the physical world. This is a dying business model — somewhat sadly.

Make Key Piece/Digital

Google is that archetypal inhabit of this box. Its product is the organization of the world’s knowledge, plus various other smaller but still monumental tasks — like Google Maps for mapping the world, and Google Scholar for organizing the realm of academic output. It is a great place to be.

An offshoot of this category is the ‘make-key-piece’ combined digital/physical platform — and that is exemplified by the iPhone/App Store combination. Apple produces a platform access device — the iPhone — which triggers use of the digital platform — the App Store — which connects iPhone users and app developers. The former brings in $200 billion of high-margin revenues, while the latter brings in $75 billion of insanely-high-margin revenues.

I think that the highest of strategic high grounds in the modern economy is the combination of Digital/Physical ‘make-key-piece’ platform businesses.

Make Most or All/Physical

This is the province of most businesses in the world. Arguably, they are platforms too. They aren’t unlike retailers who connect suppliers with buyers. The difference is that they modify what they buy rather than simply display it. They just inhabit one end of the platform spectrum.

The most advanced of them are adding a digital layer to their platform, like John Deere, which doesn’t want to just sell tractors but also be a platform for farming insights/knowledge.

Make Most or All/Digital

These are of course the software giants (Microsoft, SAP, Oracle, etc.), most of whom are attempting to become platforms on which developers create solutions that users find more valuable. I think it is fair to say that consensus is that being a closed software producer is a dying business and being an open (or at least controlled/organized) platform is the winning approach.

Practitioner Insights

That is my quick walkthrough of the platform business spectrum. I buy the argument that having a platform component to your business is going to be more rather than less important to the future of business. That having been said, I don’t believe that all platforms are created strategically equal.

Obviously (duh), digital platforms are going to win over physical ones. If you are physical and you aren’t already figuring out how to add/become digital: start today. I think lots of physical businesses will do exceedingly well by investing in a digital platform adjunct to their business. And while I am not an expert on everything John Deere is doing, I like the looks of it. It isn’t abandoning its physical strength but augmenting it.

In the digital platform world, I would aim for three strategic high grounds.

First is an awesome platform that both connects and provides insights for at least one side — but I think those that create insights for both sides will occupy the highest of this kind of peak. I am disappointed that Amazon is brazenly subverting insights for one side (buyers) to benefit itself and sellers. In my view, Amazon is actively undermining its high ground.

Second is a platform that is augmented through the production of one key piece. Google is massively successful because that is exactly what it does. Adding a unique piece that the platform produces is what makes the platform a must-have, not a nice-to-have.

Last, at risk of using a datapoint of one, I suspect that if a company can create a digital platform that is enabled by a critical physical piece that it produces, it will inhabit the highest of high grounds. Think of the monumental difference in value of iPhone/App Store versus Android/Pixel/Android Store. It is monumental. Android is an awesome platform. But Pixel is — literally — just another Google science project.



Roger Martin

Professor Roger Martin is a writer, strategy advisor and in 2017 was named the #1 management thinker in world. He is also former Dean of the Rotman School.