Adapt or Die: Do Platforms Beat Products?
Hemant Bhargava is the Jerome and Elsie Suran Chair in Technology Management at the U.C. Davis Graduate School of Management. He joined us at our recent Adapt or Die World Tour stop in San Francisco.
It was a lot of fun attending Apigee’s “Adapt or Die” event and speaking on the panel “Who Ubers Who: How to Survive and Thrive in a Digital World of Mash-Ups, Matchmakers, and Marketplaces.” The best part was the 10-minute thriller movie made by Mike Slade and starring Alexandra Grossi, whom I had a chance to meet, and featuring Apigee CEO Chet Kapoor as the muffin man! A hilarious and engaging way to get the point across about APIs, what Apigee is all about.
Anyway, back on topic: the great conversation on the panel moderated by Apigee’s Bryan Kirschner.
Do platforms beat products all the time?
Bryan kicked off by asking me to comment on a provocative quote by our colleague Marshall Van Alstyne from Boston University and MIT: “Platforms beat products all the time.” I’ve spent 10 years teaching MBAs about the virtues, potholes, and strategies of platforms, so I do buy into Marshall’s fundamental premise. Platforms—which enable value creation and value transfer, frequently in a real-time, on-demand market—are powerful competitors against companies that rely on a traditional business model of being directly responsible for the product or value (with a lot of back-end sourcing and other tightly defined relationships). Think of Uber vs. taxi companies.
When and why do platforms beat products?
But there are nuances and differences. Let’s look at the when and why of the statement. The success of platforms such as Uber, Airbnb, and even eBay reveals that platforms (the dynamic on-demand real-time connections) beat product-based ways of doing business (the rigid, hardwired relationships), but because the architecture for that traditional way of business was designed and optimized based on production factors that are 20, 40, or 60 years old. In the case of one factor —information technology—that makes for a very old design. In 50 years, we’ve had million-fold improvements in cost, power, ubiquity and ease of use of computing technologies—and that is what enables today’s platforms to succeed.
Against an old approach that might be obsolete and irrelevant, newer platforms fully exploit today’s technological and economic environment, right from the ground up. While many companies have adapted to these changes (selling hotel rooms online, in-room Wi-Fi etc.), their incremental efforts are like fixing the paint and flooring when your building needs a complete earthquake-proof redesign. A good example of this contrast, in the case of electric cars, is Tesla (computer on wheels) versus the Nissan Leaf, which is really just a Nissan Versa with batteries.
That goes to the heart of the “Adapt or Die” theme. Traditional companies have incrementally adapted minor isolated business processes, when what was needed was a complete rethink of the business itself. Many of Apigee’s customers are engaged in that type of rethink, as we heard on the panel (and elsewhere at the event)—from insurance companies to banks to innovative businesses that connect yoga studios and fitness clubs into an Uber-type model. It makes me wonder if “adapt” is too weak a term; perhaps “transform” would be more apt in some settings.
Platforms do not always beat products
So, platforms do beat products when there’s an underlying weakness in the product model and when the platform itself has strong market power. But this need not always be the case.
For a counter example, consider media and entertainment. There, platforms connect content providers and consumers. Start 10 years ago— and think of your TV bundle. You received your TV shows through a platform: firms such as Comcast and DirecTV that were aggregator-distributors of content (enablers, not makers). Even though content was important, these pipe-owner distributors were vital in the value chain with near-monopoly power. But five years later, with digital and over-the-top distribution, anyone could be an aggregator and distributor, and, sure enough, many came up—Netflix, Hulu, and a host of others. Plus, creators started going direct (CBS Direct and HBO GO, for example). That deeply hurt the market power and the fundamentals of the platform model. And in this case, it became vital for platforms to become more like (integrated) products. Witness Netflix, Amazon, YouTube and other platforms moving from just enabling to content creation and exclusive deals. This underlines the notion that platforms do not beat products all the time.
That said, I agree with Marshall that the idea of platforms is very powerful and, on its own, today, would beat the more hard-wired product way of doing business.
Timing, driverless cars, and platform ecosystems
One of my co-panelists brought up driverless cars, and this hugely interesting topic brings up the issue of ecosystems and timing of platform models. Many advances have occurred in driverless cars. But for this innovation to truly take off requires changes in other aspects of the ecosystem. Take traffic lights and other support systems for automobiles. They are designed for the human eye, not for driverless cars. Imagine, instead, a world where our road transport architecture is transformed, with all the pieces working to speak directly (APIs here!) with automated cars. Driverless cars would communicate with other parts of the system in a digital language, instead of having to sense lines on roads or the color of traffic lights. When that happens, it will be a whole different world for driverless cars. So, timing matters. If your innovation occurs when other parts of the ecosystem are ready, you have a greatly better chance of success.
Timing matters, especially to platforms because ecosystems are so vital to the whole idea. An Amazon would not be possible without widespread broadband and the WWW standards. Airbnb would not be possible before reputation and rating systems. Uber wouldn’t work without GPS and mobile. The earliest search engines—well before Google—vanished because many key parts of the ecosystem weren’t yet in place.
I should know this well. In 1996, Michael Kirschner (no connection to Bryan) and I were building HealthSelect, a web-based decision support system to help consumers and employees analyze, compare, and select health plans. But, among other things, plan description data wasn’t yet structured or available online, doctor ratings and reputation systems were non-existent, and so on. We wrapped up shop two years after trying. Imagine doing that today, under Obamacare where the decision problem has become central for millions of consumers, and all the data are available online, and with APIs becoming a standard way to connect systems and data and analytics. Perhaps it is time again to leap backward!
Most of all, it was joyful to observe the intensely positive “Can-do,” “Yes, or here’s how” (rather than “no”) philosophy and exuberant attitude of Apigee employees.
A version of this post originally appeared on Medium.