Common Sense in Technology Cycles
Entrepreneurs have been building software businesses for more than two decades which has brought us to a period of relative maturity in the current technology cycle. Over the years, we’ve distilled the strategies of the successful companies into best practices for the next entrepreneurs to follow. These methodologies are at the core of how we think about business so it’s worth taking a step back and reflecting on how we got here.
This post was inspired by Carlota Perez’s book on technological revolutions, a fantastic book that’s pushed me to think beyond the one year funding horizon and consider the broader technology cycle. I thought it would be interesting to examine where we are today in regards to her framework published 14 years ago.
Stages of Technology Cycles
The current and latest technology cycle is commonly known as the internet and telecommunications technology cycle. It began with the invention of the microprocessor in the 1970’s and has profoundly transformed the global economy.
Perez breaks down cycles into an invention stage and an application stage. The invention stage of the current cycle spans the creation of the microprocessor, the personal computer and the telecommunications infrastructure necessary for the internet. Exciting developments in hindsight but most of the world was unaware of the enormous potential at the time. In a much referenced quote, the founder of DEC remarked in 1977 that “There is no reason anyone would want a computer in their home.”
After the pieces were in place, the cycle transitioned to the application stage as new companies began to leverage the technology of the invention stage. Thus began an explosion of software in the mid-1990’s that, despite the ebbs and flows of financial capital, has continued through today.
The above graph shows the number of IPO’s from 1970 through 2015 grouped by infrastructure (semi-conductors, computers, etc) and software. While it’s clear the growing share of software (red line) as a percentage of IPO’s demonstrates the transition in the technology cycle, this graph actually underplays the magnitude of the shift for three reasons:
- Some technology companies (for example Amazon) are not assigned software SIC / NAIC codes. However these companies are prime examples of the application stage as they demonstrate applying technology to new markets and should be included as application stage IPO’s.
- IPO’s are an approximation of the technology cycle and not a perfect one. Much of the economy (corporate, government or otherwise) is not included in this graph. There has been significant economic spend on application technology in-house (think of any custom software at a large company) that would not be reflected by IPO data.
- Dynamics of the current capital markets favor remaining private for later stage companies. Many of the application stage businesses that would have IPO’d had they launched in the 1980–2000 period have chosen to pursue an alternate funding path to an IPO.
Codifying Common Sense
Underlying the application stage is our evolving comprehension of the technology cycle. It was clear after a spectacular financial crash that we overestimated the immediate opportunity for internet platforms. In the period after the crash, investment withdrew until smart phones provided a catalyst for internet platform growth. By the time financial capital returned to the technology sector, the application stage had already swung into full gear.
A few years after the bubble, forward-thinking investors and entrepreneurs began to recognize similar structures that defined successful internet platform businesses. Among this group, the investors at Union Square Ventures developed a unique viewpoint on network dynamics of internet platforms in their 2004 venture fund. With investments in Twitter, Zynga, Etsy and others (most platform businesses), this became one of the most successful venture funds of all time.
As the success of internet platforms and their investors became visible, the commonalities among the companies were highlighted and codified by individuals such as Fred Wilson, Paul Graham and Eric Ries. While much of their excellent advice can benefit the broader business ecosystem, the majority was crafted with internet platforms and software start-ups in mind. Others have built upon their work and this knowledge-base has grown to encompass almost every aspect of launching a software start-up.
Perez highlights this as a pivotal moment in the application stage when best practices become universally understood to be true (also known as a “techno-economic paradigm” or “common sense”). Paradigms are unique to each technology cycle as best practices of the prior cycle do not transfer entirely to the subsequent cycle; the paradigm needs to be discovered, designed and accepted for each new cycle. Once in place, entrepreneurs and investors can follow the paradigm to launch a business with reduced risk. However the discovery of the paradigm brings new entrants (both entrepreneurs and investors) that can lead to diminishing marginal returns given higher levels of competition and fewer market opportunities. Discovery also means that knowing the paradigm before everyone else is no longer a competitive advantage.
A techno-economic paradigm is, then, a best-practice model made up of a set of all-pervasive generic technological and organizational principles, which represent the most effective way of applying a particular technological revolution… these principles become the common-sense basis for organizing any activity and for structuring any organization.
What is the paradigm of today’s application stage? Whenever you hear commonly used business model strategies, such as product-market-fit, lean start-up, [blank]-as-a-service, [blank]-for-x , you are ingesting one facet of the paradigm. These strategies are the commonalities among two decades of successful application stage businesses, mostly in the form of internet platforms, that can be grouped by how they leverage the technology of the invention stage (zero-cost, instantaneous data transfer):
- Discovery (Google, Yahoo, Bing, Apple, LinkedIn)
- Communication (WhatsApp, Facebook, Twitter, Skype, Yo)
- Information & Media (Facebook, Instagram, Twitter, YouTube)
- Transaction (Amazon, Uber, eBay, AirBnB)
- Processing (PayPal, Salesforce, Quicken)
What’s more, the nature of an internet platform (or software business in general) has unique attributes during the growth stage:
- Historically Low Initial Capital Requirement
- Low Barrier-To-Entry (Third Party Resources, Cloud Infrastructure)
- Near-Zero Marginal Cost of Goods
- Immediate Product Iteration (Minimum Viable Product, A / B testing, etc.)
Too Much Common Sense?
Venture investment is a great indicator for the discovery of this cycle’s paradigm. Investment remained relatively flat until the outsized successes of the 2002–2007 period began to appear in 2010–2012. At which point, a flood of both entrepreneurs and financial capital began building exciting companies following the construct of the internet platform paradigm.
As available finance makes [entrepreneurs’] projects possible and as their astounding successes makes the paradigm even more visible and attractive to a greater number of people, the ranks of those that feel the calling will invariably swell.
While many point to the potential specter of a bubble, I believe it’s instead signs of increasing economic difficulty driven by competition and decreasing marginal returns of deploying mature technology. In my view, this cycle has many more years to go improving process, communication, and integrating into every last fiber of the economy. Smart investors are still funding new iterations of platforms or entrants into new geographies.
This paradigm has proven a reliable framework for designing an internet platform. However it won’t be the paradigm for the next technology cycle and it may not be a perfect template for every business today.
Chris Dixon highlighted some of the emerging building blocks for a new wave of physical technology: AI, sensors, system-on-a-chip, motor drives, additive manufacturing and advanced wireless networks. The next cycle may involve heavy service elements, large initial capital expenditure, a lack of third party infrastructure and a departure from fast product iteration. These factors go contrary to what the existing software paradigm would define has a viable, scalable business and yet are all inescapable realities of the supply chain, manufacturing and transportation industries. With the software paradigm deeply embedded into financial capital, resources and cultural structures, this dynamic may make it difficult for a new profile of technology businesses to gain support and raise capital.
Approaching the fifty year mark of this technology cycle, it’s difficult to see how long this paradigm will continue to be successful. Just like the smart phone, new catalysts (virtual reality, block chain) could provide additional channels to deploy internet platforms and extend the lifetime of the current paradigm’s utility. An important question to consider is whether the new wave of technology in the physical domain is the start of a different invention stage or the evolution of the existing application cycle. In my opinion, it does appear to depart from the nature of application of internet & telecommunications technology seen since the mid-1990's.
The resources, funding structures and organizational momentum of the current paradigm make it an incredible time to be an entrepreneur; it is easier to start a successful business than ever before. However the resources available are a product of generations of a particular paradigm which may not align perfectly with next-generation businesses. If history repeats itself, the greatest businesses are always the ones that broke the rules.
The potential of a paradigm, no matter how powerful, will eventually be exhausted. Technological revolutions and paradigms have a life cycle of about half a century, which more or less follows the type of logistic curve characteristic of any innovation.
The best thing about writing about technology cycles is that I can’t be proven wrong until 2070. :)