The Agony and Ecstasy of the Full Stack Startup
In many ways the ‘full stack’ startup represents the ideal of the real-world startup. Imagine the panacea that is controlling the end-to-end experience, bypassing third party dependencies, innovating with unparalleled speed and scope of ambition. When executed to its fullest potential, the ‘full stack’ approach affords unique opportunities to build world-class products, as companies like Apple, Amazon and Tesla have demonstrated.
The term ‘full stack’ means different things to different companies, but the main characteristic of a ‘full stack’ company is the degree to which it controls and provides the full end-to-end service or experience. Almost always, the ‘full stack’ approach requires extending the company’s core competencies beyond the traditional competencies of software development, BD, marketing and sales.
In some cases, the full end-to-end experience requires moving beyond the world of digital bits into the world of atoms — Tesla manufactures cars, Blue Apron operates massive production and distribution centers, One Medical and AltSchool run retail locations. In other cases, the end-to-end experience does not require atomic manipulation, yet nevertheless requires combining technology acumen with deep functional industry expertise, ops and service. Oscar does not sell software, it runs its own insurance company. Buzzfeed does not sell software, it runs its own media company.
So why don’t we see more ‘full stack’ startups?
Notwithstanding some glaring exceptions, the ‘full stack’ approach represents the road less traveled in the world of technology startups as the approach generally poses three existential challenges, capping its appeal amongst entrepreneurs and investors:
1. Capital intensity
Almost all ‘full stack’ companies require substantially more capital than their pure software peers to achieve their product ambitions.
With marginal costs approximating zero, software is the most wonderfully scaleable good produced by humans. When Facebook bought Whatsapp for $19 billion, Whatsapp employed a grand total of 35 engineers servicing more than 450 million people, representing what is likely the most pristine realization of software’s unique potential for scaleable value creation. In fairness, Whatsapp is an unusual outlier for its efficiency — almost all successful software companies need to employ hundreds, if not thousands, of people dedicated to less scaleable ops and services. Nevertheless, the efficiency of scale inherent in pure software products is remarkably appealing to venture investors seeking the potential for massively outsized returns on every investment.
By contrast, ‘full stack’ companies necessarily exhibit fewer efficiencies of scale relative to their pure software peers. Physical goods never have marginal costs approaching zero — an obvious disadvantage for ‘full stack’ companies engaging the physical world. But what about full stack companies like Oscar and Buzzfeed that do not produce physical goods nor operate physical locations?
Like all ‘full stack’ companies, their core products are much more than just software. Software scales wonderfully, but the human services around it? Not so much. Business development, network expansion, creative writing and editorial (as examples) can only be automated to an extent. These companies need to hire actual humans to perform core product functions, not just product support functions (as is the case in pure software companies), and this generally requires more capital.
2. Operational complexity.
All startups face challenges, but not all challenges are created equal. While ‘full stack’ startups can and should be thought of as technology companies, they are not only technology companies — more accurately they are technology, ops, service and [care/education/content/food etc] companies. In order to provide the full end-to-end experience, these companies combine competencies from a number of different disciplines and are great at an unnatural number of things. To pull off the ‘full stack’ approach, these companies often ignore the sacrosanct startup ideology of extreme focus and doing one thing really really well.
While this may seem overwhelming, ‘full stack’ companies generally embrace operational complexity as a feature, not a bug. The innate complexity often represents one of the strongest barriers to entry that few entrepreneurs are able to navigate. For those that can manage the complexity, the output can be uniquely wonderful.
3. Slow early learning cycles.
Startup ideology preaches the virtues of the ‘lean’ and ‘minimum viable product’ — build things quick and on the cheap, get them into the hands of real users to test, learn and iterate as soon as possible. Unfortunately, the world of ‘full stack’ doesn’t neatly accommodate the traditional concept of ‘minimum’. Tesla’s MVP was the Roadster. AltSchool’s MVP was a real classroom with seven students. Oscar’s MVP was an actual insurance product. That is not to say that these companies didn’t test and validate hypotheses along the way, but testing the foundational hypotheses of the business required robust investment and development far beyond that required of traditional software companies.
An important implication of this dynamic is that these companies needed to, and were able to, raise substantial capital before they proved their businesses actually worked — a very difficult roadblock for most entrepreneurs to overcome.
Implications for founders and capital
Given the dynamics outlined above, the universe of capital for ‘full stack’ startups is limited. ‘Full stack’ investments require uniquely patient capital and investors comfortable with the large dollar amounts that will be invested over years of development and growth. Pools of capital outside of traditional VC, such as private equity or strategic capital, may prove appropriate for these real-world businesses.
Regardless of source, funding these business takes unique abilities and access. Building a ‘full stack’ company is incredibly challenging, and one consistent attribute across these companies is that they are able to raise exorbitant amounts of capital in advance of proving their model works. Not surprisingly, the founders of nearly every company mentioned in this post successfully founded and exited previous companies. These are people that are able to inspire confidence of capital and aggregate the funds necessary to overcome the innate frictions outlined here.
For those with the patience, cross-functional operating capacities and ability to aggregate capital, the opportunities presented by the ‘full stack’ approach are incredibly exciting. The real world is filled with enormous industries representing trillions of dollars in spend that are ripe for massive disruption. Great companies will continue to be built selling software and services into existing industry, but the most exciting and world-transformative companies will be those that disrupt industry itself by tackling the end-to-end experience ‘full stack.’
