“Hardware is hard.” -Ancient VC Proverb
The challenges associated with producing hardware are well known. Successfully delivering physical goods to customers requires deep expertise in supply chains, logistics, and operations. It has traditionally also required large amounts of investment capital, though this is changing today. It is not wrong to say that hardware is hard, but it would be a mistake to assume that this is a bad thing. Doing the hard things well is the only way to build a long term competitive advantage.
Over the past decades, supply chains have become increasingly automated. These changes have enabled logistical complexity by driving down the marginal cost of an additional SKU. One consequence of this has been the emergence of third party logistics companies (3PL) like Flexport, Haven, and Convoy. These services change the cost structure for hardware startups operating at low volumes by reducing operating leverage, effectively reducing a barrier to entry.
At scale, the effect is sometimes reversed. For a manufacturer or supplier, automation increases operating leverage and magnifies economies of scale. Consequently, larger companies are more highly rewarded for vertical integration once they achieve sufficient volumes to amortize the fixed costs of automation. Note that this only holds at scales where production volumes exceed a large fraction of the outsourced supplier’s output such that your marginal cost-savings from vertical integration exceeds the manufacturer’s margin. This means that for widespread commodity products, the requisite scale is unachievable and vertical integration will never provide economic gains. All of these factors are superimposed on strategic imperatives which may further motivate vertical integration.
Suppliers know this well. Faced with the prospect of disintermediation by their customers, they are increasingly choosing to take equity positions in those customers instead. The most famous case was Foxconn’s 2012 purchase of a 9% ownership stake in GoPro, but examples abound. For Chinese manufacturers, this behavior is strongly incentivized by their regulatory and economic circumstances. The past three decades of Chinese growth have produced enormously profitable companies which are unwilling to keep their profits in China. Declining GDP growth, low interest rates, increasing capital controls, and a weakening yuan all strongly incentivize capital flight. For them, an investment in a US customer provides both an incentive to sell goods at zero margin and a convenient avenue to expatriate cash. This is an interesting spin on corporate venture capital which is likely to be a long-term trend.
Xiaomi represents the other side of this coin. Unlike some of its peers, Xiaomi is not motivated by fear of disintermediation or capital flight. Instead, the Xiaomi strategy is to leverage its manufacturing and distribution capabilities via partnerships and strategic investments in consumer product companies. Growth on these investments represents an extremely material component of Xiaomi’s net income. Ultimately, Xiaomi is deploying a similar set of tactics in pursuit of a different goal.
For the abundance of direct to consumer businesses flourishing in the US, it is critical to understand the scale at which vertical integration makes sense. For more specialized products, there are two paths to victory: you must either build your supply chain and logistics muscle early, or you must convince your supply chain to invest in you. Companies that do neither will be forced to pay higher margins in perpetuity and will eventually be out-competed. In some industries, an early acquisition by a large brand can be a viable solution. However, given the uncertainty inherent in an acquisition or partnership strategy, all such companies would be wise to start developing their supply chain expertise. The availability of rentable supply chains does not absolve operators from this responsibility.
When considering a company which produces a physical good, there is a significant opportunity cost. Why start or invest in a hardware company when the alternative is a pure software business with 99% gross margin that can scale instantaneously? It is always tempting to simply build for somebody else’s hardware, especially if you think you can anticipate the next hardware platform. However, this requires you to assume the additional risk that you’re wrong about the hardware. Ultimately this decision comes down to “would you rather be Apple or Uber?”
Platform technologies are the most compelling category of hardware investments. They are also few and far between. For these businesses, the hardware is primarily a vehicle which enables them to sell software and services in a way that is otherwise impossible. Platform companies can realize the best of both worlds: high-margin recurring revenues from their software products and the enduring competitive advantages that arise from doing the hard work of hardware. Apple is the example to which all such companies aspire.
Platform technologies are often conflated with the app store model. However, the two are not the same. It is usually necessary to build in-house software to run on top of the hardware platform; this can often be monetized separately. In order to support a software platform like an app store on top of a hardware platform, a product must have sufficient appeal to attract a large user base without relying on 3rd party developers. It is possible to build a very attractive hardware platform without an app store; Peloton is a great example of this. However, if an app store model is appropriate it can prove to be the difference between a Peloton-sized business and an Apple-sized outcome.
The razor/razor blade model is more common than platforms. The “razor” product is generally used to create switching costs for consumers and is often sold as a loss-leader. Though these businesses cannot achieve software economics, when done well they can generate recurring revenues with high margins.
Consumables are a third category which can generate attractive returns. On average, they are lower margin offerings than the previously mentioned models and are often extremely competitive. On the upside, they do generate recurring revenues. Brand is often disproportionately important to these companies.
Last, and probably worst, are durable goods. These are purchased infrequently, often only once. They present correspondingly few opportunities to create recurring revenue streams. In order to work, these businesses require enormous addressable markets or high prices.
There is, of course, nothing new about the appeal of platform businesses. I believe that the advantages of hardware platforms are sometimes overlooked by early stage investors. It is also true that the difficulties of starting such a company, while real, are perhaps overestimated by those whose experience with hardware production is either outdated or anecdotal. Hardware is hard, but it’s not impossible. As always, the truth lies somewhere in the messy middle. It is our job to figure out exactly where, and to make our decisions with our eyes wide open. “