Intro: Why is Hardware THAT Hard
- Yoel Frischoff
- Jan 29, 2024
- 7 min read
Updated: Mar 16
The unique challenges product leaders face with integrated hardware and software products. It emphasizes the importance of early validation to mitigate risks and the difficulties in creating meaningful MVPs due to the monolithic nature of hardware products.
Product Management for the Physical World Series 🧵: Part I
A Short Intro
Many things have changed in the automotive world over the 115 years between theproduction of the Ford Model T, the pioneer of modern cars, and the start of production for Tesla's Model Y, a battery-powered electric car connected to the internet via satellite — yet not everything has changed, has it?
All of us, contemporary drivers, can drive a Model T, and any early 20th century driver would likely master driving a modern car within just a few hours.
Despite this similarity, much has changed in the development processes of hardware products, historically plagued by (very) high business risks: Advanced design methods, easy (and realistic) prototyping, flattened supply chains, improved inventory management, and streamlined production planning — all reduce risks and improve the prospects of hardware products.
In this post I focus on smart, connected products — those that combine software, digital user interfaces, and internet connectivity.
I explore how the integration of physical and digital capabilities transforms development processes of hybrid products, as well as how we use those: I discuss how this integration reshapes lifecycles, the business models such products support, and the range of new services manufacturers can now offer to their users, over extended product life time.
This transformation, I believe, is the driving force behind the growing trend of "smart" products across many categories, as it alters the risk and opportunity landscapes for such products, enlarging the scope for competitive edge, customer and user loyalty, and, consequently, the prospect of success for ventures that skillfully integrate digital features into hardware products.
To this end, I first address some of the differences between developing software ventures to "dumb" hardware ventures, to ventures that develop integrated (smart and connected) products.
I discuss the challenges faced by product leaders during the establishment of hardware ventures and demonstrate how the trend of developing hybrid products is positively reshaping the picture even for hardware ventures.
Why is hardware so hard?
Let us first establish the problem.
According to CB Insights, the average venture success rate for hardware is only 3%, compared to 30% for software, with "success" defined as the survival beyond ten years.
This begs the question: What factors contribute to the difference in success rates between hardware to software ventures, and what insights can we draw to improve the likelihood of success for the hardware venture we are involved in as product managers?
Some of the reasons:
Sunk Costs
A sunk cost is an expense already incurred which cannot be recovered. When designing a hardware product for mass production, a significant portion of the development expenses fall into this category. Most of these costs cannot be repurposed for new products in the event of failure and are likely to be completely lost.
Take, for example, a domestic trash bin manufactured by Dolav, produced via HDPE reaction injection molding. The injection mold required for producing the main body consisted of two steel blocks weighing 80 tons in total. In addition to the main mold, several others were needed for the lid, wheels, and various auxiliary parts, weighing several tons each.
Two semi-trailer trucks were required to transport the main mold from the port to the injection molding facility at Kibbutz Dvir, and other trucks were needed to deliver the smaller molds to the factory at Kibbutz Gadot, where the smaller parts were to be injected.
The cumulative development cost ammounted to millions of Euros, a significant economic risk. A commercial or technical failure could have resulted in additional losses of several million euros due to tied-up production resources (raw materials, labor costs, injection molding machines).
Moreover, such a failure would have turned the the production tooling, ie the molds, into massive scrap metal, left to rust quietly in some warehouse, or sold by the weight for recycling.
Of course, sunk costs are rife in the software industry as well. However, the magnitude of the problem seems smaller:
Slow ramp-up, longer time-to-revenue
The production, transportation, and installation of the molds in the previous example took approximately a calendar year, a timeline almost unheard of in the software industry, but sadly by no means an exception in the hardware industries. A new car, for example, can easily take up to six years from conception to manufacturing.
General and administrative expenses, including development and maintenance costs, are, of course, common in the software industry as well. However, the time required to launch and begin generating revenue from the product—and consequently, the waiting period until the breakeven point and return on investment—tends to be significantly longer for hardware companies. This increases the financial burden on investors.
Several studies highlight significant differences between software and hardware ventures, to the detriment of the latter:
Metric | Software | Hardware | Multiplier |
Time to market | 9 | 18 | 2.0 |
Time to breakeven | 21 | 30 | 1.43 |
Time to ROI | 24 | 71 | 2.93 |
We will later revisit the impact of prolonged waiting periods on investor considerations in hardware ventures.
Lower profit margins
Compared to software companies, hardware ventures bear additional costs for production facilities, factory space, and logistics (PPE), as well as the need to finance raw materials and inventory spread across the supply chain.
These additional costs result in relatively lower profit margins. We examined ventures within one standard deviation from the industry average, covering 68% of companies under a normal distribution.
According to a study by E&Y, software companies achieve gross profitability ranging from 75% to 85%, while for hardware companies, the average gross profitability is lower, ranging from 42% to 44%, with significant variation across industry sectors.
Higher and earlier project failure
Hardware ventures face a higher risk of long-term failure compared to software startups, with the gap becoming more pronounced around the heavy investments required at the industrialization phase.
Here are several eye-popping examples of software venture failures from a CBInsights study conducted in 2017:
These examples illustrate the substantial amounts required for industrialization, marketing and distribution, all while building the supply chain and inventory necessary for hardware products - and the sheer risks hardware entrepreneurs must undertake to scale.
Slower growth
Supply chains and distribution channels characteristic of tangible products also impose slower growth and intensive cash consumption (working capital).
These entail constraints such as shipping duration and costs, inventory management, and last-mile delivery challenges — all of which are thankfully absent from the business model of most software products.
One of the main outcomes of these constraints is slower growth for hardware ventures, especially in comparison to software.
Source: Equidam and author's processing
Longer ROI period
If you also visualize a startup's success by sketching the famous hockey stick graph of its cash balance over time, you can compare the average curve for hardware and software products and see that it takes much longer to achieve a return on investment.
Note: This refers to a return on investment from organic sources, meaning from the company’s revenues and profits, rather than from an exit event where new investment or debt replace early equity.
Where capital flows
The above might be interpreted as suggesting that hardware products should be avoided at all costs... but that’s not my intention. It’s true that the factors mentioned earlier weigh heavily on the initiation, industrialization, and distribution of hardware products.
Undoubtedly, increasing competition and regulatory burdens make it even harder to successfully launch hardware products.
Other trends in recent decades, however, have had a strongly positive impact on hardware product development, and by implication, on new ventures in the field. These include improved design and prototyping processes, supply chains simplification, and, not least, the rapid growth of the global market, which has brought billions into the global economy and into production and consumption cycles.
It’s reasonable to think that these positive trends have contributed to the proliferation and success of physical products: The significant rise in purchasing power among a growing number of people needing clothing, housing, transportation, essential products, and entertainment represents a powerful positive force (yes, I am ignoring environmental externalities here).
The point is that these same trends that enabled the growth of recent decades also made information technology even more successful compared to hardware. Countless software products have been launched with reduced risk and plummeting development costs, resulting in lower damage in the event of failure.
One reason for this is that, unlike hardware, software products are developed as layers built upon the foundations established by industry leaders, who themselves continue to advance.
Phenomena like cloud-based infrastructure as a service (SaaS, PaaS, IaaS) further reduce costs and accelerate the creation of new software products.
Indeed, it’s interesting to examine venture capital flows in recent years.
As software continues "eating the world," a consistent ratio between investments in software and hardware can be observed, allowing us to infer the availability of capital for the hardware sector (and perhaps the availability of investors capable of committing to the marathon required).
(Provisional) Conclusion
Don't. Do. Hardware!... Or maybe, this was a bit too harsh...
In the upcoming posts, I will discuss trends that help alleviate some of the pains associated with establishing hardware ventures.
In particular, we will examine the effects of the convergence between hardware and software in the form of smart or connected products.
Additionally, I will explore several product management practices that can (and must) be implemented as part of hardware product development.
Read the rest of the series:
Bonus posts
Comments