Product Management for the Physical World Series 🧵: Part I
Important disclaimer:
All charts shown are for illustration purposes only, and do not represent factual data or research. They were generate via prompting GPT4.0 with the open questions in the titles, and were then developed via dialog with the LLM.
Bare metal and Software: Convergence
Compare Ford Model T from 1908... with Tesla Model Y, launched in 2020.
Quite a lot have changed, for sure, but in this automotive example, not everything, is it? We all, contemporary drivers, could drive this primitive Model-T, while any early 20th century driver would certainly master a modern day connected car, in a just a few hours time.
A lot has changed, though, in the development of hardware products, especially those who run also digital interface. Smart, connected devices bring forth a unique combination of capabilities that changes not only the nature of our involvement with them, but also their development process, their life span expectancy, their unit economics, and the assortment of services that manufacturers can now credibly offer to us.
This conversion, I believe, is the driving force behind the "smart and connected" trend as the one emerging strong for many for many product categories.
I'll try to touch on some of the differences between rudimentary and smart products, and dwell on the challenges product managers face in this journey.
Why is hardware so hard?
Let us first establish the problem. Hardware startups tend to fail more than their software comparables, with an average of 3% success rate for hardware startup vs. 10% for software.
Some of the reasons:
Sunk Costs
As an example, consider Dolav's residential waste bin I designed, back in 1996.
The mold required for injection of the main body weighed 80 metric tons - and it was accompanied by several smaller ones, each alone weighing a few tons.
It took two semi trailers to transport those molds from the port to the factory, where a two-story building sized injection machine was waiting for it.
The risk of commercial or technical failure associated could have led to losses of many millions. Furthermore, such a failure would render the production tools practically obsolete, leaving them with minimal opportunities for reuse.
Slow ramp-up, longer time-to-revenue
The creation, transportation, setup of these molds took about a calendar year, a duration that is nearly unheard of in the software industry. These were by no means an exception.
Engineering and G&A costs are common in the software industry, the time-to-revenue tends to be much longer for hardware companies, increasing the sunk cost burden.
Lower profit margins
The added cost of production facilities and equipment (PPE), raw material and inventories spread over the distribution channels also adds in. This leads to 10-40% margins for hardware products, compared to 60-90% gross margins for software
Higher and earlier project failure
Slower growth
Physical supply chains and distribution channels dictate lower, cash intensive ramp-up. They need to account for shipping duration, inventory management, and last mile distribution - all absent from most software product.
This results in slower growth for hardware startups compared with software.
Longer ROI period
If you visualize the success of a startup by charting the famous hockey stick graph of its cash reserves over time, you can compare the average curve for hardware and software products, and see that it takes longer to recoup investment.
Where capital flows
The above might be construed as if hardware products are to be avoided at all cost... But this is not my intention. It is true that intensification of competition and regulatory burden make it harder to successfully launch hardware products.
At the same time other trends played a strong positive impact: Simplification of supply chains, the rapid growth on the global market, as billions joined the advanced world economy.
I suspect these positive trends contributed to the proliferation of successful physical products: The sheer increase in number of people to dress, to house, to transport and to enable digital life for is a formidable force for good (yes, I am ignoring environmental externalities for now).
My point is that the same trends that enabled the past decade growth made information technology even more successful, enabled the launch of innumerable software products while decreasing risk and development costs. Indeed, software products are built on the shoulders of giants that continue to grow tall.
Consider venture capital flows in recent years. First, you'll notice the change in the ration between software and hardware:
But the change is even more accentuated when you measure total investments:
As software keeps "eating the world", and VC continues to grow , you'll notice the decline in hardware, with a concurrent infliction point in year 2020.
(Provisional) Conclusion
Don't. Do. Hardware!... Or maybe, this was a bit too harsh...
In the next posts I will discuss what trends can be observed that alleviate some of the pain, specifically by means of convergence between hardware and software, and furthermore, I will discuss some product management practices that can (and must) be implemented as a part of the new product category of smart, or connected, products.