HPVP’s first HW/SW (SW/HW?) investment in Protean Echo has me thinking a lot about the recent hardware boom – What it means for consumers, entrepreneurs and investors. If you subscribe to Mark Andreessen’s premise that software is eating the world, then hardware is the mouth of the beast or tip of the spear. You need a connected, sensored, mobile, low energy, attractive hardware army to vanguard the rule of software.
Consumers, entrepreneurs and investors alike are excited about connected devices. Everyone loves having something to hold, and there is clearly a boom (or bubble?). Techcrunch notes the role that Bluetooth LE and crowd funding are playing in this. Longer battery life finally means acceptable performance specs for consumers/businesses, and consumers love buying early into something cool and physical.
For consumers, this is win, win, win, win, etc…. I myself have @fitbit @sonos and recently ordered a @pebble. Of course there’s the smart phone that interfaces/runs them all. In some cases, life is much better because of these; in all cases, it is more fun.
In general, though, hardware is much lower in the stack of value creation than software. Think of it this way: value is created by people (talent) and process (software) as shown in the chart below.
You need data for insight, insight to make good decisions, and good decisions to inform action. Action creates the real value – launching a product, selling a customer, etc. On the people front, executives get paid more because they know what decisions to make and actions to take. For software, lots of “big data” companies are learning the hard way that you can’t just spit a bunch of data at customers and have them create value from it. Your software needs to provide insight and inform decision making. For all of this, however, there is a bottleneck… in the connected world you need hardware to get the data first, which is why it is the tip of the spear. Nevertheless, it is low in the stack and ultimately creates little value by itself, especially over time.
HOW HARDWARE AND SOFTWARE ARE DIFFERENT
So, the hardware story for entrepreneurs and investors is more complicated than the consumer’s. Hardware and software behave very differently over time in terms of competitive barriers, stickiness and price – three major drivers of value creation and returns. The chart below shows what I mean…
Barriers: Hardware has high barriers in the beginning – high fixed costs, long(ish) development cycles and strong patent protection. This makes it hard to launch and so gives early entrants excellent pricing and advantage. But this advantage deteriorates over time as hardware standardizes and modularizes and patents expire (if this is interesting to you, Read W. Brian Arthur’s Nature of Technology), reducing cost and speed of entry. When a new SW category emerges, however, you can build a SW MVP overnight, launch, test, iterate at almost no cost. But as the market segment matures, the concept of an MVP disappears (try building a Saleforce MVP competitor for enterprise overnight), and the fixed costs of accessing incumbents’ customer base explodes.
Stickiness: While stickiness is itself a type of barrier, it is complex enough to call out on its own. As a new category emerges, hardware has some stickiness. People/businesses buy a physical good and don’t want to replace it for a while because of the sunk cost. But in the end, it is the software behind it that makes things sticky in the long run. By its very nature, hardware has a design life. Your refrigerator’s is 10-15 yrs. Your Fitbit’s is 2-3… though they know you will lose it before then! Hardware has to be replaced eventually. When I lose my Fitbit, I will look at all the proliferating competitors – the basic function is very duplicable. Software is the other way around. I might try a bunch of consumer apps for a bit and eventually chose one to use a lot. Other people chose that one too, leading to network effects, workflow dependence, stored data growth and sometimes defacto standard development. SW gets stickier and stickier over time.
Price: The result of these trends is that the price of HW drops over time, and the price of SW stays steady or increases. Let’s look at some high profile examples starting with the iPhone:
The iPhone… not commoditized you say? When we look at the above, we see a classic skimming market entry strategy, followed by rapid commoditization at low end with textbook price discrimination among Apple religion freaks. The key here is that the bulk of the market, low to mid end, moved towards commoditization in a few short years. Other handset makers and Android figured out how to skirt patents and develop great devices, while early iPhones needed replacement in a year or two, and voila… intense competition.
Now let’s overlay Garmin nav devices for another amazing commoditization story:
Source: aaplinvestors.net, Garmin 10Ks
Arguably, this is less about commoditization and more about disruption, but a good learning story either way.
So what about Software? We see the opposite story when we look at Salesforce.com’s revenue per customer over time.
Source: ycharts, Salesforce.com earnings call transcripts
You may be thinking, yeah, but Salesforce is not profitable and Apple is highly profitable. That’s true – this trend says nothing about cost structure necessarily. But the difference in Salesforce and Apple’s profitability highlights a major difference in strategy between HW and SW. If you sell HW with no recurring revenue, you really need to be profitable. If you sell very sticky software that recurs, you don’t need to be profitable. You can always cash out the annuity later.
WHAT THIS MEANS FOR HARDWARE ENTREPRENEURS
Decide if you are a HW business or a HW enabled SW business: Is your hardware a means to an end or the end itself? I’ve seen a lot of startups confused about this. The choice is critical, informing whether the long term monetization and exit strategy are driven by software or hardware. It is fine if you are building hardware as the mouth of your software beast – invest as little as you can to get the data liberated, and then focus on the SW! Also, consider giving the HW away for free or at cost. In the end it doesn’t matter. Only the SW does.
If you are a HW business, assume your price will drop pretty rapidly: Be thoughtful about cost structure and pricing strategy, maximizing your ability to skim upfront and then going down market. The only way to avoid this is to maintain a luxury product line, which is essentially what Apple has done at its highest end. And look for sticky software models! I would be shocked if this isn’t what the CEO of FitBit is thinking about every day. It is also what Apple did with the App Store and iTunes.
Think about profitability a bit sooner than in SW: Lots of people will disagree with this and point to FitBit’s ability to raise tons of cash to support fast growth. There are always exceptions, but cost structure matters a lot in hardware. There is no annuity to cash out.
Talent: Find someone who knows something about hardware: Unlike software, where society’s knowledge base is spread pretty evenly across scrappy hackers and large companies, institutional knowledge about creating and manufacturing hardware resides predominantly in large companies. Find a COO from there to help you.
Hardware is much less nimble than software: While pre-launch SW dev costs are declining, HW dev costs are not. Don’t be fooled – there are no hardware MVPs (was the first iPhone or FitBit an MVP?), and bugs cannot be corrected overnight. Be measured thoughtful and do real old fashioned market research, benchmarking, prototyping, etc. We’re not talking landing pages with A/B testing.
Find the right investor: Lots of investors are jumping into the market, but few of us know anything about building or selling hardware. When you ask a cow to speak, it says “moo”, which is to say that most investors will probably try to force your model into the CAC/LTV software model. We did this with subscription boxes and e-commerce, and it didn’t work well in many cases. Look for investors who came out of hardware themselves or have already gone up the learning curve with other hardware investments.