Defensibility in Hardware Startups

A great business creates a virtual monopoly for a certain period of time. Usually this is not a true monopoly where the competition is prevented from selling their good or service – you’re not allowed to do that. But the most successful businesses create effective monopolies in their markets, by doing things and going places that make it very difficult for the competition to follow.

Doing something that’s hard for the competition to replicate doesn’t necessarily mean rocket-science-level hard technology. I use the metaphor of a “moat” to think about startup defensibility – it’s the barrier (possibly a low-tech one) that slows down the hordes when they decide to storm your castle, so you can keep making and selling your product profitably.

Hardware products have a key defensibility flaw: they can be taken apart and copied by your competition. In the past, the copies might not be as good as the original (e.g. HiPhone vs iPhone), but these days copies are sometimes even better! (e.g. Xiaomi camera vs GoPro) It might seem strange that it’s easy to copy hardware, since startup people usually think of hardware as being hard – but there are lots of companies in China that are extremely efficient and awesome at making perfect/improved duplicates of existing products.

The standard refrain these days about how to make a hardware product defensible is: It’s the software!

Is software actually a moat? First, let’s get specific about what this statement means. Although most hardware products have software/firmware running on the device, the software that matters for defensibility is more often a cloud service that connects to the hardware. It’s pretty easy for anyone determined enough to copy your on-device software, but the code in the cloud – proprietary algorithms, communications, unlimited compute and vast datasets – that is the defensible software.

So what are the ways to make a hardware startup defensible today? Here are a handful of strategies that I have seen play out in the market in recent years – I’d love to hear thoughts on other approaches that I’ve missed:

    • Build a brand. Win by dominating mindshare. GoPro did this successfully to beat competitors like Contour in the action camera space [1].
    • Build a community. (strongly related to building a Brand) MakerBot wasn’t the first, the cheapest, or even the best consumer 3D printer of its time. So why did Stratasys acquire it for north of $400M? Makerbot’s community-contributed cloud content platform Thingiverse was the #1 destination for 3D printing enthusiasts sharing their designs. This community, enabled by the platform, made the Makerbot experience unique in the market.
    • Be first to market. I believe that first-mover advantage in hardware is more important than in software. After all, Google was not the first search engine and Facebook was not the first social network. However, Dropcam was the first consumer-friendly WiFi security camera, and Fitbit was the first connected pedometer – these companies built category-defining products [2]. Lots of competitors followed Fitbit and Dropcam after these categories were defined, but none have achieved the same level of name recognition or sales.
    • Be (very) profitable. With high margins, you can spend more money than your competition to acquire new customers, and this can be a defensible advantage. Fitbit has always enjoyed great unit economics on their hardware, and a connected-hardware company like Square can achieve good overall economics via a blended hardware/software margin.
    • Sell your hardware as a service.  Managed hosting companies like Rackspace and AWS have been selling hardware as a service for a long time. A new wave of robot-enabled businesses (e.g. Dishcraft Robotics) is now beginning to apply this model. There are numerous benefits of the hardware-as-a-service (HAAS)/robotics-as-a-service (RAAS) model:
      • The cost of the system doesn’t hit your customers’ capex, making the sale easier.
      • If the startup operates the hardware itself (e.g. managed hosting) the hardware doesn’t have to be perfect to begin selling the service, speeding time to market.
      • It may be possible to prevent your competitors from having a close look at the hardware (especially if the startup operates the hardware itself), making the protection of trade secrets feasible.
    • Go double-black diamond.  Build something that is truly difficult to copy, via cutting-edge technology and/or high system complexity. This is more often doable by big companies, but can be in the domain of startups – for instance with a tech breakthrough transferred out of university research.
    • Collect proprietary data.  This is not specific to hardware companies, but for startups building new applications atop machine learning, having a dataset that nobody else has can be a competitive advantage.
    • Clear regulatory hurdles. (related to the double-black diamond approach) The time and expense of gaining certain regulatory approvals – for instance FDA certification – is a barrier to entry for newcomers into a regulated market, so a company that has managed to gain regulatory clearance can put distance between themselves and the competition.
    • Network effects. A product with network effects gets more valuable to customers as the customer base grows in size. Facebook and telephones work this way. This is also a key aspect of Fitbit’s defensibility in the era of cheap/copycat activity trackers – since Fitbit customers can compete against their friends (but only if their friends own a Fitbit), there is an incentive to buy Fitbit.
    • Price/market the product in ways incumbents cannot. Sidestep the constraints that bind the entrenched incumbents. For instance: Honest Tea marketed its product as a healthier, less-sweet alternative to sugary iced teas in the market owned by Coke and Pepsi [3]. These large incumbents would have had to repudiate their own product lines to follow suit. Dollar Shave Club eschewed over-engineered and expensive razor-blades from P&G, offering a simpler and lower-cost option. P&G needed to maintain high margins on their blades to support their big investments in R&D, so they could not follow [4].
    • Scary / Taboo markets. Today this means domains like sexual & reproductive health, marijuana-tech. Uber was an example of this when it started – everyone said “you’d be crazy to take on the entrenched incumbents in the space – the taxi dispatch!” Paypal Many entrepreneurs and investors are too timid to enter these markets, leaving an exploitable competition vacuum to those who will. This defensibility won’t last very long though, as evidence of wins in these spaces will encourage others to jump in.
    • Ecosystem lock-in. Products with consumables (K-Cups for Keurig, eBooks for Kindle) or product lines having shared accessories (batteries for power tools, lenses for Nikon/Canon cameras) introduces a consumer lock-in dynamic due to the switching cost to the consumer. Once the customer has invested in the accessories or consumables, they are more likely to stick with the product line.
    • Keep secrets. Don’t announce till you are 4 weeks from ship. This is antithetical to the current Kickstarter method, but is a must in a world where customer attention is short (delays will kill you) and china is good (they will beat you to market if you give them >2 months notice). It’s more difficult to keep the details of how the product works post-launch, but possible if real secret sauce is software that runs in the cloud where its inner workings can be hidden behind an API call. 
    • Win the ground game. Couple concepts here:
      • Have excellent customer service and support: This can differentiate a startup vs a larger incumbent. As Paul Graham put it: Tim Cook doesn’t send you a hand-written note after you buy a laptop. He can’t. But you can. That’s one advantage of being small: you can provide a level of service no big company can [5].
      • Regional expertise: Founders that understand their local culture well can create products that resonate with their regionally-based audience better than outsiders can. Facebook is dominant in the US, and Renren leads in China.

In closing: There are a bunch of approaches to making a hardware startup defensible. Not all of these ideas will apply to every company, consider this a sampler that lays out some of the most common approaches. The best defense is a good offense – make something great, something that your customers want to “grab out of your hands” as Steve Blank puts it, and keep moving fast. And remember that in most startups  you win by endurance [6] – you’re in a marathon not a street fight.

Further reading:


Thanks to Jeevan Kalanithi for feedback on this post.