Manufacturing for Hardware Startups Hinges on Long-Term Interest Alignment

Aug 31, 2015
Vote on Hacker News

hardware

Today it’s a foregone conclusion that hardware startups are on the rise.

Not only are there more and more exciting hardware startups springing up almost from everywhere in the world, notable exits such as Nest (acquired by Google) and Fitbit (IPO on NYSE in June 2015) have also proved that hardware startups could generate just as good returns for their investors as their lean startup brothers, if not better.

As the deal-hungry VC world wakes up slowly to this paradigm shift, the relatively old part of the high-tech industry also takes notice – namely, the EMS industry. However, as the mental gap created by over 20 years of manufacturing outsourcing is simply too huge, while the main players in the industry struggles to find a viable way to work with the startups, different “solutions” have been created by all kinds of been-there-done-that “experts” to cater to this new wave, ranging from very professional manufacturing services, such as Dragon Innovation, to brokerage platforms that claim to connect hundreds of factories to anyone in need of manufacturing even a small batch.

The landscape of manufacturing for startups is therefore fairly noisy at this early stage of the hardware revolution when most players are still figuring out how to make it work more systematically.

For young hardware startups equipped with abundant creativity, successful Kickstarter or Indiegogo campaigns but limited manufacturing experiences, the noise could be deafening. Faced with all the available platforms, service providers, incubators, accelerators or even maker spaces that bombard them with all kinds of hard-to-verify offerings, startups could at best feel overwhelmed – or at worst, take the bait and sign up with the wrong kind of services for helping on their manufacturing.

In my humble opinion, however, it’s actually not that difficult to determine which player could really help the startups achieve the goal of scaling up. It all boils down to one term actually: long-term interest alignment.

Whether it’s a small privately-owned factory in Dongguan or a big international OEM/ODM firm with an annual revenue of tens of billions of dollars, a contract manufacturer is fundamentally a fixed-cost/variable-cost business.

While in China the variable costs, notably the seasonal labors, could be managed relatively well compared to in the developed countries – at the expense of the workers obviously – the fixed costs are what make it so difficult to manufacture for startups whose volumes are invariably small to start with.

Excluding lands and buildings, which are somewhat regarded as assets that hold values relatively well, fixed costs have to be earned back over time and include mainly the machines and the management overheads. Other non-trivial fixed costs should, however, also be taken into account:

  • Downtime and extra cost due to the production line switching to manufacturing for a different product. The more such switching, the higher the fixed cost.
  • NRE (Non-recurring engineering) cost including the time/cost that the engineering team spends with a certain client to work out the DFM, DFT and testing procedures.

Another inherent cost is opportunity cost: if a contract manufacturer with limited capacity takes up a small client, it might not be able to take a big last-minute order from prestigious clients such as Apple. In practice, some contract manufacturers will just screw the small clients and switch the capacity to benefit the big clients. Most of the time the contract manufacturers hesitate in taking small clients especially before the holiday season when big orders could roll in anytime.

Given all the explicit and implicit fixed costs, it’s easy to understand that contract manufacturing is fundamentally a volume business. The manufacturer needs a certain volume for a single project to be able to recover all the fixed costs and still make money. And making money is especially critical for some of them since they might have taken loans to build the factories and buy the machines, therefore they are always under the pressure of interest and principal repayments.

A hardware startup therefore is at a disadvantage against this inherently fixed-cost business due to its lack of volume.

In fact, even as it grows to success with high volume, it might still be declined by large EMS firms simply due to the information asymmetry – the big guys have no way (and probably no interest) to verify if a young, private company really is selling $100M worth of products per quarter. I’ve personally heard such story from a regretful EMS giant regarding a hardware startup that is now public with its huge revenues visible to everyone.

How then would a genuine contract manufacturer work with a startup? It seems quite obvious that one can only sell the potential volume down the road, assuming the startup becomes successful in the future. Without the potential higher volume in the future, spending time and effort working with the startups on the current small batches doesn’t really make sense for most contract manufacturers.

And let’s not mistake a nice volume for a startup with a minimum volume for a contract manufacturer. While a startup could be excited to have a 5k-unit order out of a successful $500,000 Kickstarter campaign, these 5,000 units still rank as a pittance for most contract manufacturers – not worth the effort if it’s not already in the firm’s manufacturing process.

Just for the reference of the readers of this article, when I worked on the first contract for the HDMI chip that my startup developed in 2005, the potential order was 500k units, as a starting point.

In any case, excluding the Shangzhai model or fairly standard markets such as routers and set-top boxes, it’s hard to imagine a decent contract manufacturer could work with a startup to produce a small amount of high-quality products with brand new designs.

Unless, that is, there’s a way to align the long-term interest of the contract manufacturers and the startups.

In other words, contract manufacturers would like to work with startups that have a higher chance to grow into high volumes in the coming years, not just a one-time luck with Kickstarter or Indiegogo campaigns. By working with the really brilliant startups early on, the contract manufacturers hope to earn advantages in familiarity and specific manufacturing know-hows, so that in the later stages the said successful startups would continue to work with them on manufacturing.

The problem is: contract manufacturers would be the last people on earth to know whether a venture startup has a chance to be really successful (and hence high volume). However successful they are, contract manufacturers are fundamentally operational people with little to no understanding about the frivolous consumer market, let alone the venture startups.

As a result, however the system works, one has to find a way for the startups with real potentials to work with contract manufacturers under the long term assumptions.

If we take this view to examine the players currently trying to bridge the manufacturing gap for startups, it’s not hard to determine whom the startups should avoid and whom they could consider working with.

  • Brokers working on success fees, or worse, introduction fees, should really be avoided by the startups. In fact, even the contract manufacturers should avoid deal with these brokers. Their interest is absolute short-term and the more connections between the startups and manufacturers they can create the more money they can make, regardless of the future potential of such startups. Unfortunately, at this early stage there are many such players out there in Shenzhen. Startups just have to be very careful.
  • Expert platforms that charge hourly consulting fees should obviously also be avoided. While a French or Californian young entrepreneur might have a lot of questions regarding manufacturing, such advices should really come free from mentors, strategic investors or the contract manufacturers. There are simply too many people in Asia that could advise on these questions. Paying for such advices without any deal closing possibility is like throwing money in the water.
  • Hardware incubators and accelerators are usually (but not always) investors, so essentially their long-term interests are aligned with their portfolio companies in the form of equity ownerships. However, there are quite some incubators and accelerators who also charge fees to their startups in all kinds of excuses. Startups should avoid joining such institutions even if they do invest cash into startups. Charging the startups distorts the interest of investors and results in sub-optimal investment performance at best. Also, due to the inherent large batches of startups and low per-startup ownerships, most incubators and accelerators could only provide lists of contract manufacturers that they probably acquire from some government registry bureaus or industry associations. Such lists are very inefficient for startups since they will waste a lot of time talking to the factories that are either unqualified or have other motifs.
  • Professional services such as Dragon Innovation could be a good choice, if the startups could afford it. Note that whichever professional service provider a startup works with, it won’t change the fact that contract manufacturing is a volume business. The professional service provider should let the startups know properly why a certain manufacturer is suddenly interested in manufacturing for the startup now. It could be because the service provider preps the startup nicely in terms of DFM/DFT, thereby reducing the NRE cost for the manufacturer. It could also be simply information asymmetry, that the service provider has a network of factories that could provide dynamic capacity together. In any case it should not be simply due to that the service provider “has decades of experience in manufacturing” – you can take a car to the streets a Shenzhen and hit one person with such pedigree within 10 minutes. It’s not really an exclusive competence.
  • At the risk of self-promoting, hardware-focus VC firms should be the best options for hardware startups who aspire to become unicorns. Early-stage VC firms only realize profit when their portfolio companies successfully exit in 5~8 years. Not only will they take a very critical view to find the startups that have the best chance to succeed, after investment they will also bring all the resources they could gather to help their portfolio companies grow into dominance. Contract manufacturers could also feel more comfortable working with startups backed by such VC firms since inherently everyone’s interest is aligned: only when the startups succeed will the VC firms and the partnering contract manufacturers succeed. There’s no arbitrage here. Everyone is on the same boat.

It’s my firm belief that the eco-system for hardware startups will continue to evolve – my firm is also accelerating such evolution – so the specific details and reasoning that I mentioned above might not apply 100% one year from now. However the eco-system evolves, a startup only needs to examine the long-term interest alignment principle to have a better understanding whether the smiling man and his beautiful secretary in front of you could really help you overcome the manufacturing challenge and achieve the ultimate success.