I often ran into hardware startups who told me that they would like to engage with EMS firms only when the product designs and prototypes are close to final. While this sounds totally intuitive and reasonable, it runs against the logic in actual practices.
In general EMS firms have two business models: OEM & ODM.
OEM (Original Equipment Manufacturing)
In this model, the client has a product that's 100% completed or a product that's been in mass production for a while and is looking for a 2nd source for the product. Either way the testing programs, quality assurance, etc, have all been figured out. The EMS firms only need to mass product the product.
This is the intuitive business model that most people outside of the hardware industry would imagine. It's very common that startups believe this is the model for them.
The truth is, for OEM model the client can switch EMS firms relatively easily so EMS firms will have to make money contract by contract. This means that the project absolutely has to be high-volume.
What would count as high-volume? To give the readers a reference, during my entrepreneur years (2004-2008) I had a potential 1st order from a Top 3 EMS firm for our HDMI chips. The order size was 500,000 units.
It's clear that no early-stage startups will have this kind of volume, so startups should not even think about going for OEM model since no real EMS firms will work with you.
There are a few exceptions where the OEM model might work – when the product is based on a well-known reference design. We see this in WiFi routers very often. Most people are using reference design solutions from either my former employer Atheros or Broadcom or the ever-so-cheap Realtek. The innovation and differentiation are mostly in firmwares and softwares. In this case there are a lot of factories in Shenzhen/Dongguan that already have such reference designs in mass production might be willing to take your projects based on OEM models.
However, if you're building a robot or a special sensor-driven smart-home product, you could pretty much forget about OEM model.
ODM (Original Design Manufacturing)
ODM model, as the name suggests, means that the EMS firms also get involved in the design phase to a certain degree and for sure they're involved in developing all the testing programs.
Sometimes the clients don't even design the products but just describe the specs, functions and looks that they want. The EMS firms take care of everything for them.
Believe it or not, most of the electronics products on the markets from the big brands come from this model. This includes laptops and tablets from Lenovo, HP, Sony and Toshiba, consumer cameras from Panasonic, Fujifilm, Canon and Nikon, audio speakers from JBL, Logitech, etc., and all toys that you see in Toys R' Us.
Heck, even GoPro has always been ODM model with its long-time EMS partner Chicony!
While I'm not suggesting that startups do zero designs, I do want to say that this is the model that's most likely to work between a startup and an EMS company.
The reason is very simple: long-term interest alignment.
Startups do not have volume, especially in early stages. For an EMS company to work with a startup, the EMS company has to take a long-term view. If the startup will just easily switch to the cheapest supplier, then why should the EMS firm even work with them from the beginning?
On the other hand, by working with the startups on industrialization, building testing programs and QA programs, the EMS firms know the startups won't easily switch. That allows them to take the long-term view and hammer out multi-year growth plans together with the startups.
This is the only win-win situation.
Proofs to back this up are abundant, most notably GoPro, which we just mentioned, and Fitbit, who still has Flex as their main EMS partner.
In other words, ODM is actually the real model that might work for startups and to do ODM the startups cannot wait until the design in 100% completed but have to engage earlier, making the EMS firms their real partners.
Other reasons to engage EMS firms early
If the startups keep in mind that modern EMS firms have been doing hardware/software R&Ds and mass production for their brand clients for more than two decades, they should know that whatever in-house talent they have, there's no way they could be more experienced in bringing a product to mass production than the EMS firms.
By working with EMS firms early on and directly, the startups benefit from fewer iterations in DFM, DFT, EVT, DVT and PVT. The testing programs will also be built to cover the most likely scenarios given the EMS firms' experiences. Everything is moved forward with mass production in mind. No wasted energy.
What if the EMS firms steal my design?
The EMS firms will be killing themselves by doing so. In fact, many modern EMS firms used to be both brands and OEM/ODM companies. it's the same concern from their OEM/ODM clients that forced them to spin off the manufacturing arm to as 100% EMS firms. Just to name a few examples:
- Pegatron was spun off ASUS
- Wistron was spun off Acer
- Qisda was spun off BenQ
In other words, EMS firms' job are 100% to serve their clients. Stealing a client's design will kill its business reputation completely. No clear mind will even try this.
And I often joke that if an EMS firm steals a startup's design, the startup should be happy! Because it could sue the EMS firm and earn more money than it could likely sell its own products in the early years! Note that this should remain a pure joke as all startups should build trust with their EMS firms and treat them with respect.
Any reason not to engage EMS firms early?
If you haven't completed your prototype, don't even bother. As a startup you're way too early for the EMS firms without even the 1st prototype.
Other than this, I couldn't think of any reason not to engage EMS firms early.