China

How Much Should I Trust My CM?

Today we have a guest post by Pete Staples about finding the right levels of trust and responsibility with your contract manufacturing (CM) partner.

Pete Staples has been building embedded devices for many years. He is the President and Co-Founder of Blue Clover Devices, a full-service electronics ODM with offices in San Francisco, Hong Kong, and Shenzhen. Pete also publishes articles on the BCD blog. You can contact Pete via email.


How Much Should I Trust My CM?

Outside of the food and aerospace industries, you would be hard-pressed to find an entrepreneur who is planning to build the factory that builds their product. Everyone else relies on a contract manufacturer (CM) of some kind to put hammer to metal and bang out their product.

While many hardware entrepreneurs are close to their lawyers, investors, and accountants, often they have trouble rattling off the names of key people at their CM. While the services of an attorney are not inexpensive, they are a sliver of the spending pie next to the major wedge for the CM. Ben Einstein, now at Eclipse Ventures, wrote a useful blog about CM selection with rules-of-thumb about the economics around manufacturing. In his example, a typical hardware startup has an early valuation of $4.5M to cover the product launch and $1M (22% of their entire valuation) goes to the CM. Given the hefty investment put into CMs, why don’t hardware company leaders spend more time learning how their CM operates?

There are a couple of common reasons. One is wishful thinking. Retailers have perfected the art of making their merchandise look so plentiful and organized that we imagine the factories supplying them to be stainless steel cathedrals connected to sophisticated software stacks that calculate demand with AI and effortlessly order all of the parts to arrive at the right place at the right time. The optimistic entrepreneur thinks CMs are a lot more efficient than they really are and figures they are eager to ship anyway. The assumption is made that alignment of interests will power the project forward. It’s true that factories, like wild stallions, want to run. But factories have limited engineering resources to tackle the thorny issues that appear in the early stages of production, and this can significantly delay the launch. Product launches are a struggle and they always will be.

The second big reason for the lack of a personal connection is the effect of distance. Entrepreneurs tend to congregate around the sources of capital, professional services, and one another. Factories need more space to maneuver equipment, pallet jacks, inventory, trucks, and workers. CMs are located somewhere cheaper, and the world they live and work in is influenced by their environment. In 2016, Flex tried running a factory in downtown San Francisco (by scooping up the ashes of Quirky) but it did not survive.

Factories are run by people. We are still many decades away from robots running the show. The enduring health of your company rests upon finding CMs that you can maintain a healthy long-term relationship with. Like the pilot in command of an aircraft, someone is running your production line and their decisions will have an impact on your product. It’s expensive to keep moving production, and even if you don’t do all production with a single CM, you’ll need to invest in these important relationships.

There are many ways to manage this relationship, and the chosen style depends on how much you intend to lean on your CM’s operational capabilities.

CMResponsibility.jpg

The graph above is a way of looking at trust scenarios that we will consider. In reality there are an infinite number of possible relationships, but I’ll simplify this article by describing 3 very different approaches:

  • A: No trust
  • B: Medium trust
  • C: Full trust

In scenario A, you are calling all of the shots and essentially managing the factory yourself. You’re deciding which parts to buy and when to buy them. You provide the complete design and instruct workers how to assemble and test your product. The CM is essentially giving you a workforce, equipment, and space. With this arrangement the output is 100% your own responsibility. If you are new to manufacturing, you will learn a ton with this option, but you should also not make other plans during the build as it’s all encompassing.

In scenario B, there is a handoff generally defined by the design package. You are responsible for the design package (more on this later) and the CM is responsible for building the product per the spec. When products are returned (known as “RMA”s for return material authorization), there is an assessment of whether the return is due to a design issue or a manufacturing issue (sometimes there is no issue, often dubbed NTF for “no trouble found” or NDF for “no defect found”). The CM is off the hook if they built the returned unit according to the design. The sharing of responsibility is demonstrated by the CM managing their upstream suppliers and guaranteeing their work.

In scenario C, you are really counting on the CM to do right by you. You may be simply rebranding their existing product. You are providing some high-level requirements and acceptance criteria, and then letting the CM work out both design and manufacturing details. Even in this scenario, you will be signing a ‘Golden Sample’ but you won’t have visibility into the detailed bill of materials (BOM) or the supply chain that is supporting production. Obviously, it is nearly impossible to move the build to another CM in this scenario, but you also don’t have much responsibility to shoulder. If units are found to be defective or visibly different from the Golden Sample, it’s clearly the CM’s job to fix or replace the units.

These are all viable ways of operating, and I can think of successful companies for each arrangement described. It really comes down to being self-aware about your firm’s strengths and how much responsibility you can profitably manage.

However, there are also companies who do not see the trade-off between trust and responsibility. These companies try to push the relationship into one of two danger zones. At one extreme is the zero-zero scenario in which the brand holder does not give any trust to the CM by dictating every aspect of production. Then, the customer suddenly forgets all of these instructions when something goes wrong and claims that this was all at the discretion of the factory. This isn’t fair to a CM, who is naturally assuming that if their input is not valued that they can’t be held responsible when things backfire. At the other extreme, there are CMs who admonish their customers to “trust us, we are the experts” and work in a lot of their preferred suppliers and processes, and then don’t admit that they have co-opted the design in important ways. It’s not fair to you for a CM to insist on blind trust and then fail to step up when issues emerge. There can be higher margins for the CM in scenario C, but they must accept the heightened risk as well.

How do you even know which kind of relationship you are in? This should be discussed in the early stages of the relationship and hopefully be hammered out in a Manufacturing Services Agreement (MSA). We recommend that you have an MSA in place once the order size gets over 100,000 USD. There are complexities about warranties, excess materials, and returns that are spelled out in agreements like this, and it’s best to sort these out before the unpleasant day arrives when you have a recall on your hands.

The design package is another key element in defining responsibility. Different companies have different definitions of the design package, but here is a list of how we (at Blue Clover Devices) do it, which may be more thorough than average:

  • PRD - Product Requirements Document, overall requirements
  • CAD - mechanical and PCB design files including 2-D drawings
  • BOM - Bill of Materials, list of components in spreadsheet format
  • FMEA - Failure Modes and Effects Analysis, a risk assessment tool
  • SOP - Standard Operating Procedure, assembly instructions
  • Test Plan with specific acceptance criteria

Each document needs version control, a responsible engineer, and an approver. In smaller firms, the documentation is often managed with shared drives and a naming convention, but this approach should be abandoned like training wheels on a bike. It’s an easier way to get started, but maintenance discipline tends to break down and drags down faith in the documentation with it. We suggest using a Product Lifecycle Management (PLM) tool like Arena Solutions, Duro Labs, or Propel PLM for storing, sharing, and versioning the design package. Like starting a major construction project without blueprints, it’s reckless to begin production without a reviewed design package.

As a bonus, let me also share some of the tricks CMs play, so you can go into the relationship as an alert gazelle rather than a slow-moving calf. Smaller CMs generally have a different bag of tricks than large CMs, so I’ll group them accordingly. Note: most of these tricks are not acts of malice, but they may still have an unpleasant impact on your project.

Tricks of Small CMs

  1. Substitution - This one is really offensive but when you can’t get the component you love there is a strong magnetism to the one you’re with. When the CM is far away, there is a real risk of substituting the invisble internal parts with materials that are cheaper or have shorter lead times or -- sad to say -- are parts that came from some form of bribery.
  2. Procedure Drift - If a procedure is conducted by a person, there is a risk that it will be done differently on different days. This happens because workers are put into different positions all the time. Clearly written instructions (called a Standard Operating Procedure or “SOP” or work instructions “WI”) and a good line manager are supposed to ensure consistency, but intricate procedures are very hard to maintain and the line manager cannot simultaneously watch everyone.
  3. Skipping Tests - You might notice if a screw is missing but when you look at a product, it is impossible to tell if it’s been tested. If a tester gets up for a restroom break, there could be some units that actually skip this station. I received excellent advice from a factory manager when I was starting out. He said, “When you’re watching the line, don’t look at the assemblers; look at the testers. Really watch them, and see if they are filtering correctly.” I’ve seen lines where it was evident the tester didn’t have training to know what to inspect and just passed products randomly!
  4. WIP Growth - Work in Progress, or “WIP”, is pretty awful for a manufacturing organization. You cannot sell a half-built car. Taiichi Ohno impressed its horrors upon me in the Toyota Way, as it is an egregious form of waste. So why is WIP so persistent? In a production line, especially when people are short-handed, there is a natural desire to pick a simple process and simply “make a big batch of it”. Then we can tackle the trickier processes tomorrow. Imagine you’re the production manager and you come in and you’re tired and you have some new workers on the floor. You’d prefer to keep them busy with something that doesn’t lead to 5,000 questions and lots of line stoppages. So you put them on some easier task and they build WIP! Then when you have to integrate the whole product and make sure it all works, you might find out that the easy process actually has a little nuance you forgot to highlight and now you’ve got a giant batch of rework….more waste. You may think this isn’t your problem but it can be if the goods are no longer reworkable.
  5. Packaging - Smaller CMs have a tendency to not push for clear specifications on the packaging design and then throw it together at the last minute. The package is the protector of your precious product, and poor packaging can make an otherwise good product fail to sell through. Nobody wants goods in a crappy banged-up package, but if the packaging design isn’t done early enough, that is what you’re going to get. If you’re new to product design, you should know that retail packaging has surprisingly long lead times and may need multiple revisions. Treat it with as much care as any other part of the BOM.

Tricks of Large CMs

Large CMs usually got big by putting a lot of effort into the unglamourous job of contract assembly. If you are curious about their origin stories, check out our blog post on it. That said, they have all been around for a pretty long time and have settled into some habits that can be dangerous, if you are not a significant chunk of their revenue. If you don’t work at Apple, Google, or Amazon; then I’m talking to you.

  1. MSA - I encourage the use of an MSA, but some CMs use these as a weapon to absolve themselves of responsibility. Don’t assume that the agreement is fair because the author says it is. It might be, but you do have to read it carefully.
  2. NRE - Non-recurring expenses (NRE) can get out of hand with the large CMs. Sometimes they put together Cadillac solutions before the volumes justify such a big outlay with the attitude that “it’s not my money”. Definitely probe any line items in excess of $25K.
  3. Ownership of IP - Large CMs are powerful, and they generally build for you as well as your competitors. You need to think about which IP you want to protect as your own. ODMs (Original Design Manufacturer) are often specialists of a particular kind of product and might intend to compete with you in the future. Another term in this space is JDM (Joint Design Manufacturer), which implies the client and the CM develop some technology together. IP ownership is typically a section of the MSA.
  4. “D” Team - Big CMs have a lot of moving parts so it can be more difficult to track who is actually making decisions about ordering parts and allocating capacity for you. Talk to some other customers first to understand who is really in a position of power at the CM. You don’t want to find out too late that the only people you know there are on the “D” team, which receives lowest priority of all of the resources. All of the nimble low-volume experiments at these large CMs have started and failed several times over. They make their money from large orders from large companies period.

In summary, you can trust your CM as much or as little as you want. To build a successful relationship with your CM, regardless of how much you trust them, it is important to remember that the levels of trust should correspond to the level of responsibility you want to place on yourself or your CM. The more you trust your CM, the more you need to hold them accountable. The less you trust your CM, the more you need to hold yourself accountable. The best way to ensure success is clear communication and planning ahead of time. There are different ways to go about this, but usually MSAs (and other written agreements) provide useful guidelines that can be referenced by both parties. Remember, CMs are businesses too, and they want to succeed as much as you do!

Further Reading

Shenzhen: The Silicon Valley of Hardware

Today, I depart from the "reading" portion of the Monday Morning Reading series.

If you are not familiar with electronics manufacturing, the city of Shenzhen in southern China is the hub of all things manufacturing. Many parts vendors and CMs have offices and factories in Shenzhen. This documentary produced by Wired provides a nice look into Shenzhen.