How To Manage Your Contract Manufacturer Part 3: Quality / Testing in Production

The following post is the sixth in a series of guest posts for GrabCAD.

Manufacturing quality can only live up to the quality of the design. What you design is as good as the product will ever be.  It’s relatively “easy” to make one widget but making 100K widgets that perform according to spec under a variety of real-world conditions?

That’s another story entirely.

dragon innovation quality testing

Here are the areas that we focus on when it comes to quality testing:

  • Printed Circuit Board Assembly (PCBA) Tests:
    • In-circuit tests (ICT) – bed of nails
    • Automatic Optical Inspection – AOI
    • Functional / Built-in tests
  • Overall Mechanical Assembly
    • Subassembly functional testing
    • Overall functional / Built –in tests
  • Ongoing:
    • Torque / Tension / Drop
    • HALT / HASS – accelerated life/environmental
    • Conventional Life test

Production Testing / Final Inspection:

The design has been verified and the purpose now is to ensure that it is built correctly.

  • Normal function testing can usually be tested on 100% of the units.
  • Abnormal function testing sometimes can be tested on all of the units.
  • Abuse testing cannot be tested on any units that are then shipped to the customer.  As a result, when testing damages the product so it cannot be shipped, then it must be tested on a sample basis.

So what are the actual things you’re testing for? Here are our recommendations:

  • Mold qualification dimensional tests
  • Static Discharge Test
  • Aging Test
  • Humidity Exposure Test
  • Environmental Test
  • Life Test
  • Regulatory and Safety Requirements
  • Final Inspection
    • Sampling plan
    • Defect Definition (Critical, Major, Minor)
    • Date Code
    • Transportation Test
    • Aesthetic Inspection
    • Functional Test
    • Disassembly Check
    • Critical Measurements
    • Drop Test
    • Torque/Tension Test
    • Compression Test
    • Paint Abrasion Test

Finally, it is critical that you use documentation control for all of your testing. So critical in fact that we’ll say it again: Documentation control. It’s important.

How it’s implemented:

  • Use ECN’s for every change
  • Tie ECN’s to the date code/ production revision

Why it’s implemented:

  • You need to know what you’re building
  • Sometimes required by regulatory agencies (FDA)
  • Helps tie production and/or field failures with potential causes

Quality issues are the number one cause that can bankrupt a business. Taking the steps to implement proper testing with careful documentation will help you to avoid this less than fortunate outcome.

Questions or feedback? Let us know!

How To Manage Your Contract Manufacturer Part 2: Schedule

The following post is the fifth in a series of guest posts for GrabCAD.

Once the MSA is signed by both parties, it’s time to start the steps leading up to production.  A good place to start is the schedule.

dragon innovation production schedule

When creating a production schedule, consider the following:

  • Many consumer schedules are driven by Christmas (which has a fixed date).  You will most likely be able to sell many more units leading up to the holiday, but also it’s important to be realistic with your schedule and not ship a product of low quality.  Often, if you miss Christmas, you can use CES.
  • Plan contingency in your schedule. Things never go according to plan.
  • Allow several prototype cycles and enough time between cycles so you can fix issues that are found during testing.
  • Your manufacturing partner is unlikely to be as optimistic or aggressive as you (and this is good).
  • Have an onsite presence in the factory.  This is especially critical for complex or time critical products.
  • Track the schedule closely and take corrective actions early. Don’t expect to make up the time in the end.
  • In almost all cases, shipping known bad product is worse than shipping late product. Large numbers of returns can kill your business and reputation.
  • Avoid: “There’s never enough time to do it right the first time, but always enough time to do it again.”

The major milestones for production that will determine your schedule include:

  • Design for Manufacture and Assembly (DFM)
    • Project Kick-Off / Knowledge Transfer
    • Preproduction Engineering
    • Tool Release
  • Pre-Production
    • Tool Start (TS)
    • First Shots (FS)
    • Tool Mods
    • Engineering Pilot 1 (EP1)
    • Regulatory Testing
    • Engineering Pilot 2 (EP2) / Customer Use Test (CUT)
    • Final Engineering Pilot (FEP)
    • Production Pilot (PP)
  • Production
    • Production Start (PS)
    • First Inspection
    • First Batch Ex-Factory

While every run is different, on average the above steps usually require 6 months.  Prior to getting excited about making Christmas, don’t forget that the lead-in time for selecting the factory (RFQ), negotiating the MSA, and shipping your product would add up to two additional months to the schedule (i.e. for planning purposes, assume eight months from the post-crowdfunding to production start).  Shipping can add another few weeks to a month depending on whether you go via air or water.

Questions or feedback? Let us know.

How To Manage Your Contract Manufacturer – Part 1: The Contract

The following post is the fourth in a series of guest posts for GrabCAD.

How to Manage Your Contract Manufacturer

manufacturing contract

You’ve completed the RFQ process and you’ve selected your Contract Manufacturer (CM). You feel great about your choice; you’ve done your homework, you’ve met the management team and familiarized yourself with the people who will now become your own extended manufacturing team. You are now officially on the path to getting your product made.  Now you need to switch gears from picking the CM to managing them.  Here are the next steps:

Part 1: The Contract

The contract, also known as the Manufacturing Services Agreement (MSA), is a legal document spelling out the details of the arrangement with your CM. It is broken out into the following sections:

  • Manufacturing Services:
    • Defines who is responsible for doing what
    • Material Operations: Vendors, procurement, consigned materials
    • Subcontractor agreement
  • Forecasts and Purchase Orders:
    • Forecast planning
    • PO procedures, lead times, cancellation
  • Fees and Payment:
    • Pricing (what’s included/not – VAT, taxes, fees, etc.)
    • Margins/Markup (standard vs. consigned vs. “special”)
  • Shipments, Samples, Quality Testing:
    • Shipping requirements, schedule
    • Samples: quantity, timing, who pays
    • Quality Testing: Who does it, who pays for it, how it is done
    • How Epidemic failures will be dealt with
    • Audit: Record keeping
    • Final Inspection Procedure
  • Product Acceptance & Warranty:
    • You can return product that doesn’t meet the specified warranty
    • Warranties: Professionally made, original work, tested as specified, manufactured as specified.
    • Clearly spell out that you’re allowed to visit and audit at any time
  • Representations & Covenants:
    • This is a legal document, no conflicts of interest
    • Manufacturing in a safe environment, no child labor, ethical, etc.
  • Intellectual Property Ownership:
    • Definitions
    • Who owns what
  • Confidentiality:
    • Mutual Non-Disclosure Agreement
  • Term & Termination:
    • Defines the term of the agreement
    • How you terminate the agreement
  • Indemnification & Liability Limitation:
    • General protection for everyone involved & related procedures
  • Miscellaneous:
    • Terms of agreement, use of name
    • Successors / assignment – new agreement
    • Controlling law, jurisdiction, venue

You should not be surprised if the CM pushes back on certain sections of the MSA as part of the negotiation process.  All CMs are different and will have different requirements but in general, take into consideration their position and find an agreement that’s fair for everyone. The MSA process typically takes two to four weeks to complete. Finally, while MSAs are legal documents and considered quite common, there are many permutations so we strongly recommend having legal representation.

Questions or feedback? Let us know.

Product Cost Too High? 5 Areas to Focus On.

The following post is the third in a series of guest posts for GrabCAD.

When a product cost is too expensive, it makes it impossible to grow and sustain your business. We recommend honing in on five key areas, beginning with the easy stuff and moving into the more invasive:


  • The Packout: Look at what you’re including with your base model. Do you really need to include all of the accessories? Can some be removed and sold separately?
  • The Box: Look at the box/packaging itself, and its percentage of the overall Ex-Factory (XF) cost. Can this cost be reduced, keeping in mind the box’s main purpose it to protect the product and market at retail? If the product is being sold direct from the web, it should be simplified even more so as it won’t need fancy (and expensive) four color printing because the visual marketing is not required.
  • Labor and Markup (i.e., factory profit): Are these values in-line with industry norms? Can they be negotiated down while carefully balancing the relationship with the factory? (The factory needs to make a profit in order to want to continue manufacturing the product).
  • Follow the Pareto Model: Usually a product cost follows a Pareto model where 80% of the cost is driven by 20% of the components. In the BOM, sort by price, and look at the most expensive components.  Can any of these items (usually the electrical components) be reduced through lower cost drop-in alternatives?
  • Overall Design: Take a step back and review the overall design. Can metal be changed to plastic without a loss of performance or perceived value? Can you use discrete components rather than a module?  This will incur extra NRE and compliance testing charges, but because these are fixed prices, they can be offset with higher volumes.

In all likelihood, once these areas have been reviewed and changes have been implemented, your product cost should be significantly lower.

Design for Manufacturing Course Lecture 1: How To Select A Factory / Request for Quote Process

We recently announced that Dragon Innovation’s CEO, Scott N. Miller, would be releasing two courses: Design for Manufacturing and Mechanical Design, both of which he taught as an adjunct professor at Olin College. Following please find the first lecture for the Design for Manufacturing course, which focuses on how to select a factory and the request for quote process.

The complete catalog of videos and accompanying slides can be found in the Design for Manufacturing (DFM) course page. Please send along any feedback you have on the courses.

How to Select a Factory Part 2: Request for Quote / Contract Manufacturer Selection

The following post is the second in a series of guest posts for GrabCAD.

how to select a factory

Selecting the right Contract Manufacturer (CM) is the most important manufacturing decision you’ll make. It is critical to your success and you get one opportunity to do it right.  Here’s an overview of how to approach this process.

11 Steps for Selecting a Contract Manufacturer

1) Ask for recommendations

Talk with companies in your trusted network who have manufacturing experience to see what factories they have had a good experience with in the past.

2) Make a list

Use the information you get from your conversations with peers to generate a differentiated list of 5-10 CMs.

3) Create an RFQ

The Request for Quote package (RFQ) provides the CM with sufficient info to understand the product and provide a quote for the cost of goods sold (COGS), tooling, schedule, etc.  There is always a balancing act, as you don’t want to provide so much info that your IP is overly exposed.

4) Provide the right details in your RFQ

Overview (Word Doc)

  • Team overview
  • Funding
  • Product description
  • Status
  • Pending work
  • Factory selection criteria

BOM (Excel Doc)

  • Canonical format
  • Transparent and Formula Driven
  • Separate Std, Special and Consigned Margins
  • Include all costs to Ex-Factory
  • Fill in the blanks format


  • Ganttt chart (Cloud)
  • Fill in the blanks format

5) Narrow down your list

Down select to 3 to 5 factories based on your product category.

6) Get NDAs in place and release the RFQ

7) Visit each factory

The best time to visit is roughly one week after sending the RFQ. They have had a chance to review the documents and create questions.

8) Talk to references

Ask the CMs for intros to their customers to reference check.

9) Create a CM Decision Matrix

Base your decision matrix on the following categories:

  • Margins (Std, Special, Consigned)
  • Labor
  • Pareto of the top 5 most expensive components
  • Schedule
  • Fit criteria

10) Negotiate with the finalists

Narrow your list down to the top 2 to 3 CMs and negotiate the biggest line items.

11) Select the winner

After picking a CM, be sure to leave the others on good terms.

And with that, you are officially on the road to pre-production!

Questions or feedback? Let us know!


How To Select A Factory Part 1: China or US?

The following post is the first in a series of guest posts for GrabCAD.

Dragon Innovation often gets the question, “Should I manufacture in the US or China?” While every product and team is different, there are general guidelines to help figure out what makes the most sense.

First, we’ll cover the current landscape of China. It’s well known that labor costs are significantly lower than in the US, but often what is not considered is the expert domain knowledge coupled with a local, robust and highly concentrated supply chain.  The infrastructure, from the manufacturing to transportation, is second to none.  On the negative side, China is far away, and also has potential language and cultural barriers.

China may be a good fit if your product:

  • Is COGS (Cost of Goods Sold) sensitive
  • Requires significant manual labor to manufacture and assemble
  • Is built in high production volumes (>10,000 units)
  • Uses existing manufacturing technology (i.e., injection molding)
  • Is tolerant to supply chain and logistics lead times due to the distance
  • Can benefit from low or no cost Design for Manufacturing Assembly (DFMA) & Non-recurring Engineering (NRE)
  • Has “physically” protectable IP

Keep in mind that the labor rate continues to increase approximately 20% every year.  Hopefully, the US and China will maintain a friendly political relationship, and tariffs will not be a growing issue.

On the other side of the coin, the US may be a good fit for manufacturing if your product:

  • Can be manufactured and assembled through automation
  • Is sensitive to shipping time and expenses  (i.e., it needs to be delivered shortly after the order and the forecast is not well known, or the product is very large, such that shipping would be expensive)
  • Requires low volumes (<10,000 units)
  • Has difficult-to-protect IP, such as a fully mechanical product with no software
  • Requires non-standard capital equipment, such as a laser welder
  • Is part of a government contract regulated under ITAR

There are some exciting initiatives in the US to expand manufacturing.  Stay tuned!

The State of Wearables: UK Wearable Tech Conference

On March 18, Scott attended the UK Wearable Tech Conference where he sat on a panel moderated by Shoplocket’s Katherine Hague (who did a great job moderating – not an easy task!). Following are the questions that Katherine prepared for the panel accompanied by Scott’s notes. While not quite the same as attending the conference in person, it provides a nice overview of both the content covered by the panel as well as the state of Wearable Tech today and where we think it’s headed.

uk wearable tech conference

KH: What trends do you see emerging for 2014 in wearable tech?

SNM: The integration of design and technology, where the technology becomes invisible and style moves front and center.  Simple, discrete and elegant will not only be features, but requirements.  We are seeing significant activity in subtle notifications.

KH: What’s your best advice to aspiring entrepreneurs who want to start a hardware company?

SNM: That’s awesome!  Build a prototype quickly and get it in the hands of potential customers for their feedback. Iterate quickly.  Consider crowdfunding only if you’re fully prepared on manufacturing, marketing and IP fronts.

KH: What qualities would you say successful wearable products have in common?

SNM:  Design, wearability, and problem solving.  They have to look and feel good while offering a useful solution to a problem.

KH: How would you describe to your parents what you do? (A fun question – which sheds light on how far we’ve come from our parents’ generation).

SNM: My parents have been supportive of my hardware addiction from very early on. It may be a stretch to call them hip, but they certainly understand what we are trying to accomplish: Dragon helps hardware companies go from a functional prototype to a successful company.

KH: What’s the biggest mistake you’ve seen wearable tech companies make, and how can others avoid it?

SNM: The biggest and most common mistake is under estimating how hard hardware actually is.  Wearables take these challenges and push hardware to the limits.  I think of them as the  “Olympics of Hardware” for the following reasons:

  • Extremely compact “packaging” (not the box that protects and sells the unit, but the available volume for the housing, processor, battery, radio and other components).
  • Low power consumption.  Many devices must be worn for a week at a time between charges for a reasonable customer experience.
  • Use of flexible plastics, cloth and other non-traditional tech materials.
  • Waterproof to survive the rigors of life on a mammal.
  • Aesthetics across a wide range of personal preferences.  Often, integration with jewelry, which comes from a very different background.
  • Sufficient robustness to survive day-to-day activities.
  • Material compatible across all skin types and sensitivities.

As a result, wearables are often pushing at the cutting edge of technology.

KH: Many of the entrepreneurs I talk to swear by pre-orders and crowdfunding — does it ever make sense not to crowdfund a hardware product today?

SNM:  Absolutely – if you have not gotten to a stage where the prototype is ready to scale or done the necessary preparations. The comparison I like to make is that crowdfunding is like a power tool: when used properly, you can accomplish much more work than doing something by hand.  However, if used improperly, you can cause significant damage to yourself and your product.

KH: What new wearable company are you most excited about right now?

SNMPavlok, which is a habit-forming bracelet that closes the loop on behavior modification.  It’s a personal Drill Sergeant.  I like to think of it as a “Wearable with teeth.”

KH: We’re starting to see some convergence between fashion and wearable technology – do you think there will still be a distinction in, say 10 years?

SNM: With the direction that we’re headed, I think the term “wearable tech” will become somewhat passé. Tech will be the enabler (lower cost and power, smaller packaging) but the design and function will be what makes people want to wear the product.

Products Should Tell a Simple Story

This article was originally a guest Post for the Wall Street Journal

The two most important lessons I’ve learned about hardware design come from my experiences at Walt Disney Imagineering R&D and iRobot.

All of the Disney experiences center around a simple story. Everything in the product (e.g. a movie) is focused on telling this story to the guest.  The same storytelling approach is now becoming true for consumer hardware products. To be successful, products must tell a simple story to the consumer. Their “voice” is expressed through design.

Hardware startups need to take the same approach. In telling the product story, they need to carve away the bulleted list of 20 features to find the core three features that tell a clean, simple and compelling story through the industrial design and user experience. When done well, the result is a product that speaks visually to consumers and makes them crave interaction with it.

The lesson from iRobot is that great industrial design is closely coupled with the product’s function. You can’t silo any parts of the development process. The Roomba is a great example of htat. The key to our success was that the design of Roomba was so tightly coupled with its function. For the Roomba to complete its cleaning mission, it was critical that it never became trapped. This requirement dictated a design based around a round robot with the wheels on the diameter. The clearance under a typical couch dictated the robot’s height. Using these functional requirements, the engineering and design teams were constantly iterating to create a functional product that looks good.

But design also needs to be manufacturable in volume. It can be a very painful journey from a cost, time and emotional perspective when a company falls in love with a design that can’t be built whether due to expense or mere impossibility in constructs.

The key to being innovative, yet on schedule is leaving enough room in the schedule to accommodate the required debugging and quality testing. The best results come from close communication between the designer and the manufacturer so both can push each other productively.

Additionally, an often overlooked but important piece of the design puzzle is packaging. The out-of-the-box experience sets the tone for the overall customer experience.  At the same time that it delights the customer, the packaging must also protect the product during shipping.  Consideration towards the environment and cost must be factored into the equation.

Finally, hardware products on crowdfunding sites have to think about design a bit differently. Since the consumer can’t actually hold the product being crowdfunded, the product team needs to create a compelling design that makes consumers lust after the product by utilizing the channels crowdfunding entrepreneurs have available to them; namely product images and video. Yet, great care still must be taken to ensure the team has done their homework and not enticed backers with a stunning design that is not practical to manufacture.

Building A Hardware Company: Four Ingredients For Success

Since the inception of Dragon Innovation in 2009, the core mission of helping hardware companies succeed has guided us. From a high level, there are four critical ingredients for success:

  • Great product
  • Great team
  • People to buy the product
  • Capital to scale

We started by focusing on our clients’ products by serving as an API for manufacturing, helping companies navigate from a functional prototype to high volume manufacturing. An example is Pebble, who had run a very successful crowdfunding campaign and then needed to rapidly scale up manufacturing.

Along the way, there was a natural evolution where we would often assist our clients with building out their teams by using our network to make high value connections ranging from the placement of electrical engineers to VPs of Operations.

Last fall, we launched Dragon Crowdfunding to help connect entrepreneurs with passionate customers, eager to buy their products.

Today, I’m thrilled to announce the Dragon Million Dollar Challenge. For companies that raise or exceed $1 Million on their Dragon crowdfunding campaign, Dragon Innovation will write the entrepreneurs a $100,000 check in the form of a convertible note.

As we know from experience, capital is critical for scaling a hardware company. By being able to contribute to high potential companies immediately after their Dragon Crowdfunding success, we can give the entrepreneurs a safety cushion to hire additional people, cover any unforeseen issues, and aggressively scale their companies. Having Dragon in the Cap Table also signals to investors that the team is on track to deliver, and has a clear path to get product into the customers hands.

By supporting the four main pillars to a hardware company’s success, Dragon Innovation continues to push on our core mission while at the same time enabling the Hardware Revolution. And for us, this is only the beginning.