How to avoid common risks in hardware product development (part 2/2)

In Part 1 of this post, I talked about risks related to delays in R&D and sales start. In this post, I cover things that can go sideways during production.

Production ramp-up takes longer than planned

Production phase starts with ramping up the production. Production capacity is increased from 0 to the planned weekly amount. It is very easy to assume that production start would go smoothly but that’s not usually the case. The more complex the product is, the more likely is that there are problems in the production ramp-up. It’s just a mathematical fact, and I’ll show that below.

If the ramp-up takes longer than planned, product deliveries cannot be made on time. This usually means that all customer delivery commitments cannot be met, which will impact customer relationships. Since sales revenue (customer payments) is delayed, more money is tied in inventory.

Risk 4: Testing is not sufficient

When you take shortcuts in testing, you’ll find them again down the road. If a faulty product goes to the market, the expenses are multiplied due to customer care, warranty, product swapping costs, buyer’s remorse and even as impacts to delivery partner contracts.

These things happen quite often:

  • You squeeze off bit of a testing time. Time reserved becomes insufficient and it causes missed errors which end-up to the customers.
  • Poor pre-testing lets the product forward, but the product doesn’t pass certification. This is an immediate showstopper.
  • Low number of tested prototypes don’t reveal the variance in product performance and mass manufacturing. Some issues reveal themselves only in bigger volumes.
  • Insufficient testing may cause a need to change the product when it’s already on market, which is way more expensive.
  • And worst of all, insufficient quality builds poor reputation, which will affect sales.

Risk 5: Quality of the supply chain is not good enough

The supply chain planning is started, when product requirements are clear. The product quality is built thru the whole supply chain. Here’s a simplified example what it means:

Let’s assume that:

  • Each component has 99% statistical probability being ok and 1% to be faulty
  • Each supply chain process step has 98% yield. I.e. it fails 2 times of 100

Sounds good, we have a great process and it will produce high quality?

Let’s take a simplified example and do some math (real cases are usually much more complex):

  • There’s 50 components in the product
  • 8 process steps thru the supply chain

Here’s the probability getting a good product thru the pipeline:

Things become even more challenging if one component in the chain unexpectedly has drop in quality from 99% to 70%.

Quite disturbing? It is pretty clear that you must get your supply chain right. When the supply chain is designed, it is important to secure that each part of the chain is operating properly and the processes are in order. If one part is not able to deliver the agreed quality, it causes problems in the whole chain, creates extra expenses and delay in product deliveries.

Here’s some usual things which can turn supply chain sour:

  • Selecting the cheapest or fastest suppliers.
  • Save time and money by not travelling to the suppliers site. It is highly recommended practice to visit the supplier as often as possible, especially during main builds and checkpoints. Distance and different cultures just create amazing amount of misunderstandings.
  • Not testing the supply chain with prototypes and real production with actual/final set-up.
  • Not testing the supply chain with realistic target volumes before sales start.
  • Assuming, that subcontractors only ship agreed quality.
  • Not reserving time for product, component or supply chain corrections before sales start.
  • Underestimating the need for training supply chain personnel.
  • Assuming too ambitious production volumes for the first weeks.

If the supply chain is not secured and the pipeline is not working properly, the investments in inventory and the risks in production will mount very rapidly. And worst of all, corrections in the product or the supply chain could take several months. This leads to the next risk.

Risk 6: Not able to react challenges quickly in the production ramp-up

The product is not finished when the first device comes out of the factory. The work is just about to start to ensure timely and quality deliveries. Sales start should always feel bitter sweet for R&D guy.

Opposite to many SW problems, fixing the HW component or supply chain issues could take weeks or months. And every day you delay, your supply channel keeps pumping faulty stuff into the pipeline. To counteract this, you must reserve resources available and set them ready to react quickly to inevitable problems.

Find short term fixes to the problem but don’t forget to use root cause analysis to find where’s the real problem. The corrections must be done both to the product itself or all along the supply chain to prevent the issue happening ever again. After a sufficient time, the production will stabilize and you can go to maintenance mode.

Usual mistakes in production ramp-up:

  • Faulty expectation that production would run smoothly when the first batch is out of the door.
  • Miscommunication between supply chain people and sales/marketing about the realistic speed of the ramp-up and sellable volumes
  • Slow feedback loop from customer problems to product and supply chain development. Every day piles more failing components or products in the stock.
  • No resources reserved for product or supply chain improvement after sales start.
  • Underestimation the work load in the product ramp-up.
  • Unexpected variation of quality on components or process steps.
  • Working on the symptom but forgetting or not finding the root cause what spills the problem to the supply chain in the first place
  • Number of HW variants in the ramp-up. More different configurations you have for sales start, more complex supply chain and more errors. Try to limit this to reduce risk level. Most probably you can add more variants later when the production runs already smoother.
  • Version #2 dilemma: Distribution channel partners will be unhappy if they’ve got hands full of lower quality products while you’re already manufacturing the new version. You might need to do some reimbursements.

Risk 7: Forgetting the customer care

It’s easy to forget to plan the customer care in early concepting. But it is very important to build the draft care plan already in concepting phase. It affects the actual product design. You need to decide how the product will be serviced: should it be easy to open, do you need spare parts, does the care partner need disassembly or assembly tools? Do you need to make refurbishments? All this rules what you need to design in to product and supply chain.

Some usual fails in care:

  • Assume that product is 100% quality and none of them comes back.
  • Failing to take early feedback seriously.
  • Care channel doesn’t work quickly enough as early warning system for product and supply chain improvements.
  • Not planning how the failed product will be handled. Do you swap new product to customer, repair the product close the customer or ship back to the factory for repairs?
  • No proper feedback channels for different kinds of issues such as HW problem, SW problems, feature requests etc. There’s lot of different levels of feedback and there should be optimal channel each.
  • Spare parts are not available where needed.
  • Underestimate the need of training.
  • Return channel is not fast enough: Failed units stockpile in sales channel and distributors.

The one risk to rule them all

Hardware product development ties up large amounts of money. Not only to R&D salaries but prototype builds, testing, certification, and especially to inventory. The money tied in the inventory stock of parts and finished products is a major item to monitor throughout the product development.

Risk 8: Too much money invested in the inventory

This risk pretty much brings all the previous topics together. Hardware business is a beast of its own and that’s due to inventory. Any mistake you make very easily piles up stock.

Your key components could have 8-14 weeks lead time and it means you need to commit to the component purchases many weeks before you can put the product together and sell it. Multiply the weeks with the weekly volume and component costs to get an idea how much cash you’ll need before you can bill the customer.

Here’s again a simplified example:

  • There’s 3 key components with longest lead times
  • Average cost of these key components 5€
  • Lead time 10 weeks
  • Weekly production: 500
  • Customer payment time 6 weeks

And this was the 3 key components alone. Rest of the components have their own stocking cost as well. Product hardware variants (e.g. color, configuration) add another twist here. It’s usually guesswork to make the first orders of before you have real market sales feedback.

When the mass production is being built up, it is important to understand the exact cash flow from the moment you will order to the customer paying.  Here are some usual pitfalls:

  • Unclear understanding on component lead times. Long lead time components load your stock with staggering amounts of cash.
  • Delay in sales start. Your supply chain keeps pushing new components to inventory each week.
  • Product variants with different hardware. There’s always wrong amount of each variant in stock. You have either too little or too much.
  • Faulty or low quality products in the sales channel. If there’s a failure in a shipped product, you need to reimburse the customer and you have unsellable product in your hands. If this continues, stock levels go thru the roof.
  • Slow reaction in supply chain or product improvements. Each product manufactured with failing component increases the inventory. It’s not a nice feeling to keep shipping products when you know many of them will come back.
  • Seasonal variation of volume. Big events such back-to-school, thanksgiving or Christmas holiday season usually require inventory build-up.
  • Customer actual payment cycle – do you get money back from customers as planned?

Afterword

Even though I’ve written here about risks and things that could go wrong, there are lot of great examples of products which are very successful and loved by their customers. So don’t panic, plan well and work hard. I can guarantee, that the proud feeling is worth all of it, when your product ships.

If you need help with your hardware development, please go to UnSeen Labs to see how we can help you.

Esa Alanen

 

This article has been first published as a guest blog in Hardware Startup Finland. Read also part 1/2 of the blog here.