If there is one much overused term it's BETA. Web companies have abused the term beta release to the point that it's meaningless. Slap BETA in any piece of software and you're covered for any issues and errors indefinitely. Google keeps the label on just about any online tool it releases pretty much all the time.
The idea is speed to market. Get to market early - even if you forgot your pants. (Who needs pants? You'll make a kilt out of a tablecloth later). Where you can get away with this with code, you run tremendous risks with tangible products.
For everyone who just thought "duh, of course" I can tell you most non-engineering managers don't think that way. They think BETA.
They think speed to market. They think cycle time inventory. They think recoup of cash outlays. They think delay in sales pipelines. That creates a lot of pressure to do the wrong thing. AND - it bites you in the ass in the end anyways.
Software - particularly hosted software - has the advantage of instantaneous upgrade distribution. It's just like your website. Did I misspell something in my blog? Oh, I'll just log in and save a change. This can be automatic and behind the scenes often without the customer ever knowing. Wouldn't it be nice to push the "undo" button for all of life's mistakes behind the scenes? You just can't do that with anything hardware, tangible, real.
This messy thing called Non-recurring Engineering (NRE) often needs to be changed in order to correct or change something with your product. NRE is all the molds, stamps, jigs, testing equipment, assembly guides, stencils, etc. that are used to make custom parts and speed along manufacturing assembly. You can read more about that stuff here. Then you need to run more new parts.
Additionally, you may have a fundamental flaw in engineering, third party parts or manufacturing assembly. The more complex the product, the more likely these issues will come up.
Bite #1 Recalls
A recall is when you need to take back all the product on the assembly line, in the warehouse, on store shelves and potentially in customer possession. This can be mandatory in the case of consumer safety and it can be necessary in the case of above average product defect rates.
Recalls typically occur because of an unrecognized problem with design, parts or assembly. When you are faced with a recall situation it means that your product was never tested for some particular use (a baby will stick his head where?) or product integrity issues (they used half hard brass instead of solid brass on a part!).
Recalls have all the loaded costs of administration, accounting and shipping - in reverse. You don't mind those expenses when you're pushing to market but reverse the tide and you kill you budget. Also there is the question of what to do with all the returned stuff. Can it be fixed? Can you sell it as new again once it's fixed? Many goods today are designed to be disposable so there is a good chance it's just cheaper to trash everything. You are throwing money in the literal garbage.
Bite #2 High Defect Rates
No product is perfect or perfectly produced. The idea in 6 sigma, etc. is process documentation and process improvement. I want to point out it means you actually need to do the effort over a period of time or in succession. That means for something to improve your product - it must actually go through the process of existence in reality.
Don't use shortcuts! Engineers can do battery depletion calculations or material stress software simulations. That is only useful for R&D decisions, NOT product deployment. Engineers will say they can "stress test" products. That is not real. For example, I had a arm designed and software to track battery depletion. It took 700,000 actions to deplete the battery BUT the battery never entered a "rest state". It was just always being drained. That gave me good indicators but not reality.
There is something funny about LED light bulbs that say they'll last for 22 years. Unless you have a TARDIS and go back in time - THEY ARE ONLY CLAIMS based on some sort of calculated estimate. I already have several of those that burned out after a couple years of use.
The challenge is when you have false assumptions. You, your engineers, your manufacturers and everyone else believes a whole set of "truths" about the product. When they prove false then you're in trouble.
The longer you can field test with greatest amount of feedback the lower your defect rate will go. Every company needs to budget for some defect rate. They always want it to be well below 1%. New, just released products can have high single digit defect rates. A virgin production run can be tens of percentages. SO? Do you want that high defect rate for a million pieces or 10,000 pieces?
Bite #3 Run Away Costs
Often fixes to poorly designed and built products is a reasonable cost if you have a reasonable time to work on it. It gets to be a burdensome cost will little time. This is because you didn't plan for crap gone wrong. Shame on you. Now you need to throw people at the problem. Now you need to expedite everything - tooling, new custom parts production, additional third party parts, shipping - all at a premium to standard cost. This can easily add 20% to 50% of production cost for the next round of product.
The good news is you can manage new products and here's the tools that can set you and your effort apart.
Tool #1 Be Skeptical of Everything
You need to assume all your assumptions are wrong until they are proved correct. That engineer from MIT is a hack. The manufacturer will go out of business. The parts will never be immediately available and you'll hit the spot market. Usually everyone enters into this thinking everything is rosy. That means you need to think of contingencies, resources and alternative plans upfront. Get defensive!
Tool #2 Small Runs Straight to Market
You need to produce limited product, hundreds or low thousands, and get it into real world use ASAP. You also need time to let it all rot out there. It needs to get used, abused and bloodied. It needs to be in different places, multiple hands and constant use.
Tool #3 Method for Test and Feedback
You need a mechanism for regular feedback. It can't just be if it fails. It needs to be on a regular basis. If you can't get commitment from field users on this then move the product. Feedback needs to be quick and easy. This can be online with a simple survey or a basic email reply.
Tool #4 Autopsies
Forensic examination of field tested product (working or not) is a must. You need to examine all of it. Just because something works doesn't mean it wasn't at the edge of failure. If everything is pristine and looks good then you're starting to prove your assumptions correct. Great for you!
Tool # 5 Success Hacks
Hacks to make everything faster and easier can require upfront investment but can be well worth it. Lots of products have processor chips in them. You don't have to be making a smartphone. Create a bootloader with the code that runs the electronics. Combined with some sort of external communications like wifi, Bluetooth, heck even infrared can allow you to reprogram products and work around issues. You saw this brilliantly with NASA and the Mars Rovers. Those things lasted 10 times longer than expectations but required lots and lots of hacks to reprogram how it operated.
Make temporary tools. If you're producing parts you can spend less money and less time to produce the things to make your product. The trade off is you usually can't make much of it. This can be good for plastics parts. You may get tens of thousands instead of millions of parts but it costs less money and time. If you end up needing to throw it away - so what? Now you know for sure how to built it right.
Use rapid prototypes. Most people have heard about prototype machines that can make plastic trinkets but there are machines that can make just about anything out of any material. The only real limit is size and if you can assemble pieces to make your full size product then you'll be in luck. There is limits on how finely detailed your parts can be and the materials it's made of but it can reveal a lot about your product will little money and in a short time.
Rushing to market is like running a junk yard full of dogs. You're likely to be bit. Play it smart. It can be the difference between brilliant success and bitter failure.