Voting resources, early voting, and poll worker information - VOTE. ... Adafruit is open and shipping.

What's a Fair Warranty Policy?
Moderators: adafruit_support_bill, adafruit

Forum rules
Get help, and assist others in with open source kits and running a business! Do not ask for legal advice or for consulting services in this forum, only general biz questions!
Please be positive and constructive with your questions and comments.

What's a Fair Warranty Policy?

by bigmessowires on Sun Mar 23, 2014 9:50 pm

What's a fair warranty policy for a low-volume niche electronic gizmo?

I sell some a moderately complex bit of hardware in two versions that cost $90 or $100. I designed it myself and build them myself, and I've sold about 100 since I started three months ago. In the past few weeks, I've had two or three people whose hardware stopped working sometime after the initial purchase, for unknown reasons. I never stated a warranty policy, so now it's time.

- I want to be fair to people. Some percent of random post-sale failures are unavoidable, and a warranty should cover that.
- I don't want to pay for failures caused by accidental (or intentional?) mishandling - shorting the board against some piece of metal, dropping something or spilling something on it, etc
- I don't want to spend tons of my time troubleshooting somebody's hardware, or supporting people who break multiple units due to some kind of systematic mishandling

My thought is to give a 90 day warranty, where people can return the hardware for a full refund. If something happens after 90 days, I think odds are it was due to some kind of mishandling rather than a problem "from the factory".

I'm also thinking that a full refund would be the only warranty option, and I won't send out replacements or otherwise try to troubleshoot problems. I fear that will just be a time sink-hole. However, people could take their full refund and just purchase the product again, so I can't really avoid that completely. I'm just afraid of being caught in a never-ending CS loop of problem reports and returns from a "bad" customer (if there is such a thing).

What do you think? Is my proposed warranty too stingy? Too generous? What would seem fair to you? Is 90 days reasonable? I looked for Adafruit's warranty policy, but they don't seem to have one, so I guess it's discretionary.
Posts: 69
Joined: Sun Jun 05, 2011 7:34 pm
Location: San Francisco

Re: What's a Fair Warranty Policy?

by adafruit_support_mike on Mon Mar 24, 2014 8:54 pm

Our standard return/replace window is 30 days.

Hardware failures tend to graph out on what's called a 'bathtub curve': high for a period immediately after units go into the field, dropping off to a low-but-steady level for most of the product lifespan, then rising to high again after a certain amount of time.

The initial high period is where you discover production errors and components that don't live up to spec. The long-term high period is where parts start wearing out. The low period in the middle tends to be dominated by environmental issues that you can't control.

Estimating the lengths of those three periods falls somewhere between heavy statistics and fortune telling. The number of early failures can be controlled by having a good regime of testing though. There's a story from the 1970s about a company buying electronic assemblies from a Japanese company, with a contract that specified no greater than a 5% defect rate in units delivered. When the shipment came in there was a small box off to the side with a note that read, "we don't know why you wanted 5% defective units, but we've packed them separately for your convenience."

Broadly speaking, there are two kinds of failures: systemic and random. Random failures include things like earthquakes and lightning strikes. Almost everything else is systemic, which means it's repeatable, and you can get rid of it by finding out what causes that particular failure.

It's a good idea to track every failure you hear about, and almost always cost effective to redesign your products/processes in a way that eliminates sources of systemic failure. Your testing should look for failures that occur between the rates of 1% and 0.1%. If more than 1% of units fail for the same reason, you should stop wasting time and money on testing and redesign the product. If less than 0.1% units fail, you need to decide whether it's more cost effective to test the 999 that will be fine, or to replace the 1 that fails in the field. Even then, look for a way to design the failure away and jump on it if it doesn't cost too much.

Since you probably don't have a full set of Mean Time Before Failure estimates or the field data to support them, go with your gut and start taking notes. Within a year you should have a pretty good idea of how units perform in the field, and a list of known failure modes. IMO it's best to err on the side of generosity during that period since your customers are doing your field testing for you. Once you can make quantified statements about your failure rates, you'll be able to cut off any customers whose replacement rate seems well outside the normal values.
When you void a product warranty, you give up your right to sue the manufacturer if something goes wrong and accept full responsibility for whatever happens next. And then you truly own the product.

Posts: 61443
Joined: Thu Feb 11, 2010 2:51 pm

Please be positive and constructive with your questions and comments.