My great-grandfather, Joe Buck, was a no-nonsense man who did carpentry much of his life. In an interview I did with my grandfather about ten years ago, he described one experience with his father that particularly exemplified Joe’s attitude toward life.
While building the LDS Santa Clara chapel in Eugene, Oregon, he told his son to go take care of some job, and told him how to do it. My grandfather went and took care of it as instructed, and returned, declaring the job done.
Joe wanted to inspect the work first, though, so they went to the job site together and looked it over. Joe identified several things that needed improvement.
My grandfather, as any son would, said, “But Dad, it’s good enough!”
“No,” said Joe. “It’s not good enough. It’s not good enough until it’s right. Now fix it.”
Michael
on 04 Sep 12This makes me feel better about an enormous list of changes I dumped on my designer today.
Tanner Christensen
on 04 Sep 12The real struggle is knowing when enough is enough to launch though. When it comes to the web there is such a thing as good enough, for now. Continuous improvement is the name of the long-term success game, ala 37signals products.
Great reminder.
Patrick Harrington
on 04 Sep 12Reminds me of an old employer who used to say, “Build the A, not the A+.”
The problem with that is very quickly people started building the B+. And then the B…
Jamis
on 04 Sep 12@Tanner, I agree, knowing when to launch is another issue entirely. The entire reason we iterate after launch is either because it’s not quite right yet, or because the requirements have changed and “right” is now different than it was.
I think it’s also important to remember that “right” does not mean “perfect”. It’s entirely reasonable for something to be functionally correct, but with plenty of room for eventual improvement.
Michael Gowin
on 04 Sep 12On that same line of thinking (right vs. perfect), I know a photographer who tells his post-production studio staff, “It doesn’t have to be perfect; it just has to be really good.” Sage advice for any designer when considering the diminishing returns on time spent trying to achieve perfection.
sam
on 04 Sep 12Good enough, isn’t.
Benjy
on 04 Sep 12The beauty of the web is that you can launch without it being good enough, while still striving to get it to just right.
What can be frustrating on the web is when one builds, launches and does want to get it right while the others involved decide it’s more important to move on to the next thing and it remains close but not all the way right.
David Andersen
on 04 Sep 12Of course there is such as thing as ‘good enough’. You can go over the top like Steve Jobs did making the internals of a computer look good even though it served no practical purpose. As long as it was easy to assemble and service that’s good enough, otherwise it was Steve indulging his fetish. In rough carpentry there’s no reason for the level of precision that finish carpentry requires. Sure you can do it but it’s wasted time and money.
Jamis
on 04 Sep 12@David, I never said there wasn’t such thing as “good enough”. :) All that was said was that “good enough” is only good enough when it’s doing what it’s supposed to do, i.e. when it was “right”.
Obsessing over perfection is the danger, but so is the opposite.
MR
on 04 Sep 12@Michael — Hopefully you shared with your designer why your changes make things right before dumping said list. ;)
Joe S
on 04 Sep 12Reminds me of when I worked fro an electrician the summer before I went to college. He asked me how some work I’d done was… and I truthfully said “its halfway decent.”
His response: “Halfway decent is the same as halfway shitty”.
Michael
on 05 Sep 12Mig, I tried to! I think I’ve still failed for not communicating that earlier in the project.
RDO
on 05 Sep 12This reminds me of a blog post I wrote couple weeks back:
http://ruben.orduz.info/2012/08/a-little-south-of-perfection-idealism.html
Anonymous Coward
on 05 Sep 12I like this post and this is a classic quality and work ethic lesson. However, I think it also is adverse to the MVP – minimally viable product – idea that 37 also promotes – that you can keep tweaking and adding features in a more perfectionistic way – you need to ship the software ASAP and then tweak. Maybe both of these ideas can live in harmony somehow
Don Schenck
on 05 Sep 12Damn, folks, it’s not rocket surgery!
Build, ship, improve. Sheesh.
Jamis
on 10 Sep 12This doesn’t contradict the idea of a “minimally viable product” at all. “Right” can have a more granular meaning; a minimally viable product is “right” if it accomplishes the task it is intended to, in a way that you feel good about.
My grandfather didn’t say exactly what the problems were that his father identified in his work; they may have been things we’d roll our eyes at, or they may have been things that would have caused the structure to fail down the road. The point, though, is not what specifically was wrong, but that the craftsman wasn’t satisfied. This happens at 37signals often enough, too—a programmer or designer will take a stab at something, and feedback will come, essentially, “it’s not right yet, fix it.”
This discussion is closed.