The innovative online rental company Netflix is pretty great. They send me a CD in a little sleeve, I watch it, then I send it back. We’re only on the one movie plan, so I always chuckle when I see Netflix’s elegant solution to losing an envelope or DVD sleeve.
I’m guessing most people have the two DVD plan. So, Netflix’s solution works for most people. Is that good enough?
I’m a big fan of my Mac. One of the best things about Apple is that they provide generous support via their retail stores. The other day I was calling in and received the following message:
“We regret that Genuis Bar reservations can only be made online.”
That works for me. I have a couple other computers. What about someone with a fried motherboard on their only computer? I guess they go to the library?
Again, most people probably own or have access to a second computer. Does that mean Apple should make some customers jump through hoops to get access to their otherwise great support?
This is a decision we all have to make on the Web every day. Do we support people who have disabled Javascript? What about people with small monitors? Are we alienating a percentage of our users who don’t know what “FAQ” means by using the term?
Nobody can be everything to everyone, obviously. These decisions have to be made at some point, but they need to be made consciously. Who is our audience and how much work is it worth to support the outliers? Is the good for most solution good enough?
This is definitely an interesting point, and one thing I’ve noticed is that successful companies often understand and accept that they can’t help everyone. I look at Basecamp by 37 Signals, they outright admit that they aren’t going to solve 100% of your problems, but they nail 85%. While I’m often frustrated by the things they ignore, wishing they’d solve that one problem I still have, I know that it’s probably why they’re successful. When I’ve worked on applications, I quickly find that if I try and imagine every single situation where someone might be alieanted, the project never moves forward, and I spend an extra 70% of time working on a 3% problem.
37 signals is run by a small, lean group of people, and I think they can be so lean because they know when it’s time to let things go.