Month: May 2009

This is about sandwiches. Yep, sandwiches.

A good sandwich is a work of art — where the meat is nice and lean and the tomato is ripe. Any place that makes a good sandwich is worthy of a repeat visit, especially if I can make it there and back during lunch.

When I go to a sandwich shop for the first time, I check the menu.  What kind of sandwiches do they offer?  What kind of toppings?

What kind of bread?

“We have sourdough roll, multi-grain roll, and dutch crunch.”

Unless they make their own bread, all the sandwich shops in the bay area serve the same breads.  All of them.  And they all answer that question in the exact same order: sourdough, multi-grain, and dutch crunch.

What can you learn from this?  First, be a bread maker; it’s a good business to be in.  In San Francisco, if you’re eating a sandwich, you’re probably eating Wedemeyer Bakery bread.  And remember — there’s room for more than one breadmaker in a town.

Second, if you’re not a skilled breadmaker, make a better sandwich.  Mr. Pickles is good for the speed and cost, but Bonne Sante‘s chicken and prosciutto is a sandwich that I would drive for miles to get.

And there ya go. Make bread or make a better sandwich.

It’s amazing what sandwiches have to teach us.

Did you hear the one about the Google designer who didn’t want to improve his designs with data?

The New York Times covered the story of the Google designer who defected to Twitter because Google was hellbent on testing designs to pick the best one.

Deciding which design is “best” depends on what school of design you come from. Here’s a cheat sheet for the big three — their philosophy, criteria for good design, and a phrase you might hear such a designer say.

Data-centered design

Philosophy	Test everything and let the numbers be your guide.
“Best” criteria	Crunch the numbers and find the best combinations.
Key quote	We’ve got web tracking. What do the metrics tell us?

Design driven

Philosophy	Designers know best; follow your heart and best practices.
“Best” criteria	It just feels right.
Key quote	That’s doesn’t work for me. Can we try it one more time?

User-centered design

Philosophy	You gotta find the problem before you can build the solution.
“Best” criteria	It solves their problems, and they love it.
Key quote	You can’t read their minds. Can we talk to them?

But these design philosophies don’t live in separate silos. They can easily coexist.

For example, a data-driven designer should be very interested to know why the numbers ended up as they did. To find out, you can turn to your underlying design principles (design-driven) or ask people (user-centered).

Similarly, a design-driven person should make sure the designs resound with real people (user-centered) and that the results pan out when launched (data-driven).

And a user-centered designer should look at statistics to see where problems are occurring (data-driven) and trust gut feelings when translating user feedback into feature ideas (design-driven).

There’s room for everyone in the design world. How narrow-minded do you have to be to ruthlessly test everything? Or not to trust your design instinct? Or not talk to people about their problems?

One way of coping with the reality of design is to leave your company because you couldn’t find a way to balance your design sensibilities with the company goals. *cough* doesn’t play well with others *cough*

But I think there’s no such thing as the “right” way of approaching design problems. You should always look to expand your design horizons, even if you have to venture into new disciplines or combat your own opinions.

Design is as much about your opinion (design-driven) as it is your audience’s (user-centered) and your employer’s (data-driven). The best designs come from a healthy acceptance of all three.

By the way, does anyone know which line size — 3px, 4px, or 5px — worked best? I’d love to know.

in flagrante delicto — caught in the act

Here’s a quick lesson on the importance of watching people while they use your products in the context where the products are used.

Some time ago, I was leaving the movie theater in the mall when I passed by some people using the touch screen information kiosk.  There were three or four of them huddled around the screen — talking, pointing, touching — trying to find something.

Two of them in particular were the primary drivers of the touch screen.  Frequently, they would both touch the screen at the same time in different places, leading to random results; the next page that displayed was different than what either person had touched.

Unfortunately, this screen was built to operate on a single touch.  If two people touch it at the same time in different places, it records a touch at the midpoint of those touches.  The two people at the screen that day didn’t know it but their information retrieval goal was subverted by the screen’s designers.

More recently, I was using a Coke touch screen vending machine when my nephew came over and started playing with it at the same time.  The same thing happened; the screen didn’t handle multiple touches, and the result was neither my nephew or I got what we expected.

Two lessons here.  First, single-touch screens are obsolete.  If you’re building a touch-screen device, multi-touch is the only way to go.

Second, always watch people use your products in real situations.  Unless you sat and watched that kiosk all day, you would never know that single-touch was causing so many problems.

No usability test or lab setting can reproduce the infinite variety of reality.