A Kobo Parable

I love cof­fee. Always have, as far back as I can remem­ber actu­ally know­ing what cof­fee tastes like. So when I was look­ing for a job dur­ing a break from school, it was only nat­ural that I should apply at Starbucks. I did and was hired. Actually, I worked at Starbucks sev­eral times, as well as a few other cof­fee shops, but this story takes place at the Starbucks store at the Perimeter Pointe shop­ping cen­ter in North Atlanta.

The last time I did a stint at Starbucks was the sec­ond half of 2001, and the man­ager of our store was a woman named Sherri. She was nice-ish, but the two of us had prob­lems get­ting along. Though we both had good inten­tions, we seemed to go about every­thing in totally dif­fer­ent ways. Of course, my way was infi­nitely superior…

At least my drinks were bet­ter, and this I know because reg­u­lar cus­tomers and other staffers often told me so. The thing was, nobody seemed to under­stand why. I would often ask them how they could tell the dif­fer­ence. We both used the exact same recipe for every drink on the menu. Even my cowork­ers were hard-pressed to put their fin­gers on exactly what is was about Sherri’s drinks that didn’t taste right. The answer was obvi­ous, once you thought about it.

Sherri and I were the two fastest drink mak­ers in the store, and we both worked morn­ings — the busiest shift. When you have a line that stretches to the side­walk, you always want the fastest per­son pos­si­ble to be mak­ing the drinks, because that’s the most labor-and-time-intensive part of giv­ing each “guest” the “Starbucks expe­ri­ence” (you have to use this kind of ter­mi­nol­ogy when you work for a cor­po­rate chain). To be even more pre­cise, the thing that takes the most time is steam­ing the milk for things like cap­puc­ci­nos and lattes.

This is a fairly innocu­ous step in the mak­ing of an espresso-based bev­er­age. All of the Starbucks lit­er­a­ture goes on and on about the care­ful roast­ing and spe­cial pack­ag­ing of the cof­fee beans. The cups describe the “per­fect” shot of espresso — a twenty-second pour of one ounce of 192-degree water through I for­get how much ground espresso beans and what pres­sure. Nobody ever really says much about the milk because milk is milk, and you can’t screw that up, right?

The process is actu­ally very sim­ple. Cold milk is poured into a stain­less steel pitcher. Then the pitcher is brought up under the steam wand. The tip of the wand goes a lit­tle under the sur­face of the milk, and the “barista” opens the valve to allow hot water vapor to pass through the milk. The air froths the sur­face of the milk up a bit, and then the wand is pushed down into the bot­tom of the pitcher to heat the milk evenly from the bot­tom. Once it reaches about 140 – 150 degrees, it’s con­sid­ered done, but it’s still good up to 170-something degrees, at which point it burns.

Now there are two ways to go about this. First, I’ll tell you how Sherri did it. When things were busy, Sherri didn’t want to have to wait for milk to heat up before putting it in drinks. So she decided that the most effi­cient way to do things was to fill the pitcher with as much milk as pos­si­ble and steam it until it was as hot as pos­si­ble. That way, she would have plenty for sev­eral drinks, and it would stay above 140 degrees for longer.

On the other hand, I would only steam small amounts of milk at a time, and only to the bot­tom end of the tem­per­a­ture spec­trum. It would seem that this would require more steps when mak­ing a lot of drinks, but it actu­ally requires much less time to steam a small amount of milk than it does to steam a whole pitcher, espe­cially to a lower tem­per­a­ture. Instead of putting the pitcher on the rack at full steam blast, I would actu­ally stand there and hold the pitcher while I kept the steam at a low level to heat the milk grad­u­ally. This made all the difference.

With the Sherri-method, you end up with milk that is unevenly heated. Some parts of it are nearly-scalded, and some of it is nearly-raw. This mixes together and gives a sour fla­vor to the drink. You can’t detect this right away, but the more you cook milk, the faster it goes stale as it cools. So what often hap­pens is that you buy a drink at a place like Starbucks and, by the time you get to the bot­tom half, it doesn’t taste good anymore.

However, if the milk is steamed grad­u­ally and evenly, and doesn’t ever get any­where near it’s scald­ing tem­per­a­ture, it retains its fresh fla­vor for much, much longer.

So what, if any­thing, does all this have to do with kobo? Well if you’re prac­tic­ing kobo cor­rectly, it’s a lot like grad­u­ally heat­ing and froth­ing milk. If you prac­tice the way most peo­ple do, you’re try­ing to build algo­rithms for jis­sen, and those algo­rithms can never be ade­quate prac­tice for some­thing so dynamic. What about the scald­ing milk? That’s what hap­pens when you try to learn jis­sen with­out a ratio­nal pro­gres­sion through proper use of kobo methods.