Picture the scene. It's 49 years ago -- May of 1961. The Cold War is threatening to thaw. President Kennedy addresses a special, joint session of Congress. He's about to call for the most ambitious project of its time. And he says...
Our mission is to become the international leader in the space industry, using our capacity for technological innovation to build a bridge towards humanity's future.
Say what? Oh sure, it expresses a goal and a direction and such. But, what do you do with that sentence? Let's try again. What he really said was...
I believe that this nation should commit itself to achieving the goal, before this decade is out, of landing a man on the moon and returning him safely to the Earth.
That's much better. It's simple, and concrete. We can all put our heads around what that means and start making plans. But more than being understandable, the idea sticks.
Some ideas are stickier than others. Urban legends seem to spread across the globe, die, morph and return every few weeks. Other ideas can get all the press you like, and have no one remember them the next day. They just don't stick. In their book, Made To Stick, Dan and Chip Heath explore what it takes to make an idea stick.
One of the things that they uncover is a notion they call the Curse of Knowledge. Once you know something, once you really understand a thing, it is very difficult to imagine not knowing it. Learn something else that builds on this first thing -- now it gets really tough because you've got to remember not knowing both of these things. The effect of the Curse of Knowledge is that the more expertise a person has on a topic, the more difficult it becomes for them to communicate to someone who isn't an expert.
Have you ever watched over someone's shoulder and cringed when they double-clicked on a web link? Now, try to explain to that same person when to click and when to double-click. Unless you are a really exceptional communicator, you'll run into the effect that I'm talking about in short order. (Me? I gave up on that one ages ago. I just cringe and keep my mouth shut.)
Let's try another example, that's a little closer to Performance.
The frequency distribution graph of the response times for the transaction in question exhibits a bimodal distribution. After the change, the 95%-ile response time dropped from 30 seconds to 26.
Gee. That's nice. So what?
One of the great myths of technical presentations is that a presentation that is easy to understand was easy to create. I find it is just the opposite. So, how do we work around the Curse of Knowledge? The curse happens, in part, because of the abstractions that experts use to communicate ideas quickly. To combat this, we use just a couple of easy rules.
- Keep It Simple. Remember what is important to our audience and focus on that.
- Keep It Concrete. How do we make things like response times and CPU utilization concrete? Tell stories. People think in stories. We remember stories. Use that to your advantage.
Let's try describing that last graph again.
This is a histogram. The X-axis represents response times for the screen under test. The Y-axis is the percent users who saw that response time. The red curve is what response time looked like before this change, the blue curve is after the change. The shape of these curves is important, but the thing to remember how the shape changed. After the change, the whole curve moved to the left, bunching up right here. That is, after the change, more users are getting faster response from this screen.
Notice that making this more understandable does not always mean giving less information. OK, so I'm waving my hands over the details of what was going on with this screen and why the shift. That's a discussion for later in the talk. It is simple -- which is to say, it is devoid of our usual jargon. And there is enough of a story to hang the details on and to remember what is going on.
How effective is it? You tell me.
