I mentioned in my last post that I use a game to coach my teams on estimating. I use games for many things actually. I blame my past life as a Cub Scout leader for teaching me how to make up games on the fly. Feel free to steal & modify my idea if it makes sense for you.
Story Points don’t inherently make sense to people. I was met with no small amount of resistance when I first mentioned using them recently. Like many people I decided to start with something that does make sense and we talked about t-shirt sizes. Easy. Everyone has worn a t-shirt at some point.
We started with a simple scale of sizes (S, M, L, XL, 2X) and some ground rules. The team has 30 seconds per estimate to decide how big something is, and a consensus is not required. Majority opinion is fine.
First up: I ask my teams to tell me what MY t-shirt size is. My actual size is irrelevant. I don’t care if they decide I am wearing a shirt the size of a circus tent. Having a thick skin and a sense of humor is an important skill as a Scrum Master, by the way. I write my name on the board beneath whatever size they give me.
The important thing is that I am now officially a reference story. And that reference story is something they know, not an intangible something. Now that I am established, the real game can begin. My mind works in silly ways, so I latched onto the term “relative sizing” and decided to size my relatives…
With that in mind, I show the team a photo of myself with my younger son. What’s *his* t-shirt size? My son is taller and leaner than I am. A lot taller and leaner, actually, because he’s young and in shape. The photo I use was taken the day he graduated from US Army Basic Training, in fact. Does his added height put him in a larger shirt, or does the fact that he doesn’t have the dad bod put him in a smaller size? Once again… the truth doesn’t really matter, whatever the team decides is perfectly correct. His name is now added to the board.
Next up? Here’s another picture of my son with two of his best friends. His friends, of course, are not close to the same size he is. One is far shorter, and one carries more muscle mass. Because chances are the team has different sizes for my son and me, this now gives them two points of reference going forward. Names added once again.
One more picture. My son with my sister and my brother-in-law. My sister is several inches shorter than I am, and my brother-in-law is built like a college football lineman (which, in fact, he was). By now, the estimates are all over the map, and we have a set of six names on the board that have been sized.
It’s now time to blow up the scale. To this mix of people I add two more outside my personal circle of family & friends:
- Peter Dinklage
- Andre the Giant
Every time the photo of Andre appears when I run this game, the room is laughing. How do you possibly guess what size shirt Andre the Giant wore? And yet, we find a way to expand the top end of the scale and arrive at an answer.
The scale of sizes has now changed from where we started. We often have an XS at the bottom, and a XXXXL (the actual number of Xs may vary) at the top. At this point, I start putting story point numbers alongside the t-shirt sizes (1,2,3,5,8,13,20,40,100) and the teams can begin to see how just thinking about the size of something translates easily into story points.
The beauty in all of it is that none of the estimates need to be correct. In fact, they usually are way off, but that’s totally okay. The teams’ perception of the size of the work is all that counts, and those estimates are good enough. What one team decides will almost certainly not align with what another team decides, and that’s perfectly great too!
The same rule applies when we start estimating our actual User Stories. We don’t need to be perfect, we just need to be close enough, and in agreement on what a certain number of points means, relatively speaking.