Tuesday, May 31, 2011

Woodland Creature Story Sizing



I'm not a big fan of story "estimates". In my opinion, it conveys a notion of accuracy that simply isn't possible. I prefer story "sizes" instead…still not perfect, but better.

I'm also not a big fan of using numbers to size stories, again, because numbers encourage folks to do math that doesn't make sense.





So what do I recommend for sizing?

Woodland creatures!

Here is my "normal" list:
- Shrew
- Squirrel
- Badger
- Boar
- Deer
- Bear
- Elk
- Caribou
- Dragon*



The reason I like woodland creatures is:
1) you can't do math with them (though they do multiply!)
2) they imply a size or category, but not an accuracy, and
3) they keep the team from taking the activity too seriously, but differences in opinion still get discussed.

Personally, I've only ever made cards and used them for various planning activities. Others have suggested finding various animal stickers and sticking them right on the card (I like this idea but have never gotten the stickers). The list of animals obviously doesn't matter. They simply need to go from small to large if used for sizes, or any animals if you're simply assigning a category or class of work.

*Note: You may have noticed the last animal in the list above was a dragon, and might be thinking, "wait a sec! dragons aren't woodland creatures!"…and you would be right :) However, I usually reserve this special animal for stories the team think needs to be broken down further. Then we either assign a "dragon slayer" to ensure the story gets broken down, or we simply avoid it for now.

My preferred planning activity is "Creature Sizing". This is basically T-shirt sizing; assigning a category of work to a story. Select a handful of woodland creatures to use. Normally three is fine to start off with (I use Shrew, Badger, and Bear). Next, a subset of the team (about 3 members) is asked to discuss key points of the stories. If another team member with special knowledge is needed they are "tagged in" at the last moment. These team members then assign a woodland creature to the story. If the team is less mature or lacking confidence then the whole team could be used for this activity.

I should also mention, that when possible, I like the categories of work (and therefore number/types of animals) to emerge over time instead of selecting them up front. This isn't always possible if an entire initiative is needing to be relatively sized for a governance decision. Also worth mentioning is that this level of planning can be continuous, with little disruption to the team.

I've also used the woodland creatures with Planning Poker (my least favorite sizing game) and 1-N Table Walking (which I like better than Planning Poker, but still sub-optimal/less mature in my opinion).

So there is it! If you think your team or your product owners are getting too hung up on the "accuracy" of your "estimates" or start doing "crazy velocity math", consider using Woodland Creatures as an abstraction that may help :)

12 comments:

  1. I was on a team a few years back that used frogs. Sadly, the owner of the company could not get past the desire to equate estimates to hours and he eventually created his own fictitious mapping. Imagine the team's surprise when we were informed that a bullfrog should take no more than 4 days to complete or it was clearly a Goliath.

    ReplyDelete
  2. So, I would suggest that in this case, the team was better off 'cause at least they got a good laugh out of the owner taking frog-based categories so seriously :D

    Perhaps using cycle-times could help alleviate this? It would provide an actual average that at least wasn't completely made up…of course then that can cause other issues :/

    ReplyDelete
  3. This is absolutely adorable and cracked me up! Sadly, I don't know the size difference between a Shrew and a Badger, and this is pretty upsetting to me. I must go into the woods and sort this out. Heaven help the person going out to catch mad badgers due to your coaching. :)

    Well. People love to predict and manage. I think that there is a business culture issue that runs so deep that it is hard to overcome. When you ask those who are "paying" employees for time or expertise to trust what they are getting without them micromanaging it, that is scary. Then there is the idea that if they don't truly understand what is going on in detail, that maybe they aren't a good manager or good business owner. It's really a difficult cultural challenge. I wish I knew a way to keep these people calm, or distracted, or at least useful in some way without disrespecting, disregarding their feelings, or not being understanding. I think that even woodland creatures may not be abstract enough, but I love that you are adding fun.

    One of the happiest things about my assignment at PG was a person who told me that I taught them about "clownery". Indeed. This is one of the things I most want to convey wherever I go. Fun and celebration aren't optional. They are a vital part of having a healthy team that others want to work on.

    ReplyDelete
  4. Love the post, especially the dice pictured. You can even have some fun with character of the animals:
    - shrews are so small they can slip in around the edges and hardly be noticed.
    - assigning T-shirt sizes (XS=Shrew ...XX=Bear) works for me, and having bear sized story seem completely appropriate ("That one will be a BEAR").

    With a straight face now, two things.
    1. I suggest the menagerie has too many woodland creatures to be useful. This is after all estimating and should be quick.
    2. Oops; Caribou are smaller than Elk, per top weight ranges on Wikipedia. Elk: 500-750 lbs, Caribou: 200-460 lbs. What a drag.

    Cards with woodland creatures names, picture, and characteristics would certainly be a lot more entertaining than your standard planning poker deck.

    ReplyDelete
  5. @CuriousAgilist --

    Huh…good point on the Caribou. Perhaps I should swap it out for a Moose…those things are HUGE! :)

    As for the quantity, I mentioned for T-shirt sizing I like to select three (Shrew, Badger & Bear). I use the full gradient of the menagerie for the 1-N table walking game. The gist there is that all unsized stories get ordered relatively to each other from smallest to greatest and then grouped into sizing piles. I've found that the menagerie provides useful context for that game.

    Thanks for the input though :)

    ReplyDelete
  6. Cute.

    so how do you figure out from these animals when you're releasing and what will be in it?

    ReplyDelete
  7. @James Peckham --
    Using the animals as categories of work or sizes, one would measure cycle time per story (hours/days from start to done), and then average cycle time per category/animal.

    Determining your release plan is simply sizing enough stories and doing the math (avg cycletime * # of stories in a category / WIP bottleneck or avg WIP). I use a similar method for "epics" or large chunks/animals. Again, this forecast is not "accurate" but is to inform a general idea. You get more accurate forecasts with smaller amounts of time.

    If you need to be using iterations and a "velocity" mesure instead of flow and cycletimes, I have a method for doing that too, but it gets a bit hairy.

    Feel free to hit me up on Twitter or mgbarcomb at gmail dot com if you have specific questions though :)

    Cheers,
    -Matt

    ReplyDelete
  8. "If you need to be using iterations and a "velocity" mesure instead of flow and cycletimes, I have a method for doing that too, but it gets a bit hairy." -- Yes.

    ReplyDelete
  9. I have a similar sizing method:
    ant, mouse, cat, dog, horse, elephant, whale

    Might swap Dragon for the Whale though, as the imagery associated to "slaying the dragon" feels like it could be useful.

    Do you have any advice on how to nudge people away from obsessing over whether a story is 2 days or 3 days of effort? I've found that the arguments over those definitions become very circular and wasteful.

    cheers,

    Paul Hodgson

    ReplyDelete
  10. I've got a set of animals to make the sizing point as well. We can all thank Mike Cohn for the the original dog sizes. The ambiguity in Mike's set (e.g. poodle) makes his the best I've seen.

    I do love the dragon and dragon slayers.

    James

    ReplyDelete
  11. I look forward to having number fixation as a problem worthy of our attention in our Scrum implementation. Respect! :-)

    ReplyDelete
  12. We read this post about a year ago. 12 months on we are still slaying dragons! :-)

    Thanks for sharing. I've written a few notes about our experiences here: http://workingwithdevs.com/woodland-creatures/

    ReplyDelete