Timing: 30 mins


  • 20-30 balloons per team
  • Supplies for each team: construction paper, rulers,
    scissors, markers


Start by showing the teams a balloon that you would like created (or draw one).

Test Balloon with Acceptance Criteria
Test Balloon with Acceptance Criteria

The balloon has a face made up of two round eyes, a triangular nose, and a semi
circle mouth. Without any further instructions, tell the teams that they have 2
minutes to create as many of the balloons as possible, then have them bring the
balloons up to be ‘accepted’. Eliminate any balloons that do not meet your criteria
of ~10 inches wide, ~2 inch eyes, ~1 inch gap between eyes, ~1.5 inch high nose,
and ~4.5 inch wide mouth. Very few teams will have balloons that meet the criteria.
As you reject their work (waste), ask the teams if they’ve ever had a similar experience
in software development. Before the second round, give the teams 2 minutes to discuss
how they can improve for the next iteration. They should start asking more questions
about the acceptance criteria, which you will happily offer. When round 2 starts,
the teams will now apply the acceptance criteria to their work and some will even
start building ‘test harnesses’ (e.g. paper templates for face, quick ways to measure
balloon width, etc.) . The results should be better in round 2. Discuss how they
changed the way they worked and what improvements they would make the next time.
If needed, play one more round. This time, every team should be using a test harness
and should therefore be producing balloons with much more efficiency and quality.

Learning Points:

  • Defining acceptance criteria is not the same as writing
    tests, only to be applied after something is produced. They can be used as requirements,
    as tests, and as a target for developers.
  • Automating acceptance tests (or executable requirements)
    can be very useful, as demonstrated by the test harnesses produced during the game.
  • The investment in creating and automating acceptance
    tests is worthwhile and has a high return.



21 thoughts on “99 Test Balllons”

  1. If I am understanding correctly.

    This exercise will help team to unveil scope by defining DoD & AC.

    We can run this for two iteration and help them learning Groomed Stories, Planned Stories and Complete Points.
    As well as Velocity.

    We can ask them to have Face of this kind.

  2. Hi Eric,

    I usually just get my balloons at Walmart: http://www.walmart.com/ip/12-8221-Royal-Blue-Balloons-10ct/35775038
    You can find them in the party section. They can blow up to 12” and come in all sorts of colors. I like giving 1 or 2 packs of the same color per team so they can rally around their color. 🙂
    And teams can blow up the balloons before putting the face on them. But I don’t tell them how to work, just what I want as a customer. They figure that out pretty quickly. Especially after I flatly reject the ones that were blown up after the face was drawn on. Not because of the process that was used, but because it just looks ugly.


  3. Hi. Thought I had posted my comment over a week ago, but must not have saved. Where can I find balloons that are this large? I checked a huge local party store, Walmart, Farm Fleet. Are the balloons supposed to be blown up all the way? To draw on a flat deflated balloon is hard enough. Introduce the curvature of the blown-up balloon, and even the template becomes a major difficulty. Maybe I’m just missing the obvious somewhere. Please reply soon, as I’d like to use this for a client engagement next week. Thanks, Eric.

  4. I don’t see the point of how this game is relate to testing… if there’s they are fairly weak points.
    I tested this with my team, all proud Indian, the first question being asked is, what’s the measurements?
    I evaded by pointing at the balloon printout and my team walks up with a paper and trace the measurements.
    I guess I just have smarter people in my country.

  5. We played this game with scoring: +1 point for valid balloon, -1 point for invalid balloon. Lots of negative scores after the first round.
    Obvious learning point is that sometimes it is better to do nothing than to do the wrong thing.

  6. I tried this game with a room of 60 people (15 teams of 4). Quite a challenge and would only recommend it if you have one or two helpers to answer questions along the way. Even if instructions are crystal clear, some people will raise their hands. Because of the noise and excitement, it’s hard to manage such a big group.

    To make it more interesting, I added a scoreboard. At iteration 1, I make all teams fail by rejecting all balloons unless one of them already figured it out. Helpers can also count the balloons so the facilitator doesn’t have to run around. Between iteration 1 and 2, I make sure that teams get the trick. I then go for another run and see who wins. No prizes. Just pride for the winning team.

    I still wonder how the hotel staff manage the 500+ balloons that were lying on the ground 😉

  7. **Instead of slides –> Draw on whiteboard and evolve**

    Draw a balloon and smiley on a whiteboard
    1. Draw a balloon and smiley on a whiteboard
    2. “I want this”
    3. Iteration 1
    4. After Iteration 1 you have a clear picture and know what you want a bit more
    5. Start specifying with the people (adding inch/cm)
    6. Iteration 2
    7. …

    Instead of preparing everything build it up from scratch, which might be more real life…?

    Did someone ever try that variation?
    Is it worth doing? Different learning points?

  8. During the 99 balloon game we had the idea of doing the game differently:
    Something along those lines
    1.Give them tools (paper, scissors, balllons and pens)
    2.Don’t tell them about the balloons.
    3.Tell them you want to have this face (don’t mention the balloon)
    4.The people start to use the balloon ( but you didn’t actually say anything about the balloon)
    5.After 5minutes stop and ask: Show me your product
    –> I didn’t want a balloon. I want a round smiley face

    Learning: People tend to use the tools they have or that they know

    Does a game like this already exist?

  9. I think I will try this with my Team next Week. It really sounds very interesting for me, because it seems like some of them are on a very slow motivation Level….

  10. Hi Robb,

    I’ve been meaning to post something about this…
    I first used to just rush participants in to the first iteration.
    However, after getting more and more confidence with this exercise, I now just ask them if they have any questions and approach it like a true customer would.
    Common questions and my answers:
    Q: What color? A: I don’t care. I don’t see color.
    Q: How big should the balloons be? A: About this big (I hold my hands out to show an approximate size – a customer won’t know exact measurements, just what feels right.)
    Q: Does it have to be on an actual balloon? (thinking I’m trying to trick them) A: Yes
    Q: Does it have to look like the picture? A: Yes, like that.
    Any questions referring to precise size, I just point to the face I drew. In the debrief after round 1, I ask the audience if they think I, as a customer, should have known the exact sizes off the top of my head. The answer is obvious – no.

    I’ve found the questions naturally taper off and people start playing around with their materials. So I say “Ok, let’s try a round and I’ll come back to check on you in a minute.”. Even with the questions answered, they still are all over the place with their balloons. In fact, I don’t even do all the rejections after round 1. I ask the rest of the audience, who reject them based on their own ‘feel’.

    Btw, I’ve started doing smaller iterations (60-90 sec). Especially for larger crowds.

  11. I did this with my team, great results.

    Iteration 1
    Team1 – velocity = 1
    Team2 – velocity = 0

    Iteraiton 2
    Team1 – velocity = 12
    Team2 – velocity = 7

    Team 1 used a template as a test harness and got much better quality balloons for it. No one asked for acceptance criteria up front, they just got stuck in blowing up loads of balloons =)

    We also did a ‘how many balloons can you pop’ iteration to clean up afterwards.

  12. Tried this with my teams – a lot of fun, but there were questions of acceptance right up front. So, it’s good to have a strategy in mind to deal with that beforehand.

Comments are closed.