Monday, September 3, 2018

Testing Tour Stop #18: Mob Exploring with Amitai and codecentric

This stop on my testing tour was really special. Why?
  1. I saw Amitai Schleier again! We had read each other's tweets since longer and first met in real life at the Mob Programming Conference 2018.
  2. Amitai went on a coding tour this summer, joining several companies for one week each to learn together (find more about it on his blog). Most of the tour happened to be in Germany, and the last two days in Munich at codecentric. He knew about my testing tour and imagined us to have a common stop in my home town.
  3. We mobbed. My original hypothesis for my tour was about either pairing or mobbing, however, I only had the chance to pair before.
  4. The people on the mob identified as developers and had no experience with exploratory testing. My tour was about learning from testers, still, this promised to be interesting. I did not join the mob myself but stayed the facilitator. Though my tour was about hands-on testing, with me included, I learned a lot about teaching testing.
  5. My previous stop with Simon Berner was about pairing and mobbing itself. Only a few days later I could put our exchange into practice!
At first I saw this session as guest visit on Amitai's coding tour. Amitai, however, wished for counting it as stop on my testing tour as well. Even though I was not on the mob myself, this indeed felt like a bonus stop on my tour. So be it! :)

Setting the Stage

Amitai and Thorsten Brunzendorf already welcomed me at the entry. We had a nice chat, set up the room intended for our mob and waited for interested people from codecentric to join our session. They came, and even brought a test target we could use where they would value feedback on. Great, I really like to have our sessions to not only have the training effect but also have real-life impact.

After Amitai and I introducing ourselves and sharing some words on our tours, the session started. I decided to have a short introduction round in the mob as well, asking them for their name and core expertise. I knew everybody was identifying as developers, however, it was still very interesting to hear the different backgrounds, interests and specialties. Some people were into DevOps, some preferred backend development, or even tended towards security testing. I was pleasantly surprised to hear most of the group already had mob experience, only two persons didn't. After giving them a short heads-up, I went over to introducing testing, and especially exploratory testing. I presented my preferred definition by Elisabeth Hendrickson, as well as her charter template that I find useful especially for people new to exploration.
"Simultaneously designing and executing tests to learn about the system, using your insights from the last experiment to inform the next."
"Explore (target)
With (resources)
to discover (information)"
(Elisabeth Hendrickson, Explore It!: Reduce Risk and Increase Confidence with Exploratory Testing)
I wanted to keep the basic theory short and have people learn by hands-on testing as soon as possible. I asked the ones who provided the test target to shortly introduce it and afterwards asked the mob to write their charter. They, however, were really unsure what they wanted to focus on as they felt they didn't know the product well enough yet. So I decided to have them do a recon session for the first rotations in the mob, mapping out the system. I also recommended to already jot down notes together as a mob, for example in a mind map. And off they went!

Let the Mob Explore

It was really interesting to observe this mob explore a product most of them saw for the first time. They were unsure about their procedure, trying to align themselves as a group. They jumped all over the place, already finding great issues; not taking notes of them. They started a list of assumptions and wishes they had for the product. And they shared lots of implicit knowledge with each other, just as I have witnessed it in so many mobs before.

After nearly one full round of rotations, I had them stop. I shared some of my observations and provided recommendations. I also asked them whether they now felt ready to write a charter so we could do more focused exploration, using a simple Roman vote for quick decision making: thumbsup - I feel we're ready, thumbsdown - we need more information, thumbs sideways - not sure which way to go. Half of the group voted up, half sideways. I could have used the mostly positive result to have them write a charter now, but decided to have them finish a full round of rotations before asking again. Now all thumbs went up! Chartering didn't come easy to them. I helped them through, trying to not enforce my own suggestions or opinion. They managed to come up with a first charter to start with, providing them the needed focus. Interestingly, they decided to explore user experience with the help of the source code! That promised to be interesting, and it was indeed.
  • They used the code as oracle - great.
  • They looked for existing unit tests - nice.
  • One person shared: "I still don't understand the test, can you shortly walk me through?" - liked that.
  • They fed different values to unit tests, in an attempt to use it for exploration - awesome! Unfortunately, they did not follow up on it, or left a remark for future sessions.
  • They said they should also test the feature from user perspective - exactly! It might be the code they saw was actually never executed, right?
Towards the end of our mob session they started a discussion of whether a certain feature was useful at all. They brainstormed ideas how a different solution might be way more valuable to the user in the end. They questioned the context the product might be used in. Many great points were made.

Retro? Debrief, Discussion, Feedback

On my testing tour sessions, I normally do a short retrospective in the end so we can learn how to improve (I learned their value on my very first stop with Maaret). This time, I did a debrief of the testing session with the mob instead, to have them reflect on certain aspects. It proofed really valuable to get them thinking about what they did. Mental note to myself: I should do that more often for my own testing as well.
  1. What was our mission?
  2. What did we test and what not?
  3. What did we learn?
  4. What feedback can we provide about the product, can we quickly derive it from our notes?
  5. What to focus on as next highest priority with the knowledge we have now? 
The response on the last question was my personal favorite. They wanted to talk with the product owner as next step to see if they were going in the right direction or whether other areas might be more risky.

After a short closing, the mob session was officially over. Still, the group stayed and started an informal discussion. One of the persons who proposed the test target now shared that he didn't want to spoil the session, but the tested product version was outdated already and they had indeed agreed to remove the challenged feature - so the mob's feeling was indeed correct here!

It was great to hear people's experience with testing, mobbing and knowledge sharing. I really enjoyed the conversation to conclude a great session. Afterwards, Amitai shared personal feedback with me. He said he could clearly see I did these sessions very often already. This feedback took me by surprise. I had facilitated several mob sessions since beginning of the year: three in my company, one at a meetup, and three at conferences. So this session had been my eighth only. Amitai shared it was received as really professional and it was exactly what he was hoping for. This showed me I had learned a lot and it also triggered me to reflect. The session had indeed felt good. Maybe because I taught developers new to exploratory testing? Would it have been the same in case they would have been other testers? Still to be answered. This I know for sure: I had shared a lot about testing and testers during the session while people were doing hands-on testing. The mob approach is just an awesome way not only to learn but also to teach testing.
Bonus: Amitai also shared what he learned on his blog as well as recorded a conversation with codecentric sharing their highlights from this stop; including my exploratory testing session! :D

No comments:

Post a Comment