Wednesday, 25 January 2017

Test Strategy in 10 minutes

This post covers both the workshop I did on test strategy at Solita as a part of the Testing Tuesday and the extempore workshop I arranged at the Nordic Testing Days in 2015. While both had the same agenda, they were vastly different. (Why this post is published now is because it had converted into a draft which I noticed only now.)

The one at Solita

The workshop was not supposed to be a slide show, like none of the workshops arranged at Testing Tuesday. Once again I drew stuff on the white board and let rip. I started by describing testing strategy quickly. Then we discussed about project elements, product elements and quality aspects. These were loosely based on the Rapid Software Testing models. The audience gave ideas on what kind of elements we need to take into account when choosing a test strategy. There is no comprehensive list on what we came up with, but we had a whole lot of ideas ranging from schedules to "pissed-offness" and expected user behavior.



After we covered most of the areas, I split the audience into three groups. They all were supposed to create a testing strategy in ten minutes. The product I chose was a web shop that sells knick-knacks and mails them to people. After ten minutes I had three completely different strategies.

The first one was focused solely on money. They mapped out every aspect that could hinder the income of revenue and prioritized them according to importance to the "owner" of the store.

The second one was a "software engineer approach". That group mapped out all the aspects that were important to different kind of engineers. There was an aspect of security testing, transaction testing, happy day testing of known processes and some complementary approaches. There was no prioritization but there was definite focus on tools and know practices.

The third one was focused on business processes. They mapped out as many user stories as they could and dissected them to steps. They then tried to figure out what kind of a prioritization would make sense to test the processes.

In 10 minutes we had three outstanding drafts of testing strategies. All of them had different approaches and they the strategy supported the focus they had chosen. They were not comprehensive, nor should they have been, but they were "good enough" to start testing the most important thing as soon as possible. We discussed the fact that when combined these three could actually complement each other. In a coffee break, we can create a draft of a strategy and even introduce that draft to a potential customer to explain our strategy to test their product.

All in all, the workshop was a success, since it spawned a new way to think testing strategy. I never thought I could choose a simple focus like "money" as the guideline of my testing. It does reveal the importance of talking to stakeholders to understand their values and needs. Without any templates or predetermined practice every team could conjure an awesome strategy which seemed even executable after they explained what they thought.


The one at Tallinn

Feeling confident on the outcomes of the workshop, I agreed with Helena Jeret-Mäe that I would do an extempore workshop during a coffee break of the Nordic Testing Days (see here). The workshop was supposed to be on the last day. I dragged a white board to the hallway near sofas and gathered people to do the exercise. I was able to gather a couple great minds of software testing on the sofa including Erik Brickarp and Santosh Tuppad. The workshop started by me explaining the idea and then I gave the same task to them: "Generate a test strategy in 10 minutes."



The problem was that the pro testers weren't too happy with the vagueness of the context. They wanted to know more about the product, more about the stakeholders, more about the project. In the end we didn't actually generate a context map for the product (plus some testing strategy elements).

The best outcome for me was (once again) to observe experts in their work and how the dynamics within a group actually work. Also the lack of structure was something they pointed out. Erik mentioned that we could have described four key aspects of the product and then figure out how to test those. Where the first three groups I had at my office all chose a focus, we had none. We could have spent a few minutes in the beginning to define where we would like to focus (even an arbitrary focus) and then create a strategy based on that. There were hints of focus surfacing when they interviewed me about the product, but none that was chosen as the key thread.


What next?

Having had two vastly different workshops on the same subject, I think it makes sense to arrange these even more. The original idea to this came from Fiona Charles at EuroSTAR 2013, but I think we approached it quite differently. I am planning to do these extempore workshops at every conference I attend, since it makes sense to give people a chance to try their skills at this. Every time I get a huge amount of pointers, ideas and lessons learned. I would think that it is the conversation that ensues rather than actually creating a strategy that counts.

The key lessons learned in both sessions were:
- The dynamics in the group determine a lot what kind of strategy is created
- Know your stakeholders!
- Choose a focus as a skeleton and then fatten it up
- Don't over-do it! Ten minutes might be enough to start testing the most important thing.
- Have some structure, but keep ideas flowing

When you see me at a conference, come ask when am I gonna pop out the white board. It might be the next coffee break. ;)

- Peksi

No comments: