Thread by @michaelbolton

@MeekSoItIs 1) Clearly I have to write a blog post on this.

Instead of wondering about how to *measure* performance, consider how to *assess* performance.

As testers, we don't measure the quality of the product. We assess it.; we evaluate it.

We don't worry about "objectivity", either. /

@MeekSoItIs 2) When we evaluate a product, we're trying to identify problems that matter to the people who will use it and to the people who are producing it. That is, we're trying to see if there are (potential) problems in the relationship between the product and someone who matters. /

@MeekSoItIs 3) One really good way to do that is to identify *quality criteria* — things that people might want from the product. In RST, our list includes capability, reliability, usability, charisma, security, scalability, compatibility, perfomance, installability... /

@MeekSoItIs 4) If one of those attributes is less that what it could be or entirely absent, someone may suffer loss, harm, bad feelings, or diminished value. One might say that's not "objective", but if a problem affects someone important, who matters, objectivity is irrelevant. /

@MeekSoItIs 5) We have categories by which we can evaluate a product relative to people who matter (and, of course, different people may have different categories, which is fine). Can we apply those categories to testing work? To the work of a a tester? We could, couldn't we? /

@MeekSoItIs 6) Is the tester capable? Reliable? (Hmm.. what would "usability" map to for testing work, or for a tester? Maybe: is it easy to learn how the tester works? Is it easy to get stuff from him or her, with a minimum of effort or fuss? Is the tester accessible, adaptable?) /

@MeekSoItIs 7) We could go right down the HSTM's list of quality criteria for a product (satisfice.com/download/heuri…, page 5) and see parallels. We should be careful to remember that testers are human beings, not products, but a tester's work represents a product in some sense. /

@MeekSoItIs 8) It might also be to helpful have a list of what we might consider to be elements of excellent testing work. Maybe something like this: satisfice.com/download/eleme… Maybe some of those things are important to people that matter; maybe not. You might prefer your own list. /

@MeekSoItIs 9) With a set of quality criteria, skills, tactics, work products etc., we can go through each one and ask people who matter "are you okay with the tester's performance in this category?" If the answer is "Yes, Very", issue praise. If the answer is Yes, move to the next item. /

@MeekSoItIs 10) If the answer is No, dig deeper. Ask for specifics about the mismatch between the tester's work and the needs or desires of the person who matters. That's an important first step towards addressing the issue on way or another. *It's a lot like a bug report, right?* /

@MeekSoItIs 11) Gathering up all the answers into an assessment adds up to something that looks a lot like an evaluation of a product. In this case, the product is the tester and his/her work, but the general patterns are the same: we're seeking to evaluate quality and address problems. /

@MeekSoItIs 12) Remember that quality is *value to some person(s) who matters*. To evaluate the quality of testing work, identify people who matter; identify what they value; observe and examine the tester and his/work work products; look for problems in them that matter to people. /

@MeekSoItIs 13) Finally just as you'd apply the three-part testing story to an evaluation of a product, apply the three-part testing story to your evaluation of the tester and his/her work. developsense.com/blog/2018/02/h… In this case, Part 3 is about the quality of your evaluation. /

@MeekSoItIs 14) People are not products. Their behaviour is not deterministic; they have motivations, feelings, and fears that affect their performance, and that must be considered seriously. Yet if you know how to evaluate a product, you've got a good start on evaluating testing work. =