In some companies, the testers are expected to do Quality Assurance. In my opinion, this is not the way to go. Testers normally don't have the mandate to push things thorugh on a higher level. Because of that, it is impossible to perform decent QA.
QA should be done by a QA manager that consults the testers. What do you guys think?
As you already stated, testers (normally) don't have the power, budget or other means to actually improve the quality themselves (as in rewrite code, extend deadlines, reduce features).
It can be argued, that by giving information to the decision makers (stakeholders) we do improve quality indirectly (if the people listen to our advice and act upon it).
By putting the decision not on our shoulders (or ours alone), one can share the responsibility and create a commitment for all parties involved.
Thats what I did in my current company.
A) It takes the heat and blame of the testers and my team.
B) I created a team for the release decision making of product management, developers and testers. C) The collegues involved get a better insight, what is he current state of the product.
D) If they still need to release it (market, special feature, etc.), they decide so better informed and conscious.
E) By involving different aspects for the decision (market view, technical view, user view, aka product management, developers and testers) we made sure, that we cover all important aspects which should be seen.
I really like the other approach: QA can stand for Quality Assistance. QA role is to inform what quality is the product. I think PM should be responsible to make decisions based on this and other information he/she has.
There isn't a stock answer to this question - it really depends on the organization and the product, and even then I'm sure there are multiple philosophies that can be successful. I suggest focusing on pleasing your customers, and iterate your processes just as you iterate your products.
Separate roles for persons who execute and plan tests are often unnecessary, especially in a small team. These are two closely related functions, so I don't see any disadvantage for a tester to write a test spec, or for a test manager to do some exploratory testing.
In addition to that such strict separation adds bureaucracy and reduces team flexibility.
