I'm running a team that takes care of 3 products that are distinct in language/product type so that 2 out of 6 developers I manage are only ever locked to coding on 2 of the products (thick client applications). The other 4, are the web developers and look after a very rich/complicated web application.
The process the team are used to doing is story pointing based entirely on complexity alone, but the web team and other 2 dev's point their own user stories for their products. So 4 dev's point their web stories, and the 2 point their other product stories. The team then total those up, and the velocity calculation is based on all 6 of those developers. Further to this, there are 3 QA's who bounce between being able to test either web and the other products, so don't strictly belong to the web developers alone, or the 2 other devs.
I'm really struggling to grasp that this is the 'Agile' way and what I want to do is split the two groups, lock 1 QA to the 2 thick client developers, and then lock 2 other QA to the web team. This way, I can get independent velocity and manage the time of both groups better.
Another problem I'm having is having no visibility of when a release is QA heavy or development heavy and light on QA. Thus having developers and/or QA running out of work. It's driving me a little crazy that I can't see any estimation of QA efforts along side development effort.
Can someone tell me where I'm going wrong in my thinking, or this teams process flawed? I just feel the 'Agile' scrum process here has such lack of clarity behind the individuals skills/time management.