Personally I'd say that unless you deal with particularly dynamic requirements, 6 - 8 hours a week is massively massively excessive.
What you need to ask yourself is how many of those hours are recovered by the efficiencies those tools offer you? If you use a tool 10 hours a week if it genuinely makes you 10% more productive then, assuming you keep a tool for a year, that's a 52 hours saving so, so long as you spent less than 52 hours on finding and selecting it then it was a good thing.
But I'm thinking that sort of saving is going to be rare. Think about the following:
- What's the upskilling time? Sure you're going to be more productive eventually but for the most part that won't happen instantly and you'll be less productive to start. You need to figure in training time (zero productivity) and the bit when you're achieving less than you did, as well as the bit where you're more productive.
- If you fiddle with tools a lot you're likely to swap them a lot. If this is the case then the payback time has to be a lot shorter as you probably won't have it for a year.
- If you swap tools a lot you probably never become an expert in any of them. Perhaps becoming better with your existing tool is a cheaper way to productivity?
- You probably work for 40 - 50 hours a week. You need to understand how much of that you spend on each task and how much it's really going to save you. If a tool allows you to do a database upgrade in half the time but they only take an hour and you do them once a month it's probably not that great.
- Look at the most time consuming tasks you carry out. What are the things that take you most time to do? The more time you spend doing something the greater the potential for savings. If you spend 10 hours a week doing something and can save 20% that's massive. Conversely if you do something rarely and or it's very quick (say half an hour) even eliminating it might not be worth the effort in doing so.
- What are your pain points? Sometimes tasks cause disproportionate pain - perhaps they cause interuptions or make customers unhappy. These are things where it might be worth investing time but only if that time pays for itself in an increase in productivity greater than solving the problem or a similar increase in customer happiness.
- Are you hearing really good things about something. And I mean really good things. If so then take a look, but be critical and look for credible people with solid stories and a consensus, not just people hyping something new.
- Understand the difference between better and different. This is the biggest thing. Novelty is lovely but all too often it's not actually better. Programmers are the worst in the world for reimplementing something with a relatively minor change and touting it around. Different is interesting but it's not always better.
- Set up a periodic review for each item in your tool chain and research it at that point. Personally I'd say more than once a year would be over the top. Somewhere around 18 months - 2 years would probably the point where things will have improved enough to justify a change.
Against all this, if you want to fiddle because you enjoy it, then by all means fiddle. Just don't kid yourself that it's productive when it's not and (as a development manager) please don't try to convince me it is unless you've been through the list above.