One of the great frustrations of working in IT is that you very often know lots of things the bosses don't, but they still get to review and possibly override any recommendations you make - even if those recommendations are directly in your area of expertise.
One result of this is that people who want to keep their jobs don't repeatedly make recommendations they know the bosses are going to turn down - and that reality, in turn, affects how people trying to sell to your company behave.
Thus one of the things big, well established, sellers do is to play this out against both you and your boss -often meaning, in practice, that their mid level sales people go golfing with your bosses, and some juniors come round to your shop to listen attentively to your every need while quietly guiding you to issuing just the right purchase recommendations at just the right times to meet their boss's quotas and expectations.
So, assuming you actually know better than to want the stuff you're buying, how do you beat this strategy without prematurely ending your budding career in IT management?
One approach involves applying IT marketing's 80:20 rule (bearing in mind that the numbers used in the 80:20 rule aren't remotely point estimates - just magical stand-ins for comparing an assumed overwhelming majority to whatever's left after that majority is accounted for - this rule states that 80% of new sales are made on the basis of the 20% of product claims that cannot be realized through your use of the product) by finding the intersection between the 20% that can't be realized, and the companies your bosses want to buy from.
Imagine, for example, that you want to put a couple of your people to work with some new gear to develop and prove out some magical new application for the organization; think the right infrastructure for this is SPARC/Solaris and open source; but your bosses golf with some locally senior people from IBM and a guy who used to represent HP locally when that meant something positive, but now runs a consulting services company making its money by promising Wintel reliability and security.
Remember, when thinking about this, that there are two main reasons your bosses golf with these guys:
In this situation the dynamics are clear: the right technical proposal is a career limiting move for you, while one based on buying wintel gear from IBM and having the ex-HP guy front some people to work on them will cost the company much more to produce much less, but some of the instant and enthusiastic support this will garner from your bosses will rub off positively on you.
So what do you do? Try splitting the project into multiple pieces with the biggest chunk committed to the 80% you can achieve and the remainder divided in ways designed to gain senior executive support.
In the example, therefore, you would format your proposal to minimize the importance, value, risk, and cost of the 80% of the project you're going to do internally on Solaris/SPARC, hype the skills and risks associated with the remaining 20%, list price it, and then carve it up between the Wintel wizards and IBM.
In doing this you'll rely on part of the sales dynamic: the fact that the sellers will be perfectly aware of what's going on and will adjust prices accordingly with the no hope seller cutting things to the bone, and the guys with the sure thing padding things out where ever possible.
As a result your 80:20 split can often be proposed as much closer to 20:80 in both costs and risks with your hacks limited to the 80% in boring low risk work they're capable of and the 20% suppliers delivering the high risk, high cost, high excitement, stuff - and when you apply the same rates to cost out what the high price guys would want to do the 80%, your bosses can intuit huge cost savings from splitting the deal: meaning that approving the thing gets them what they want:
Obviously the hows and whats of this strategy are going to depend on the situation, but in general the bottom line on bosses who override technical recommendations on non technical grounds is that they're usually reacting to either/or choices by preferring group and personal reassurance and emotional support over the possible corporate benefits associated with a decision they can see as technically risky - and that you can generally beat this by breaking each half of the either/or into the 80% that's common between them but cheaper your way, and the 20% that's sufficiently risky you'd rather have somebody else be responsible for it anyway.