Most consultancy firms that don't respond to RFPs eventually die. Of those firms that survive, they could be viewed as a collective of successful freelancers. There typically is not enough revenue to grow the company beyond the founding consultants because the project budgets that they deal with are typically $15K to $50K. You can definitely get away with not responding to RFPs if you're only ever dealing with budgets of tens of thousands of dollars. Also, an RFP is a way for customers to weed out the "I just quit my day job at an agency and started my own consulting company" firms. These customers want to deal with bigger firms because they expect these companies to be around for a while and provide support years down the road. A bigger company can afford to devote a few employees for a couple of weeks to address an RFP properly.
Depends on the project. A lot of website RFPs can be done in a crunch session of 1 to 2 days. I've had to bid on systems that interacted with assembly lines and ones that involved dealing with the government and getting security clearances. These latter two took 2 calendar weeks although the people involved in putting together the RFP were definitely not working full time on it until the last minute. All in all you might have 1-2 days of actual work, 1-2 days spent Q&A attending meetings with the client and other potential vendors, etc. The time can add up.
I have seen one ridiculous "RFP" that a government agency up here put out recently in which they wanted potential vendors to essentially do 6-8 months worth of work and if they landed the project, they'd end up with a long term multi-year contract. We're talking tens of millions of dollars here, but still, that one struck me as quite the lottery and I'm pretty sure the only companies that responded to it were the big names consultancy firms.
I've done some work for a Federal government agency a few years back and a provincial government within the past year. eHealth was exceptional not because of the cost but because nothing got done.
Here's where posting anonymously is beneficial (but alas since it's anonymous, take my account with a grain of salt). My opinion is that a select few large IT companies in Canada have a virtual monopoly on all government work. In fact, the only reason that I was involved in these projects was because certain teams from these companies screwed up so badly that some bureaucrats formed a skunk works side project to move things along. There was an unspoken understanding that the government wouldn't sue and shake the boat, but that they'd screwed up badly and needed to let another party try to resolve the situation.
Pretty much the only way you're going to get access to government contracts is via sub-contracting for these firms. To bring this post back on topic, the reason that this happens is because in a bid to try to account for government spending, the government issues these RFPs that are so lengthy and cumbersome to accomplish that only the big companies can actually pull them off. But it's a premature optimization and means that the government spends more money than it needs to. It's not unheard of for a simple web application to have a budget of CDN$1M. This would be something that your typical YCombinator team could pull off in 3 months and US$17K.
Some of the more successful government projects that I've heard of are ones where you have a bureaucrat that hires acquaintances without an active bidding process, because they can vouch for their expertise. Stuff gets done quickly and correctly the first time because there's little red tape to work around. Unfortunately, eHealth involved friends hiring friends and stuff just not getting done. I always thought that the criticism of eHealth folks hiring people they knew was less of a concern than the fact they simply didn't get stuff done. I mean think about it, in the private sector do you get people to actively bid on every little position and project or do you have your go-to guys?
Anyways, my opinion is otherwise largely positive with government work here in Canada. They seem to spend more money than I think they would if they let smaller more agile companies into the fold more often. But stuff seems to get done and when it doesn't (e.g., eHealth) you hear about it. Most of the government people I've worked alongside care very much about their job and doing it well. It wasn't unheard of for them to reply to my off-hours questions on their Blackberries on a weekend. So all is not lost. If I'm a little bitter, it has to do with what I perceive to be a rigged RFP process that favours these big vendors.
I used to work at a company that sold a software product to provincial/municipal governments in Canada, and the RFPs were often rigged. Sometimes in our favor, sometimes for a competitor. There were cases when the RFP was written in concert with our sales team to ensure we would 'win'. At other times it seemed genuine, but you can never be sure. Often the customer knows what they want, but they're required by law to go through the hoops.
This is a really well thought-out response. Even though people would probably like to describe eHealth as an exception, I think that situation really exemplifies the problems I was trying to expose in the RFP process. It's just another reason it's a bad system.
No, it exemplifies what's wrong in the GSA-style procurement system, where the process is deliberately complicated and biased towards a small number of institutionally-accepted contract holders. It has nothing at all to do with the bare concept of an RFP, and does not reflect the nature of RFPs outside of FedGov.