In my experience, the problem is that those most in need of such a service are those who are least likely to use it.
In my peer group, there are some developers who treat their employment like a business and negotiate aggressively. They read up on tactics and practice them. I'm not sure agents would actually be that helpful to such people.
Another whole group basically acts like they're paid "enough" and don't negotiate at all. They're unlikely to see the utility in professional negotiators since they're not even willing to read up on tactics themselves.
I see these things too. Most developers believe they are much smarter than the average recruiter (which is certainly true) and probably think they can negotiate at least as well with even minimal training or reading. It's a matter of convincing a smart developer that the negotiation will likely provide a better result than would have happened without the assistance, and for a price that seems both fair and reasonable when compared to the amount negotiated.
I tell you what I think I can get out of an employer. If you think you can get more, you take me on as a client and receive x% of the additional amount.
I like that idea, and I definitely would like to see an element of incentive like your model includes, but I imagine I'd also spend a fair amount of time either (a) saying no to people with unrealistic expectations or (b) trying to manage expectations down for those I'd consider overly optimistic, which could be perceived as opportunistic. It's tricky.
If you think most people's expectations aren't reasonable, then maybe your service isn't needed. Isn't your whole thesis that lots of engineers are getting less than their worth due to their inferior negotiation skills?
By "fair amount" of time, I wasn't trying to imply "most". I think at least a slight majority of the population have fairly reasonable expectations based on my years of experience. For them to be a client though (based on your pricing model), their expectations have to be reasonable and a bit lower than market or unreasonably low. Reasonable and a bit high doesn't work.
A small percentage tend to have expectations far below market.
There are thousands of engineers getting less than they could be getting if they had better negotiating skills, better marketing skills, and better information.
The number of underpaid engineers doesn't need to be anywhere near a majority for the service to be a good business opportunity, or needed. There are plenty of engineers that could use the help - the question is whether or not they'd seek it out.
In my peer group, there are some developers who treat their employment like a business and negotiate aggressively. They read up on tactics and practice them. I'm not sure agents would actually be that helpful to such people.
Another whole group basically acts like they're paid "enough" and don't negotiate at all. They're unlikely to see the utility in professional negotiators since they're not even willing to read up on tactics themselves.