It is important to take the time to figure out and focus on your target market. You can do this in part by documenting your successes and challenges and finding out what works for you.
Have you ever written down on paper why certain clients return to you on a regular basis? Have you thought about why the RFPs that you respond to don't always work out in your favour?
Taking stock of why you "win some", "lose some" is an important way to identify your target market. Identifying those that help you win also can be useful in winning more and losing less.
Let me tell you a story on how I took stock: I responded to an RFP for a large Canadian NFP to modernize their website to use a content management system. This was about 5 years ago when CMSs were still a bit of a novelty. During the Q&A session of the RFP process I was surprised to find well over 40 vendors in the audience. There were vendors of all sizes from small (I was partnering with an organization called VentureLabour.com whose system I used in another project) to huge vendors like Vignette and IBM.
Needless to say we didn't make the short list. The larger vendors were seen as more reliable and a sure bet, even though they potentially were not the right solution to the NFP. But as you may have heard "noone was ever fired for hiring IBM". In this case the winning vendor could not sign an amenable deal with the NFP and the 2nd place vendor, an integrator out of Ottawa who implemented RedDot did not even get a chance to respond. For some reason the NFP chose vignette, probably for the name and probably because they provided the software for *free* but charged for the services. I don't know the exact details, but I can honestly tell you that unless I had a direct connection to get the deal I had no chance in competing for this deal. The most amazing thing is that the current home page design for this pan-Canadian NFP is still using the design we came up with -- for free. But we didn't get the implementation services work. On the win/loss analysis I put down -- don't compete with the likes of IBM and Vignette on RFPs no matter how good you think your technology is.