You seem to be assuming that a large number of probably lower-quality leads is better than a smaller number of probably higher-quality leads. Usually, the more you know about a lead, the better the lead, in the sense that you can select out leads based on their attributes, e.g. 5000 leads that match this description "male between the ages of 25-40 who has completed college, is early adopter of electronic gadgets, and has shopped online at least 5 times in the past six months" are usually better than 20,000 leads that match this description "males between the ages of 25-40". I don't have any stats to back this up, but I worked doing data warehousing for about 5 years for a direct marketing company with clientele with a national presence, and the head of the operation held that opinion. But there's probably a "Sweet Spot" where you haven't asked for too little information and haven't asked for too much. Knowing in advance which questions to ask and which to jettison is not always easy, but your sales department should be able to refine things with each iteration.
Moreover, the quality of the questionnaire experience can be more important than the absolute number of questions asked. If the server is sluggish, even a few questions might prove too much. But if the server is snappy, and the screens are pleasant to use, then users might be willing to stick around longer.