This thread, motivated by "fight[ing] professional and/or personal isolation," has focused on sharing clients. Many have pointed out the difficulties and limitations of this. However, there is much that can be done short of that, including collaborative marketing efforts, providing mutual support, and passing along opportunities.
"Collaborative marketing" can include low-key efforts such as joint research, teaching short courses together, and teaming on proposals, as well as more overt marketing such as meeting jointly with prospective clients. The low-key stuff works well when you have common expertise whereas the more formal teaming is best when you have complementary expertise. (I work extensively with consultants in other fields, especially engineering, telecommunications, and environmental consultants, where there tends to be little overlap of expertise, even when those consultants already employ statisticians. The trick is to identify ways that you, as the statistician, can enhance the capabilities and credibility of your partners, and to exploit that to pursue consulting opportunities.)
"Providing mutual support" can be as easy as calling someone up and talking. Make it clear in advance whether the advice is billable or not and what the rates will be. When it's billable, the advisor sends a simple invoice and the advisee pays it (and issues a 1099 at the end of the year). It doesn't matter whether you're incorporated or not or how you're incorporated. Some of us, myself included, will provide some telephone or email support to consultants (and potential clients) at no charge on the premise that it helps establish a working relationship and may eventually create opportunities to work together. Support can also consist of reviewing reports, programs, analytical strategies, proposals, and occasionally more extensive hands-on work (more about those prospects later).
"Passing along opportunities" is perhaps the easiest way to share work. Colleagues, clients, and former clients frequently bring requests for proposals, grant opportunities, and other such material to my attention, but I can pursue only a fraction of these. In some cases it's not a good fit. On occasion, conflicts of interest arise: a lawyer is looking for an expert who has not worked for one of the parties in a case, for example. It these situations it's good to know of others who might be more suited, because a good referral helps the client.
I would like say a few more words about subcontracting. Some projects may be large enough that specialized portions could usefully be separated and farmed out to another statistician. In my experience (23 years, 18 as a small business or sole practitioner) this is rare. A typical project requires a good understanding of the client's needs, knowledge of their business, a willingness to apply (and first learn, if necessary) whatever technique will address their problems, an ability to interpret and explain a statistical analysis, and good writing skills. Usually, only a small part of the work actually consists of statistical calculations: the bulk consists of identifying what the *real* problem is and then helping the client use your statistical insight to take some action or make a decision. Basing a consulting business on the hope of getting subcontracts that require only statistical analysis, but none of this other stuff, may be unrealistic for most.
-------------------------------------------
William Huber
Quantitative Decisions
-------------------------------------------