But this is a little like everyone programming their own desktop publishing tool. There is a reason why PageMaker and Quark Xpress dominate their industry. Much of what we publish is similar in presentation or style to other publications, and standard tools facilitate productivity (availability of trained workers, etc.)

If your content is truly different, then build your own CMS. But keep in mind that you may be more attracted to the technology (the medium) than to your content (the message). You would be better served by getting an experienced open-source CMS consultant to build it for you.

Far too many of those building a CMS think that they will sell it to the rest of the world and make their fortunes. It's a buyer's marketplace, but there's a very high chaff-to-wheat ratio.

But in the short term huge revenues flowed into market leaders like Vignette, Documentum, Interwoven, and Broadvision. If they have as many failures as success stories, the cause was very likely these impossible market pressures, and not the suppliers fault.

The vast income has been used to build (with research and development) and buy (with mergers and acquisitions) major technology improvements, so surviving companies are likely to provide solid products in the coming years of CMS realism.

You might also "buy" an open-source CMS by going to the many independent CMS consultants who will build it to your specifications.