Despite widespread acceptance in the scientific field and with some publishers, and recent advances in the humanities, TeX-based systems have largely been publicised by word of mouth, and no-one can tell how many users there are. The commercial versions advertise where they deem fit but generic publicity for TeX is not common.
Users can be TeX's best advocates, but formal training is rare. Users learn mostly from colleagues -- themselves often ill-taught -- and acquire bad habits which are hard to overcome. The results are often responsible for the poor image TeX has had among most printers and publishers. If TeX systems are so goodat typographics, the question is often asked, why does the output look so poor? Although TUG runs courses, it is hard to cover such a geographically dispersed user population.
Support for TeX via the Internet is excellent, often far superior to that of other products, but there is always a need for more introductory documentation aimed at the beginner and the non-scientific user. Some installation help is also still needed, especially for the first-timer: the assumption that everyone is already skilled in the principles of computing no longer holds.
This paper argues that the biggest need is for distributable publicity targeted at identifiable markets backed up by presentable documentation. More of the power of LaTeX should be made use of in creating these documents if it is to regain its market share.