The transcription of records for indexing is also valuable, but less so, given that volunteers were working from projected microfilm copies of (usually) hand-written records (some in beautiful but dated hands).
OCR in the 1980s was not adequate, but Soundex indexing helps a lot, and if you need to search a small area exhaustively, you can see many of the microfilm copies of records through a site like ancestry.com or from the digital media available for use at local LDS family history centers.
The LDS was also strong in producing the GEDCOM data standard, which is widely (almost universally) used for family tree exchanges, and the PAF software package (which now has many competitors.)
Note that research into ancestry has an interesting network effect - while people have 2^N N-parents, once N gets over 4 or 5 it can become challenging to trace individuals - but the chances that some other active researcher shares some of those 2^N N-parents start to increase - and ancestry.com has done well at providing links to other researcher's public trees when they become available.
This pattern has the effect of encouraging subscribers who have already found most of their ancestors out to level N, for whatever N, to continue to subscribe in hopes of finding a new link from another researcher, or perhaps a new clue in a newly digitized document set.
It will be interesting to see how the decreasing cost of genome analysis affects family history research - perhaps helping find distant cousins, or perhaps revealing surprising differences between a presumed history and the biological record.