Are there other factors unique to the Humanities that mean the book is a preferred publishing route?
The boundary between Humanities and Social Sciences is fuzzy in many places, and indeed the phenomenon of theses commonly being parlayed into books happens in some social sciences as well: e.g. history. But there are some factors common to these disciplines as opposed to non-STEM ones, yes.
- The length of the thesis.
Average (in any of several senses) thesis lengths vary widely by discipline: see here for a nice treatment. Theses in history (the longest) are almost three times as long as those in mathematics and bio-engineering (the two shortest). (Note that history is followed by anthropology and political science: all social sciences. English comes in at sixth longest in the list of 50 disciplines.) In order to have math theses average about 100 pages, they must be counting pages with respect to the spacing one uses in the officially submitted copies. So when math theses are single-spaced -- as they would be in any published form -- the average is more like 50 pages. See here if you don't believe that a math PhD thesis could be much shorter than that.
Anyway, the moral is clear: in some disciplines a perfectly good PhD thesis has length closer to (or equal to) that of a journal article than a book.
- The independence of the chapters of the thesis.
It is my understanding that in many STEM disciplines, a perfectly good PhD thesis is several independently written articles stapled together. (This is less true in math but not unheard of.) But in the humanities a thesis is usually one long, sustained argument such that it would detract from it to publish it in parts.
- The prior publication status of the work of the thesis.
It is my understanding that in many (most?) STEM disciplines outside of mathematics, the sense is that a PhD thesis should contain some already published work! Whether the work got published in a reasonable place becomes part of the committee's evaluation. Clearly this works against publishing as a book.
- The amount of time that one spends doing the work and writing of the thesis measured as a portion of one's academic career.
In the humanities and many social sciences, postdocs are no longer unheard of but are still quite rare. A lot of STEM PhDs are awarded to people in their mid to late 20's at the point where they have demonstrated that they can do one novel, sufficiently substantial piece of research. They then go on to 2-5 years of temporary faculty positions where the real work is done. In the humanities (and...), getting a PhD much below the age of 30 is pretty rare, and most tenure-track jobs -- even very strong and research-intensive ones -- hire straight out of PhD programs. One might say that a PhD in the humanities is more like a European Habilitation in the sciences: it's just a more mature, substantial piece of work.
- The feeling that someone outside of academia might just possibly be interested in buying a copy.
I have actually occasionally read and even bought books that I presume originated as a humanities / social sciences PhD theses. Though I am having trouble thinking of a truly gripping read, some of these books certainly make more of a concession to the general reader than STEM theses do. I have never bought a book describing five experiments the author did in someone else's lab...