A prospective client e-mailed me recently, asking that I describe my expertise indexing books of a certain discipline. It turns out I have none, but the author hired me to index the book anyway. Here is my reply, edited for anonymity:

My process when indexing does not depend so much on the discipline in which the book is written as on its subject matter. I try to approach each book with the needs of the person who will use the index in mind: as a reader who needs to find an anecdote that she recalls, as someone who picks up the book and flips through the index to see if it includes a topic of interest, or as a researcher who needs to find a certain statistic. I index people discussed in the book, as opposed to those whose research is cited (though sometimes the two overlap).

Anything that surprises or delights me also finds its way into the index—or at least into the draft that I send to the author for approval, which I try to make overly inclusive so that edits to it are mostly deletions (although one recent author convinced the press to keep my index exactly as I sent it rather than delete anything for the sake of saving space).

I provide that draft in Word with Track Changes locked on so that I can see what you would like to change and apply it in my indexing software, which ensures that alphabetization and page numbering remain correct even after any reorganization.

The formatting rules required by the publisher and Chicago style are my guidelines for providing subentries for any entry that shows up more than five or six times, as space allows. I also try to double post rather than use “see” references for important topics; if we think of something as “Voodoo economics,” we might be a little irritated to be sent over to “economics, Voodoo” instead of to the pages where it’s discussed. I try to use “see also” references to draw out threads for those who use the index as a research tool, and I also try to use them to create an impression of the narrative.

I use Sky indexing software with a macro that copies directly from the book’s pdf, and I have found that this combination, along with the close reading required for indexing, often reveals minor errors (a nickname spelled in two different ways, say, or a date whose digits are transposed but a hundred pages apart) that are otherwise very hard to catch.

I charge per indexable page, which excludes the frontmatter, any blank pages or pages that have zero entries, and bibliography.