FAQs

Can you index my book?

Probably! It will depend on several factors, including the subject matter and whether I can fit your project into my current schedule. I’m happy to discuss whether I’m the right indexer for your book. You can get in touch with me through the contact page.

 

What do you charge?

I generally charge by the indexable page. My rates vary from $3.00 to $7.50 per page, depending on the book’s audience (children, trade, or scholarly), the subject of the book, the density of the text (number of words per page), the complexity of the material, and the turnaround time.  If you would like to discuss rates for a specific project, please contact me via the contact form.

 

My manuscript is finished; can you get started now?

I find it works best to index from the final page proofs rather than manuscript pages. When I work from a manuscript, I spend a lot of extra time changing all the locators (page numbers) once the text has been typeset and any major changes have been made. That extra time translates to additional cost for you, so it’s best to wait until the book is in proofs. Many publishing houses also prefer to have the index prepared at the same time as the proofreading, though usually not by the same person. However, you can (and should) try to schedule your project as soon as you know when the page proofs will be ready, since I often book projects several months in advance.

Why do I need a human indexer? Can’t a computer do the job?

If what you mean is, can indexing be done automatically by a computer, the answer is no. As of September 2025, even with recent advances in AI, no software is yet sophisticated enough to read and analyze the text and make all the informed decisions that a human indexer must often make. In addition to understanding the basic rules and structure of an index, a human indexer keeps in mind the needs and expectations of the reader, evaluating each concept or piece of information for its relevance in the context of this specific book, and designing a conceptual structure appropriate to the book’s content.

For several decades, some computer programs have claimed to produce an index, when in reality they generate a concordance: a list of specific terms that appear in the book, and which pages they appear on. This can lead to entries with well over 100 locators without any subheadings, which isn’t very helpful to a reader. Programs of this type also fail to include important discussions of a topic where the given search term did not appear on the page (a failing shared by search functions.)

Generative AI — Large Language Models, or LLMs — don’t do an adequate job, either. For a recent white paper on AI-generated indexes, Elisabeth Bartmess and Michele R. Combs (both professional indexers) tested several LLMs, feeding them the complete text of one or more public-domain books and asking them to create a “professional-quality index” for each. The indexers then compared the AI-generated indexes from each LLM with the book’s original human-created index. The AI-generated indexes failed in several areas. They were incomplete, skipping some or most of the pages of the supplied text and leaving out up to two-thirds of the main headings found in the human-created indexes. They were not easily navigable; several failed to identify or index the metatopic (the main topic of the book), and they struggled with cross-references or left them out entirely. And the AI-generated indexes were not at all accurate; they failed to include pages on which a heading was discussed, added page numbers on which the heading did not actually appear, and used the LLM’s own terminology instead of that of the author. One LLM failed to provide page numbers at all, even when specifically asked to do so. Bartmess and Combs concluded, “At this time, AI cannot replace professional book indexers and we are doubtful that it will be able to do so soon, or even at all.”

It’s also worth noting that due to copyright concerns, publishers are understandably reluctant to allow their works — especially not-yet-published works! — to be used in training Large Language Model (LLM) AI systems. Some publishers have begun including in their contracts language prohibiting the indexer from using generative AI at any stage of creating the index.

Further reading:

But you use a computer, don’t you?

Yes, we do. Thirty or forty years ago, creating an index meant writing each entry on an index card, sorting and filing those cards manually, and then typing up the index from the boxes and boxes of index cards. Thanks to computers, we no longer have to deal with index cards. Indexers today use dedicated indexing software programs that take care of the “grunt work” of alphabetizing, sorting, and formatting the entries. This allows us to focus our time and energy on the skilled aspects of the job: reading and analyzing the text, creating the overall index structure, and choosing terms—deciding how to phrase each entry and subentry so it is succinct, clear, and easy to find.

There are four major indexing programs: Cindex, Macrex, SKY Index, and Index Manager. Think of them as the indexing equivalent of word processing and database software combined: they can’t create the content, but they can help organize and format it. All four tools are flexible, powerful, and specifically designed for indexing. Using one of these programs, an indexer can quickly reformat an index from indented to run-in format, for instance, or switch the alphabetization style from word-by-word to letter-by-letter. Once the index is complete, the software can then generate an index file in Word, RTF, or any of several other formats, ready to send to the book’s editor. Almost all professional indexers are proficient in least one of these programs; I use SKY Index.

 

My question isn’t covered here.

I would be happy to answer any other questions you have. You can get in touch with me through the contact page.