I think every indexer will, at some point, experience a baptism by fire from updating a previously written index. It is tedious. It is time-consuming. It is mental gymnastics trying to get into the mind of the original indexer and trying to figure out what has changed. 

My first freelance project was like this. Thankfully the budget that the client offered was not a pittance, but it was still a hard grind. I should have suggested that the index be re-written from scratch.

Which is the usual advice, when it comes to updating an index. It may take as much time to write a new index as it would take to update, but it will feel a lot easier because it is yours and because you won’t be constantly second-guessing yourself, wondering if a correction was overlooked.

I have come to realize, though, that not all updates are equal. And while they are usually still tedious, in some cases I think updating is the best option. To help me decide whether or not to accept an updating project, I have developed a set of criteria, which I try to rigorously stick to.

1. The first is, how extensive are the changes? Understanding this is key to estimating how much work will be involved. If chapters have been added, removed, or shuffled around, then definitely re-write the index from scratch. If it is just a handful or couple dozen changes throughout the text, then the main issue will probably be reflow. Which is not a lot of fun to fix, but it does have the virtue of being fairly straightforward and can usually be done within a day or two.
2. If it turns out there are only a few changes, I then ask if the client is able to mark all of the changes on a PDF. This removes the guesswork for me. Part of what takes so much time, otherwise, is trying to figure out what has changed between the two versions of the book. If the client can show me what has changed, I save time, the client saves money, and the changes are usually pretty easy to make. If the client cannot tell me where the changes are, then I will pass on the project.
3. I also consider whether or not I have written the original index. It is a lot easier to update if I have, because I understand how the index is written and I may even have some memory of writing it. Deconstructing someone else’s index can be an interesting learning exercise, but trying to update or fix someone else’s index is a lot more stress and mental work. In a way, I prefer it when clients ask me to update rather than finding someone else because even though I don’t really want to do it, I know I can probably do a faster and better job because I am already familiar with the index.
4. My relationship with the client is also important. This does not have anything to do with the index itself, so perhaps should not be considered, but I think it is important to consider a project in light of the larger relationship. I am more likely to say yes to a long-term client. And, last year, I did update an index for a regular client that I did not originally write but still met my other criteria. Even though I was paid, I also saw it as a favour and good customer service. If I did not write the original index and I don’t know the client, I will probably say no. 
5. Lastly, I estimate how much time it will take. For it to be worth it, it needs to take me less time then writing from scratch. This usually means that I want to finish within a day or two. Since I also charge per hour for updates, completing the index within two days should also save the client money, as a brand new index would typically cost more. 

Thankfully, I don’t get asked too often to update indexes. I definitely prefer writing fresh. But when I am asked, it helps to have a plan for how to judge and complete the work. The pain and frustration does not need to be all-encompassing.