I'm in the process of wrapping up RTK1 (on number 1800-something) and I'm curious to hear how other people have gone about introducing themselves to new kanji. I feel that my progress is quite slow (I generally work through about 10 kanji/hour) but I have no idea as to what is average or not. I do see people claiming to have finished the book in three weeks, which is insane to me: at my pace that would be 10.5 hours a day every day, not even counting reviews.

I use the 6th edition Anki deck which already includes the two highest rated stories for each kanji. My process is:

1) Check RTK for new primitive/meaning if applicable.

2) In Anki's card browser, check the story for the new kanji, modify as necessary. If I don't like either story, I'll either check the study section for alternates or come up with my own. Bold all element/keywords.

3) Do a quick search of my study decks for words which use that kanji. I have finished Core2k and my sentence deck has ~3100 mature cards, so I usually find a few usages. I also frequently check WWWJDIC for nouns and add those as well.

4) Add vocab using the new kanji (in kana) to the 'Japanese' field I have added to my Anki deck. This helps me distinguish between linked keywords and also reinforces vocab.

5) Once my stories and Japanese are all sorted, I write each kanji 7-8 times in my kanji practice books from the 100 yen store. I repeat a key mnemonic as I write each kanji.

6) Unsuspend new kanji in Anki, test myself, ensuring that I read the stories carefully for every new card.

7) Custom step on the Anki deck means I see the kanji again in 2 hours, but in general I'll see them either later in the evening or the next morning.

8) I always hit "hard" the first three times I see new kanji, along with ensuring I fully and carefully reread/imagine the corresponding story.

This is a lot of work. I am sure that most people do not go this far in adding new kanji, and I'm also sure I could be moving much more quickly. That said, my RTK deck's stats are quite good: 95% retention on new and review cards, 98% on mature cards. I've also found that, thanks to adding the Japanese field, kanji come much easier when thinking about vocab. I can write a lot of what I know from memory without checking, which is nice.

What is everybody else's process for adding new kanji?

I'm doing it similarly. I only use the book to look up new primitives, and only when I don't get what's up from my ANKI cards, which include stories. I don't necessarily alter the stories, but I bold the kanji keyword and make the radical keywords italic. I also delete the weaker one of the two stories, or rewrite it completely in cases where a story does nothing for me or I have a good idea.
I don't, however, care about Japanese vocab. My vocabulary is pretty good, and I keep the "ahh, so that's how it was" fun for later.
I do a card at 1 minute, then ten minutes and then 1 day, if I passed them. If not, I reset them to 1 minute, then 10 until I got it. Sometimes I edit or rewrite the stories again or concentrate on visualizing.

I use the SRS of this website instead of Anki. My process is a bit more basic:
1) Work through a chapter of RTK and jot down the stories/associations in pencil in between the panes, while typing them into the Kanji.Koohii SRS at the same time. Or I simply copy good stories/associations by other users.
2) I add the 6th edition cards directly after having provided them with stories. The regular ones follow in one go at the end of the chapter.
3) I "learn" the new Kanji for the first time the old-fashioned way with a piece of paper covering both the symbol and the stories/associations and then checking if I recreated them correctly in my mind. Initially, I did actually write them down a couple of times after having imagined the stories with closed eyes as Heisig suggests and reciting them as I drew the Kanji, but since I saw little added return, I abandoned that approach. Either the story/association works, or it doesn't and then no amount of imagining/telling/writing down the Kanji will make it stick.
4) I review the new Kanji from the "review" tab, and from then on they are part of the SRS-cycle the way Fabrice designed it. The ones I got the first time show up again the next day and then again, with the 4th time about 30 days from the day they were first done correctly. And the ones that went wrong go into the "restudy" stack. Until they go right and then enter the SRS-cycle.

My retention is about 70-80% for Kanji recently learned, but generally only about 63% for the mix of newer and older Kanji (expired from Boxes 3 through 5). I have spent some time on analysing what causes the mistakes and have come to the following conclusions:
A) Abstract concepts very often fail to trigger the stories that lead to the Kanji. E.g. "peaceful" and "quiet" are completely neutral concepts (to my mind) that don't allow specific stories to be attached to them for very long. No matter how clever, weird, shocking or brilliant they are.
B) "Close" meanings that interfere with each other, so that elements of two or more different stories bubble up when trying to recreate the Kanji. At first I thought this was a limited problem that I tried to solve by making a separate list with clusters of "close" Kanji. Now that I am close to filling my 14th page in Word with such clusters (while having a little under 1,600 Kanji in my deck) I am not so sure that this is such a "limited" problem anymore. (And I'm having a hard time believing that Heisig never encountered this problem when he first learned all of the RTK1 Kanji using his own method in - as he claims - about a month, but that aside.)
C) (This is probably somewhat related to point A.) Short term retention does *not* equal long term retention. Some of the greatest, most memorable stories/associations that stick so well at first turn out to have completely vanished from my memory upon the 3rd or 4th review after about 3 to 4 weeks. I chalk this up to long term memory (LTM) not being as impressionable as short term memory (STM). A story/association being weird, funny, shocking, personal or whatever is simply not enough to impress LTM, I have found. In order to do that it has to be so logical as to be a striking illustration of the key meaning, or it has to provide some new fact about the key meaning. Only that sort of thing seems to impress the gatekeeper that guards the entrance to my LTM. Or a key meaning has to feature in a larger story with other key meanings in a sort of "first this happened, then that and then that" style. For some reason, a chronological ordering of events seems to provide a free pass for the gateway to LTM. But it can be rather cumbersome to recall a Kanji from such a larger framework.

Personally, I wouldn't be surprised if the mystery of people who have claimed to have finished RTK and then suddenly vanish from this website and never report anything about their more advanced experience with learning Japanese has a lot to do with the transition from STM to LTM. I suspect that many of them found that they weren't as "finished" with RTK as they thought they were and that their retention dropped dramatically as soon as they stopped pumping the Kanji they thought they had mastered into their short term memories every few weeks as they fell back from the more mature stacks.

Guess I went OT there. Sorry about that.

What I also found is, that the keyword for the Kanji has to be the conclusion, the punchline of the story.

There's a story for *concept* that goes like "the concept of love is hearts interacting" or some such stuff, and since it targets *love*, it doesn't fire when i encounter *concept*.
What would work for me here would be a story closer to something like "you have to interact to reach a conclusion that leads to a usable *concept*".

This might only be me, but I rewrite stories to match this kind of format to always have the same train of thought.
element (radical) + element + element = keyword (meaning)

This is in agreement with the poster above. I may find a story like "you have to put a few fingers into her to extract her juice" and it works, because a) it's hilarious to phrase it like that, and b) it's a conclusion. I can remember it even after I didn't see it after a month, or I just relearn it in an instant.

I also made leeches do nothing on ANKI, as removing them from the deck will only save time, but it won't make me learn them. Something like 曜 has to be learned though, it's an important character. And if that means I "waste time" on it, so be it. It only means the story I need to make it stick has to be more streamlined.

I will consider "being done with RTK" as soon as I initially learn the last Kanji, but I will not think of that as "done with the Kanji", as that's a whole new dimension.