When I was a newbie writer, I hated editing. Writing was the fun part (all right, it wasn't always fun, but editing was WORSE). I had no patience for the edit; I grimly hoped that my first drafts were dashing and brilliant, and in no need of revisions. I think many newbie writers suffer the same delusion, much in the same way we squeeze our eyes shut and hope our insufficient efforts are enough in any other part of life. It would be nice. But most of us find out we're WRONG. And for me, there's no better way to demonstrate this than with data. I've previously blogged about my love of spreadsheets; I will not repeat it here (evidence will come shortly in any case). I've also previously blogged about the editing iceberg with my debut novel, Ryders Ridge, but I did that with time estimates. Having just turned in the revised edition of the follow-on, Iron Junction, I thought I'd break down the editing process (this time with numbers) and show you all how much a manuscript (at least one of mine) changes after I write THE END. Now, you may be out there as a brilliant new writer and every word comes out golden (if you really think this, I think you should take a long, hard, sober look at your writing). But I'm not like that. I rather hope it's the other way around – I'll change less as I get more and more experienced (I can hope).
So, here's what I did. I went through the 12 versions of the manuscript I had and collected data: word count and % change from the last version. For % change, I basically did a document compare in Word, and estimated the changes by counting pages (i.e. it's NOT the change in word count, but the overall change throughout the text). Yes, it's a rough guide, but representative. I also made notes about the different drafts. The data, in tables and a graph, are below, but here's the major conclusions:
- My own edit of the first draft produced the most change in the whole process (62%)
- Having said that, later drafts still changed significantly, up to 25% of the text being altered
- Big structural changes can happen even late in the process (see bold) – this happened after trying to make something work from the beginning that just didn't fit. It was so much better afterwards.
I guess the lesson here is: you probably need to edit. And you should be brave to edit substantially if needed. Delete characters. Delete subplots. Write new scenes. I've heard Terry Pratchett credited with saying that the first draft is you telling the story to yourself. So remember it needs a lot of work before you're ready to tell it to someone else.
Table 1: Editing data from Iron Junction
Version | Word Count | % difference | Notes |
1 |
94,479 |
- |
First draft |
2 |
90,121 |
62% |
My changes entered – draft beta readers read |
3 |
94,724 |
29% |
Changes from beta reader comments (round 1) |
4** |
96,928 |
38% |
Changes from beta reader comments (round 2); submitted version. |
Revised** |
97,560 |
42% |
Changes from publisher's comments; re-submitted version |
**versions 4 and Revised both had a number of sub-versions; see below.
Version | Word Count | % difference | Notes |
4.1 |
99,014 |
14% |
This is the difference to v3 |
4.2 |
98,954 |
3% |
First chapter changes only |
4.3 |
96,895 |
23% |
Prologue added; back-story emphasised |
4.4 |
96,928 |
4% |
This is version 4 in the table above |
Revised v1 |
99,905 |
25% |
Major structural changes |
Revised v2 |
100,099 |
3% |
Scene movements. Version re-read by beta reader. |
Revised v3 |
97,280 |
11% |
Back-story deleted. Version re-read by beta reader. |
Revised v4 |
97,819 |
1% |
Additions of small details |
Revised v5 |
97,560 |
6% |
Last revisions - my read-through and beta reader comments. Re-submitted |