54 Responses so far

1. 1. Alex on Feb 7th, 2005 at 1:25 pm

   Wow. Thanks again for doing this.

   Did a lot of people give (you) their name or did most people want to remain anonymous even from you?

   Alex

2. 2. tobias buckell on Feb 7th, 2005 at 1:41 pm

   Ah, I forget to address that. 60 people gave me their names, 14 did not. I was expecting a higher ratio, and looking over the data the 14 didn’t skew anything, so for now I’m going to keep it as authentic since it’s such a smaller group…

3. 3. John Scalzi on Feb 7th, 2005 at 1:46 pm

   Whoever is earning $600,000 for their fantasy novel is substantially skewing the average results for that genre. In that genre, the median is almost certainly the way to go.

4. 4. tobias buckell on Feb 7th, 2005 at 1:57 pm

   Duly noted and corrected to show the difference, John.

5. 5. Jon on Feb 7th, 2005 at 4:19 pm

   A $600,000 advance for a first novel? Sweet Jeebus, that’s impressive. Also terrifying, if they didn’t come close to earning it out.

   At any rate, that’s some interesting data, Toby. Thanks for doing this.

6. 6. Jon on Feb 7th, 2005 at 4:22 pm

   That’s some impressive data collecting, Toby. Thanks for doing it.

   And as an aside, a $600,000 advance for a first novel? Holy crap.

7. 7. Naomi Kritzer on Feb 7th, 2005 at 4:46 pm

   Unless I’m reading it wrong, the $600,000 advance was for a “most recent” novel, not a first novel — probably by someone who is very well-established in the industry with a history of bestselling books, rather than a very lucky first-time author.

8. 8. tobias buckell on Feb 7th, 2005 at 4:55 pm

   Naomi is correct. The highest first novel advance is not anywhere near that:

   The range was from $0-$40,000 for an advance on a first novel.

9. 9. Trent Walters on Feb 7th, 2005 at 7:15 pm

   Interesting tabulations, Tobias. Thanks. You might want to add–and I’m not sure how you could do this without time-consuming software and telephone solicitations–that these are not random samples. Still it is informative and something to think about.

10. 10. Jon on Feb 7th, 2005 at 7:52 pm

    Phew. Also, sorry for the double post. Had some weird internet thing going on, didn’t think it took.

11. 11. Dawn B. on Feb 7th, 2005 at 10:20 pm

    Nice work.

    I would (personally) prefer to see a representation w/o the $600,000 advance because it is skewing it something awful. Especially if it was an anon entry. Also, noting that it is not shown on the publisher page [could be one data point for DAW or something I guess which doesn’t have five] makes me less likely to believe it. Any way you could get a hold of that author to double check for a typo of $60,000?

    Confirms what I knew about the first time novel game. Thanks.

12. 12. Dawn B. on Feb 7th, 2005 at 10:23 pm

    Also: Re 4000 writers.

    Locus is a bad indicator, IMO, for who is a writer. I sub to Locus and I’m not published. Plenty of other people I know who aren’t published sub there too. And I’d be willing to be there are several writers who don’t sub to Locus.

    I would suggested a value of 3000 NOVELISTS in the field, as a short story writer could sub to Locus and/or be a member of SWFA w/o publishing a novel.

    Have you asked about this survey being linked the SWFA bulliten?

13. 13. tobias buckell on Feb 7th, 2005 at 11:44 pm

    I would (personally) prefer to see a representation w/o the $600,000 advance because it is skewing it something awful. Especially if it was an anon entry. Also, noting that it is not shown on the publisher page [could be one data point for DAW or something I guess which doesn’t have five] makes me less likely to believe it.

    Okay, to the validity of the 600K, it is not an anonymous datapoint (I pointed out that the anonymous data provided didn’t significantly skew our data). It is not a typo either.

    Taking it out of agented submission does change the equation in averages, but that is why I did medians as well, always look to the median for an accurate snapshot of your stats data. Here are the new results:

    The range in agented advances is now from $1000 to $260,000
    The average agented advance is $24,800
    The median agented advance stays the same at $12,500

    The range in unagented advances is from $0 to $21,500
    The average unagented advance is $14,143
    The median unagented advance is $7500

    Not having an agent still significantly alters the advance it seems.

    I also received a note from an agent that said I’d see even more dramatic differences had I included some way to track foreign rights sales…

    Locus is a bad indicator, IMO, for who is a writer. I sub to Locus and I’m not published. Plenty of other people I know who aren’t published sub there too. And I’d be willing to be there are several writers who don’t sub to Locus.

    I would suggested a value of 3000 NOVELISTS in the field, as a short story writer could sub to Locus and/or be a member of SWFA w/o publishing a novel.

    Fair enough, I wanted something that was obviously way higher than SFWA membership as I know not everyone who writes genre is in SFWA. 4000 was as good as any 3000 just makes me look better, as it pulls down the potential margin of error. I wanted something high enough that no one would accuse me of trying to make this look more statiscally significant than it was, so I went with a ‘worse case scenario’ number…

14. 14. Gerald Weinberg on Feb 8th, 2005 at 1:39 am

    First, as I’m just publishing my first SF novel (after 40 non-fiction books) this is immensely helpful. Thank you.
    Second, there are several ways this survey might be skewed. I’ll mention a few that haven’t been brought up so far:
    a. People who got smaller advances might be less likely to want to report them in a survey. If that’s true, these are overestimates.
    b. People who got large advances might be less likely to want to report them in a survey. If that’s true, these are underestimates.
    c. Agents may not be better at getting money. Perhaps they are just better at spotting and taking on novels that are likely to earn larger advances.

15. 15. Dean Wesley Smith on Feb 8th, 2005 at 1:47 am

    Very interesting work, Tobias. I also liked how you tracked it by both years in publishing and numbers of books sold. A note, however, in sf and fantasy, there are a lot of authors with far more books published than the top of your graph. Did you limit that for a reason or just because that was the most published author you got into the mix?

    It will be interesting to see how this tracks with even more authors on board. (I’ll try to take a minute to fill out your survey as well. )

    Great work. Thanks for doing it.
    Cheers
    Dean

16. 16. tobias buckell on Feb 8th, 2005 at 8:44 am

    a. People who got smaller advances might be less likely to want to report them in a survey. If that’s true, these are overestimates.

    TB: True

    b. People who got large advances might be less likely to want to report them in a survey. If that’s true, these are underestimates.

    TB: True, or both could be happening.

    c. Agents may not be better at getting money. Perhaps they are just better at spotting and taking on novels that are likely to earn larger advances.

    TB: Maybe, but agents have a wider web of connections and deal with publishers more frequently, and have the leverage of other authors in their stable to use in negotiations and understand where they might get more out of a publisher.

17. 17. tobias buckell on Feb 8th, 2005 at 8:47 am

    Very interesting work, Tobias. I also liked how you tracked it by both years in publishing and numbers of books sold. A note, however, in sf and fantasy, there are a lot of authors with far more books published than the top of your graph. Did you limit that for a reason or just because that was the most published author you got into the mix?

    TB: That was the most published author in the mix (35 books) who reported to me.

    It will be interesting to see how this tracks with even more authors on board. (I’ll try to take a minute to fill out your survey as well. )

    TB: Agreed, and thanks for filling it out

    Great work. Thanks for doing it.

    No problem!

18. 18. Nick Mamatas on Feb 8th, 2005 at 10:50 pm

    The other issue, re: agents, is that they do more than get an initial advance. Agents get subsequent advances for paperback or other formats, foreign rights sales, handle excerpts, film options and other subrights, etc.

    My advance for my first novel was X, but thanks to my agent I also earned foreign rights sales X/3, and a serial rights sale of X/5. Not a ton of money, but it took me from under the median to over it on Toby’s scale.

19. 19. Jeff B on Feb 9th, 2005 at 11:32 am

    This is an interesting statistical study, and it would be most useful to people who are just getting started in publishing in these genres. I suspect that, among people who have been writing for a living, the study would say more about how much money they save then how much advance they could get. It’s possible that as a writer builds up a cash reserve they would be likely to trade a small or non-existent advance for a deal that paid better overall. This might explain the data that shows some of the writer’s advances getting smaller as they publish more books. An interesting followup study might include information on whether getting a big advance would actually mean that an author gets less in total profit.

20. 20. Bran P on Feb 9th, 2005 at 1:38 pm

    I would hazard a guess that the $600,000 culprit has initials RJ?

21. 21. Tim Bartik on Feb 9th, 2005 at 2:22 pm

    As an economist who does work with statistics, and reads but does not write science fiction (unless you consider economics to be science fiction), I have three brief comments:

    (1) You should adjust for changes in prices given the range of years represented in the data. Simply download the CPI-U for each year from the BLS website, and multiply each observation by the ratio of the 2004 CPI-U to the CPI-U for the year in question. The data then is represented in 2004 prices.
    (2) Whether the true population size is 3000 or 4000 writers is irrelevant to the standard errors.
    (3) With this type of data, the median is far more relevant. It would also be relevant to look at other percentiles, e.g., 25th percentile and 95th percentile.

22. 22. Nick Mamatas on Feb 9th, 2005 at 2:27 pm

    t’s possible that as a writer builds up a cash reserve they would be likely to trade a small or non-existent advance for a deal that paid better overall.

    Why would they give up guaranteed money for only the potential of more money in the future? That is, even if it were so that the margins of the book trade were sufficient to make royalties as fungible as you’d have it, which they generally are not. Giving a writer a 25%-30% royalty in exchange for no or a very low advance is just not in the cards. It’s also worth noting that as an advance is an advance against royalties, there’s no real gain to be had from delaying payment of them by shifting the money away from the present and into the near future (6-18 months) anyway.

    There are fa more likely reasons why some writers’ advances are trending downwards: they’re not selling as well as they did, their entire category is not selling as well as it did (this is especially true in horror), and they sold poorly enough to lose their commercial appeal entirely, so they sell books to specialty presses with little in the way of liquid to pay out advances.

23. 23. Jeff B on Feb 9th, 2005 at 3:02 pm

    Nick: I don’t work in or even know much about the book trade (and I’m not even sure what fungible means for that matter) so I won’t address that. But, I was under the impression that an advance was, more or less, a loan given to an author so that they could write a book. It is my understanding that in the record industries it is not uncommon for a band to end up owing money after an album is released.

    That being said, a publisher takes a risk by paying an advance on a novel. I assumed that they collected some kind of premium from the author to compensate them for the risk. An author who was confident in the sales of his or her work might be willing to forego paying that premium. If it doesn’t really work that way then it doesn’t. It just occurred to me that this might explain the discrepency.

24. 24. Fred Ollinger on Feb 9th, 2005 at 4:12 pm

    Toby:

    1. Great research. I’m glad that you did this.

    2. Are the people like Jeff B correct in stating that an advance is just like a “loan” to write a book? If this is really the case then the advance just means the amount of confidence that they have in a book, but if this needs to be “earned back” on sales as someone said then in one sense, a big advance is a big risk. The person who got 600k is going to have to sell a lot of books.

    Of course, I have heard that publishers put more money into promoting books where they gave a bigger advance so it is also a self-fullfilling profecy (sp?), but still, if the above is true, then we should be looking at how much is earned in total sales for each book, which is harder to do.

    I thought that you did well with the stats.

25. 25. Fred Ollinger on Feb 9th, 2005 at 4:22 pm

    Sorry to be post happy, but there is a similar article related to music earnings here:

    http://www.firstmonday.org/issues/issue10_1/kretschmer/

    It is interesting to compare results. Due to Toby’s results it’s not so surprising that only 1200 musicians can support themselves full time in Germany.

26. 26. libraryhead on Feb 9th, 2005 at 5:37 pm

    As a sometime publishing insider, I can offer a clarification: a book advance is not a loan. As long as the contractual provisions about delivery and acceptance are met, the advance never has to be paid back, regardless of sales or lack thereof. However, it may still be in an author’s interest to take a lower advance in certain cases. The problem arises when an advance is too large and does not earn out, the publisher’s “expectations” for an author are revised rapidly downward. So, even if you manage to get a contract for the second novel (with a much lower advance, natch) most likely they won’t be making much of an investment in the marketing and publicity the second time around, and you’re looking at a downward slide toward pseudonym city. On the other hand, accept a reasonable advance for the first book, earn out quickly, start cashing those royalty checks, and everyone will be pleasantly suprised. Mad Max at BookAngst 101 has been talking a lot about this lately. Check out http://bookangst.blogspot.com/ for more inside baseball.

27. 27. Seth Breidbart on Feb 9th, 2005 at 5:49 pm

    How did unagented first novels compare with first novels by those who got an agent during the sale?

28. 28. Nick Mamatas on Feb 9th, 2005 at 7:20 pm

    Right, as mentioned above, publishers make no attempt to reoup advances. That’s just money in the bank for the writer, and a cost for the publisher. Nobody would write otherwise.

    It’s also worth noting that the writer’s margin on the sale of a book (royalties) is lower than the publisher’s margin. That means that a book can be profitable even though the advance doesn’t “earn out.” The advance is paid back on the royalty — only after one earns out one’s advance does a royalty check come for units sold after the earn-out. However, as a publisher can be collecting 15-20% of the cover price while the writer is only collecting 4-15% of the cover price, the book can turn a profit anyway even if the author never sees another penny after the advance.

    As far as the idea of taking a lower advance anyway in order to hedge against inflated expectations, advances tend to be low, as the chart shows. A 3000 word article sold to Spin or Rolling Stone beats the median first-novel advance.

    If one does get offered a high advance and the book is a flop, publishing again can be hard, but turning down that high advance and essentially demanding a place on the midlist is:

    a) fairly arrogant, as it is essentially the writer claiming to know more than the publisher about the publisher’s business,

    b) silly, as less publisher risk means that they’ll expend less effort trying to cover that risk, and

    c) fairly self-destructive, as your agent is paid on commission; asking him or her to take a large pay cut because you don’t think your own book can sell more-or-less guarantees that you won’t be a client he or she will go to the wall for.

29. 29. Luc on Feb 10th, 2005 at 2:17 pm

    A quick comment about advances: in at least a couple of places you talk about a difference in agented vs. unagented advances being worth well over the agent’s fee. However, the advances are not total earnings from the book; if the book earns out, the advance amount doesn’t matter in terms of the total payment.

    That said, I still think agents are terribly useful and wouldn’t want to be back in the soup without one.

30. 30. tobias buckell on Feb 10th, 2005 at 3:38 pm

    I’m falling behind on keeping up with comments:

    My advance for my first novel was X, but thanks to my agent I also earned foreign rights sales X/3, and a serial rights sale of X/5. Not a ton of money, but it took me from under the median to over it on Toby’s scale.

    TB: Good point Nick, and many have pointed this out, so I’ve changed the survey to accommodate that kind of information.

    Tim Bartick: thanks for your comments and for your follow up explanations and assistance via email.

    There are fa more likely reasons why some writers’ advances are trending downwards: they’re not selling as well as they did, their entire category is not selling as well as it did (this is especially true in horror),

    TB: SF declined .5% as a percentage of overall booksales to 6% this year, part of a general decline, so that could be part of the trending downwards.

    1. Great research. I’m glad that you did this.

    TB: Thank you Fred. Long time no see, drop by more often!

    2. Are the people like Jeff B correct in stating that an advance is just like a “loan” to write a book? If this is really the case then the advance just means the amount of confidence that they have in a book, but if this needs to be “earned back” on sales as someone said then in one sense, a big advance is a big risk. The person who got 600k is going to have to sell a lot of books.

    TB: Yes and no. You don’t have to give the advance back if you don’t ‘earn out’ so it’s not technically a ‘loan.’ But the person with the 600K will have to sell a lot of books.

    Of course, I have heard that publishers put more money into promoting books where they gave a bigger advance so it is also a self-fullfilling profecy (sp?), but still, if the above is true, then we should be looking at how much is earned in total sales for each book, which is harder to do.

    TB: Right, which I’m trying to fix by adding a category in the survey now that asks about earning out. Not perfect, but a step in the right direction.

    The problem arises when an advance is too large and does not earn out, the publisher’s “expectations” for an author are revised rapidly downward. So, even if you manage to get a contract for the second novel (with a much lower advance, natch) most likely they won’t be making much of an investment in the marketing and publicity the second time around, and you’re looking at a downward slide toward pseudonym city. On the other hand, accept a reasonable advance for the first book, earn out quickly, start cashing those royalty checks, and everyone will be pleasantly suprised.

    TB: Exactly. Thanks!

    How did unagented first novels compare with first novels by those who got an agent during the sale?

    TB: Anecdotally: it helped some of the increase their sales, or got others better royalty rates and kept them more rights.

    However, the advances are not total earnings from the book; if the book earns out, the advance amount doesn’t matter in terms of the total payment.

    TB: Good point…

31. 31. bluejack on Feb 10th, 2005 at 3:51 pm

    Quick question: what do you mean by median? I have always understood median to be the exact midpoint between lowest and highest value, so in a range of 0-40K, the median would be 20K. Are you using it to refer to the most common value, which I have previously heard termed the mode, or modal value?

    Everyone except for me seems to understand what you mean, including that statistician, so I probably have a wrong understanding, but this reference ( http://www.bartleby.com/68/32/3832.html ) does seem to articulate my definitions.

32. 32. Scott Nicholson on Feb 10th, 2005 at 3:55 pm

    Speaking from experience, my first advance was very low yet the book seemed a “success” to the publisher because it earned out so fast, and advances went up accordingly from there (so far). People with the same publisher, and with what I expect were similar deals, have since been dropped from the line. I’m glad for that small advance now, because every book looks like a winner on the ledgers. I have no doubt that bigger advance equals bigger print run equals almost-guaranteed better sales, but any midlist writer better watch that dreaded slip. But I don’t even mind that, because you’re supposed to always be improving and expanding your audience anyway, and if you can’t do that, why should the publisher or readers care?

    There was an SF/F agency roundtable published recently where the agents pretty much agreed the “average” advance was $7500. Don’t have a link for it, but I remember Donald Maass was one of the agents.

33. 33. tobias buckell on Feb 10th, 2005 at 3:59 pm

    http://mathforum.org/library/drmath/view/52755.html

34. 34. bluejack on Feb 10th, 2005 at 4:17 pm

    Ok. Thanks! My math is better now.

35. 35. Scott Nicholson on Feb 10th, 2005 at 4:26 pm

    Here’s the link to the 2003 agent article I mentioned:
    http://www.computercrowsnest.com/sfnews2/03_april/news0403_1.shtml

36. 36. Aaron Schutz on Feb 10th, 2005 at 10:10 pm

    Hi, interesting data. Thanks for putting it up.

    Regarding “margin of error.” I don’t mean to critique, since you have admitted you’re not a stats guru. However, I think it is useful to know that margin of error is calculated with sample size only, as an earlier respondent noted. For example, national polls only sample 800-1000 people in the country, but have a margin of error of ~3-5 points, generally. What “margin of error” generally means is that (about) 95% of the time given random error, the real number in the population is captured within the margin of error. (Thus, 5% of the time the real number is outside of this range). The bigger your sample size, the smaller your error. That’s why medical studies use large numbers of people (they are very risk averse). Of course that assumes that your sample is random. But samples are almost never random and people mostly just “assume” that the requirements of the test that produces the “margin of error” numbers are right. (For example, polls don’t capture people without listed phone numbers or only cell phones, and lots of people refuse to reply. They try to adjust for this, but you can’t say much about what you don’t actually know.)

    Also, with sample #s around 5 or so, I think the statistical assumptions that produce “margins of error” really break down–in other words, you can’t really calculate them. I’d actually advise you to take this data down until you have a more robust sample. Whatever you say, people will walk away with some sense that they mean something, and there’s no way to know if they do. As I remember from grad school, the magic # necessary to apply the assumptions is around 30.

    I’m rusty with stats, however (I’m trying to teach a high school student stats at the moment, but we haven’t gotten to actually calculating errors yet, and I won’t remember how till we do).

    I don’t actually think you need to calculate the “real” margins of error. I think you’ve given an adequate sense of the data with your comments, actually. So this is meant mostly as friendly info.

    Anyway, yoeman work, here. Thanks!

37. 37. Michael Pullmann on Feb 12th, 2005 at 10:48 am

    “I would hazard a guess that the $600,000 culprit has initials RJ?”

    My personal guess was “SK,” but that’s a good one, too.

    In any case, I found this terribly fascinating and informative. Now if only I knew how to write a novel…

38. 38. Terry Whalin on Feb 12th, 2005 at 12:04 pm

    Fascinating work, Tobias. Thank you for putting together this information. I’ve been in publishing over twenty years and haven’t seen this type of information. Great effort. Terry The Writing Life

39. 39. Jemiah Jefferson on Feb 12th, 2005 at 1:11 pm

    Thanks for putting this together - if nothing else, it’s reminded me that I need to get an agent sooner rather than later.

40. 40. Rikk Gned on Feb 12th, 2005 at 7:56 pm

    Firstly, my advice (and I should practice what I preach!), is to finish your novel before worrying about how much you’ll get paid. Write well and true, and you may be one of the lucky ones who get a nice advance and contract.

    Secondly, in regards to the $600′000 advance, there is an author who received such an advance. Cecilia Dart-Thornton, an Australian fantasy author, was awarded a $500′000(australian) three-book contract with only a few paragraphs posted on an internet fiction-sharing site to go on. Now I’m confused — is this an advance, or will this be paid out over the course of several years whilst the three books are being written? I’m guessing the advance she received would have been a percentage of this contract. All I can say is I hope her books are selling well!

    Cheers for the info. It’s sobering and motivating information, and I’m even more determined now to make a living from writing fiction.

41. 41. Cece on Feb 13th, 2005 at 11:47 am

    Tobias…thanks so much for this! It’s nice to know that women’s fiction and/or romance writers aren’t as far short-changed as I’ve been told/lead to believe.

    Cece

42. 42. Eric Plunk on Feb 21st, 2005 at 9:11 am

    Tobias,

    I’m not a member, but just read through this web page that someone sent me via email:
    http://www.tobiasbuckell.com/archives/001423.html

    Would you please post my response for me so others can benefit from it? Thanks.

    I went to a seminar where two agents gave presentations. One explained that advances can turned down by the author by choice. An agent was working a deal for an author. The publisher agreed to buy the book. The agent called the author to find out his/her preference for payments. The author requested the third option below:

    “As much money up front as possible because I’m living off of credit cards.”

    “Some now, and some later.”

    “No advance, and no money until next year. I have enough money in savings, and I don’t want to pay the next higher tax bracket.”

    Thanks,
    Eric

43. 43. tobias buckell on Feb 21st, 2005 at 9:12 am

    Sure, but how is that a benefit to the writer?

44. 44. Eric Plunk on Feb 23rd, 2005 at 12:13 am

    I don’t know how this instance benefited the author. I can only guess that the author approached writing as a profitable business. As such, he or she managed the flow of money and decided that a delayed payment was the best option.

    It wasn’t my intention to submit a example that needed to be justified for its benefiting the author in question. My purpose was to allay any fears of future authors with regards to low advances. Basically, the author has some input regarding how money is paid to him/her. The author is not totally at the whim of an agent or publishing house.

45. 45. tobias buckell on Feb 23rd, 2005 at 8:35 am

    Granted. I’m not sure where I suggested that the author is at the whim of them, and I can see where bringing down a crazy large advance could allay the problem of not earning out, but I fear this could be used as an excuse by authors to ‘aim low’ or get taken by a publishing company that is more of a vanity press. If its a small press and the author knows what they’re doing, sure…

46. 46. Eric Plunk on Feb 23rd, 2005 at 9:48 am

    You _didn’t_ suggest that authors are at the mercy of other’s whims. But that doesn’t stop people from believing so, especially if one is new to an industry.

    It is the agent’s job to make sure that the author is not taken by the publishing house. Your results above confirm this (higher advances for authors with agents.)

47. 47. Michelle .K. Jones on Mar 1st, 2005 at 1:38 pm

    wow………. i never thought there would be a ite like this. im trying 2 become a writer myself. this site is real cool and i’ll tell all my friends about it.

48. 48. Richard on Mar 15th, 2005 at 6:25 pm

    I am new to this game. Have a triology written. Trying to get copywrite info and other info like I just read so I can get published and make money at it. Plus I want to find out how to keep rights such as movie and toys andclothing. Thats all send input

Pingbacks

1. Boing Boing Trackback on Feb 9th, 2005 at 9:03 am
2. Peers and Peerless Trackback on Feb 9th, 2005 at 12:07 pm
3. ctype(yourBrain, myBrain) Trackback on Feb 9th, 2005 at 9:19 pm
4. the0phrastus Trackback on Feb 10th, 2005 at 1:45 am
5. MemeMachineGo! Trackback on Feb 11th, 2005 at 1:17 am
6. Big Dumb Object Trackback on Feb 15th, 2005 at 3:16 am