The Tip Jar as Revenue Model: A Real-World Experiment (Internet 1.0 Perma-free Book Publishing Technique)

Peachfront’s Note: This is a reprint of the classic Tip Jar experiment first published on and widely discussed all over the internet. First date of publication was April 27, 2003 on I’m not sure a writer would get the same outstanding results today, with dozens if not hundreds of free and perma-free books appearing on Amazon and other distributors every day!

* * * * *

Localroger’s Note Dated April 27, 2003: It’s been three months since I posted my novel online for the world to read for free, with a tip jar as compensation medium. I think we have enough data now to tell how well it might work for other artists.

I will summarize the aspects of my experience which are most relevant to any artist publishing a work:

  • Distribution
  • Feedback
  • Revenue

Distribution is the number of people exposed to the work. For some artists this is more important than revenue, since they regard spreading the message as its own reward.
Feedback is how the artist learns how well his work was received; knowing that N people saw your work is not the same as knowing that 0.8*N liked it or that 0.5*N think it should be burned. Again, for some artists this is more important than revenue but it might also be more important than mere exposure; the artist might want to feel that he positively influeced the readers, or made them think.

Finally there is Revenue, the question of whether your art can pay the rent and keep the wolf from the door. Most artists would prefer not to focus on this but the circumstances of life often force us to give it top priority.


Tracking my distribution is pretty easy since Rusty kindly configured a Webalizer report for the novel directory. Although there is a lot of ambiguity in web statistics we can learn a lot from these data.

As anyone who has ever set up a website knows, simply putting it online does not bring readers. My novel has benefitted from several major sources of publicity:

  • Rusty’s Jan 12 introductory article on K5
  • A Jan 14 mention which Rusty solicited on Cory Doctorow’s site BoingBoing
  • A Jan 16 comment thread on Slashdot started by readers under, ironically enough, a front page article introducing Cory Doctorow’s free online novel
  • A Feb 7 ani-gif banner ad on low-tech left-wing political site which I bought as an experiment
  • A Feb 21 (drumroll please) front-page review on Slashdot
  • Subsequent blogspace activity, revealed in the referrers or by google search

These publicity sources were fortunately pretty much separate from one another so we can see how effective they were. In judging readership from the Webalizer report, there are two main download threads. I tend to regard anyone who downloaded all eight chapters as a reader, so the stats for the least-popular chapter file (mopi1.html through mopi8.html) bear some resemblance, if not exact, to true readership. In addition the all-in-one files including the two versions of the .zip archive ( and mopi.ZIP) and the all-in-one HTML file (mopiall.html) represent a mixture of people who have and haven’t already read the chapters, and who will or won’t actually read what they’ve loaded.
There is a very consistent trend that about half the number of people who load the index load chapter one, and about half the number of people who load chapter one load all eight chapters. I have used this in some cases to tease results out of the referrer logs.

So taking into account all the fudge factors, I officially and by fiat declare that I have the following readership results so far:

  • From Rusty’s intro: about 1,000 readers
  • From BoingBoing: about 100 readers
  • From the Slashdot comment tree: about 100 readers, which is interesting since this was a comment thread beneath a story that sank pretty quickly
  • From the Bartcop ad: about 300 readers
  • From the Slashdot review: about 3000 readers; this might be much higher since Slashdotters were much more likely to load the .zip and /all versions
  • Overall, including all sources: between 5,000 and 10,000 readers

I am defining “readers” here as people who read most or all of the book. This is a little harsher than even publishing industry statistics; if you paid money for a printed copy, took it home, and couldn’t finish it this scheme would call you a “non-reader” even though you paid. But it’s a more accurate determination of an artwork’s influence on the world.
So one could say that MOPI has effectively sold out a typical first printing run for a non-best-selling author. However, this is a little deceptive; the “marketing” I received was very targeted, and benefitted a lot in the beginning from the reputation I have for my other writing on K5. It has also been read all over the world, from Japan to the Netherlands to South Africa to Slovenia. So in many respects I consider this a best-case distribution; most people would not have gotten so much distribution so quickly. It would also be more difficult if instead of a 250K fileset it was a 30 megabyte set of songs.

On the other hand, distribution by regular publication is glacially slow by comparison. If I’d signed a contract with a publisher on Jan 12 I would still be reviewing galleys and the first books would probably be on shelves toward the end of Summer, if that soon.


I have received about 200 private e-mails about the book. It’s very easy to describe them: They almost exactly mirror the tenor of comments which were posted after Rusty’s introductory story.

The particular work at hand was never designed to be “friendly.” It is just about guaranteed that any particular reader will find something within it revolting or insulting. In this regard it is very similar to some Indie music which is trying “not to be Pop.” The feedback reveals that this work does not appeal to everybody, but it does appeal to a significant number of people a lot.

Because of the way the book is, this is a thing that was hard for me to find out any other way. Before Jan 12 exactly eight people had read it, but that isn’t much of a sample size and they all knew me. Now that thousands of strangers have read it I have much more confidence that it accomplishes what I thought it did.

If you peruse the google search you will see a similar pattern. The fact that the K5 comment, e-mail, and blogging response sets are all very similar convinces me that the novel is not riding on my previous efforts or my reputation on K5.

Again, had I published conventionally it might have taken me a year or more to find out how well my book worked before a live audience. But electronic publication has the advantage that feedback is always just a click away for the reader — either to tell me to go to hell or that I changed his life.

Nearly all of the mail I’ve gotten has been positive, some of it a little embarrassingly so. I credit this to the fact that someone driven off by the first pages of chapter one probably wouldn’t bother writing at all. I have received a few negative reactions, such as this one. The comments following the Slashdot review were also more negative, but they were dominated by people who had obviously only skimmed the first chapter so as to post quickly.

(I’m frankly glad a few of the negatives bothered to pipe up, because if they hadn’t I would seriously have to wonder about people :-)


As I write this I have received about $760 in tips from 85 people. Thanks guys!

About half of the tip money came from the initial introduction, indicating either that K5’ers are extraordinarily generous or that I was getting extra bonus points for my other work here. K5 also generated my largest tips, three of $20, a $30, and one of $45. (Did I mention that some people seem to like the book a lot?)

Slashdot accounted for most of the rest simply because it accounted for most of the rest of the readers. It is worth mentioning that some people, especially in the Slashdot crowd, tipped me just for trying the model out. I got several “haven’t read it yet, but here’s $1 on principle” messages.

An especially interesting data point is the Bartcop ad for which I paid $80. (I consider this an experiment and I don’t account it against tip revenue.) This was supposed to guarantee 15,000 hits and all indications are that I got a fair click-through for my dollar. But as far as I can tell I didn’t get a single dollar in tip money. Even if a couple of small tips came from Bartcop click-throughs it doesn’t come close to covering the ad, which was extremely cheap by Web advertising standards.

A typical royalty deal from a publisher is 10% of gross, which would snag me $0.50 to $2.00 per reader depending on what kind of book it was. On the other hand much of my distribution depended on it being free; most of my readers didn’t bother to pay. (This is not a complaint, it is just an observation.) So it’s hard to tell if my royalty revenue would have been higher with conventional publication.

One thing I can tell, though, is that I didn’t make enough to cover a traditional advance. Most non-best-selling authors never pay out their advances anyway, so that’s the real metric for publication payment. You generally get $5,000 for an “ordinary book” contract. This is a sensible amount for a work that is expected to take months to complete.

Noted from a comment: I put the tip jar on the top index page so as not to impose on readers. Some may not have realized it was there or forgotten by the time they read the whole thing. So I may have been a bit too polite and discreet by not reminding people that the opportunity was there when they finished the book.

In my case I have no complaints; I never expected to make even the $760, never expected to have more than the eight readers. Because of the book’s “difficulty” I was not able to get it in over the transom at a real publisher when I tried in the late 1990’s. This way I have readers, I have feedback, and I know the book does what I hoped it would do. And all this happened in a very short time.

On the other hand I could not honestly advise anyone to try making the Internet tip jar their income source for the rent and groceries. All indications are that I got extraordinary results, and it just wasn’t enough.

The Future I: Free *

The Free movement (as in Free software, music, art, and so forth) has work to do. As of now the tip jar model doesn’t work economically as a substitute for conventional publication. (Free publication does have its upside, which I have especially appreciated since I am not depending on an advance to pay the rent.) But the next generation of artists — new artists — are the ones most in need of some kind of patronage. If they aren’t supported they will become CPA’s and forget about their art. We need a better plan.

The Future II: Me

On the novel site I made an offer to self-publish the novel in book form if I got enough support. As of shortly after the Slashdot review I made that goal, and I am keeping the tip money aside to fund this project.

However, the extraordinary success it has had in distribution and the feedback have convinced me to follow the advice of some of my readers, and try again to have it published conventionally. Right now I have a very highly regarded agent who has expressed interest in reading it, which is under normal conditions difficult to arrange. So my Web adventure has already put me in a place I couldn’t get to when I didn’t have thousands of readers, at least one of whom is an editor.

But we are back in the glacially-slow conventional publishing space and now that I’ve synopsized it and formatted the first three chapters appropriately and sent it off it will be a month or two before I know whether I have a shot at being represented. (Yes, the agent wanted a standard submission: “I can’t (won’t) read a whole novel off the screen.” And they make the rules.)

I will give this process until the end of the year. At that point if I haven’t made any progress I will self-publish the damn thing since I promised I would and I hate not fulfilling a promise. I’ll keep the ball in motion in conventional land but the self publication will not help my cause there, which is why I am delaying it.


Please don’t leave a comment complaining that this article doesn’t include a link to the novel in question. It’s not about plugging my novel, it’s about the process and the results.

* * * * *

Peachfront’s PS: That’s the end of the original 2003 article. A short time later, Roger published The Metamorphosis of Prime Intellect, and…SF history was made.

The Road to Schenectady — Part 3

Peachfront’s Note: This is part of a series Roger is writing on how writers get their ideas, based on his own experiences. Here’s Part 1, and here’s Part 2. Today he’s going to talk about the role of community, especially the community, and how it encouraged him to create.

* * * * *

So far I’ve written mostly about the things I wrote for myself which happened to find a public audience because the site existed where they could be published, if not for money, then with an expectation of a wide audience if they survived the very democratic upvoting process.

Rusty Foster also published The Metamorphosis of Prime Intellect on his servers outside of the usual kuro5hin voting process in response to strong community support for the idea. There was a period of time around 2002 when Rusty openly mused about having a “stable of writers” he would support outside of traditional publishing venues. There were at the time dozens of frequent kuro5hin contributors who might have been candidates for such a venture.

Of course, it was not to be.

What I have not written about much are the things I wrote only because kuro5hin existed as a likely outlet. All of the Passages stories except the first, The Happiness Broker and its sequel, the popular Plant essays and even the series that forged my reputation on kuro5hin, A Casino Odyssey were all written because kuro5hin existed.

I suppose it’s easy to say that makes me an egomaniac who only cares about being fluffed but from my end of the telescope it seems a lot more complicated.

Some of my stuff got a pretty hostile reception, like my libertarian-bashing sendup of Yet Another Effort and my musings on astrology and mysticism, but still I wrote up the next idea that seemed worthy.

I only stopped when one of the trolls created a votebot to make it impossible for even a popular and anticipated story to be legitimately voted up, denying me access to the front page.

I find writing easy but it’s still an extra effort to take a cool internal vision and turn it into something that holds up for another person who might not share my background or interests. So if it wasn’t to fluff my ego, why did I write for K5?

As with the very different compulsion to write MOPI I’m honestly not sure. It’s as if some ideas have a life of their own and aren’t content with living inside of my head. But like people they differ in their level of drive; MOPI wanted out no matter what but The Fifth Gift would have been content to stay a personal fantasy if I hadn’t been asked to write something “positive about transhumanism.”

I suspect I’m not the only one like this. Kuro5hin and its Scoop-based progeny seem to have brought a lot of people like me out of the woodwork.

Rusty Foster created and its software engine Scoop as a reaction to the unpleasantness of what was then the most popular tech blog on the Internet, Rob “CmdrTaco” Malda’s The slashdot story submission process was believed to favor clique members and the not very well community-moderated comments were a cesspool of trolling and linkbait.

Scoop was Rusty’s answer to this situation; kuro5hin went live in late December 1999 and managed to draw off a healthy following of Slashdot refugees, eventually including myself. I made my grand entrance with A Casino Odyssey in July 2001.

In June 2002 Rusty launched a fundraiser which brought in over $37,000, and for awhile there was talk of expanding kuro5hin’s scope to create a nonprofit foundation to explore the possibilities of community edited media. It was in the wake of this that I published Passages and then MOPI there. But the foundation was never quite able to get its footing, kuro5hin stayed pretty much the same, and there was a sharp dropoff in story queue submissions after 2003.

Of course, kuro5hin was not the biggest success of Rusty Foster’s Scoop software, and his other successes probably helped doom kuro5hin by drawing his attention away from it. In October 2003 Markos Moulitsas decided to move his DailyKos blog from Movable Type to Scoop.

“Kos” didn’t enable the story voting feature that had attracted me to kuro5hin, but he did leave in place — while openly wondering if it would serve any useful purpose — the Scoop “diary” feature that allowed users to post unmoderated stories that appeared on a sidebar.

The diary feature turned out to be wildly popular, with many vibrant sub-communities forming around popular writers and meeting in the comment sections of their diaries. DailyKos went from being the most popular Democratic political blog to being the “largest progressive community blog in the United States” with an estimated 2.5 million unique viewers a month.

Other Democratic politicians took note, and when I made my way to Maine and met Rusty in person in 2008 his main source of income turned out to be consulting on Scoop blogs for Democratic political candidates. This involved a lot of travel and apparently not much time for administering kuro5hin.

Some writers whose stars like mine had flared on kuro5hin now migrated to a scoop site set up by k5 member “Hulver.” But HUSI never attracted the wide readership or incredibly high Google pagerank that kuro5hin had enjoyed.

(Amazingly, kuro5hin has incredibly high pagerank even today despite being little more than a hollow shell of its former self for more than half a decade, a phenomenon nobody seems to be able to explain.)

Others migrated to sites like Everything2 or Wikipedia, or just dropped off the face of the Internet as far as anyone could tell.

Sam “zenofchai” Montgomery-Blinn started a science fiction magazine, Bull Spec, which bought the reprint rights to publish my story Mortal Passage in print form.

In my own Internet travels I have never found a home comparable to kuro5hin that inspired that urge to write down and refine what might otherwise have been idle fantasies. I have noticed that being a popular writer on a big site like DailyKos is pretty much a full-time job, one I don’t have time for since I have a real full-time job.

Places with low barriers to entry also tend to have low readership and participation. The new trend toward “social media” like Facebook has filled the Inter-ether with hucksters intent on gaming the search engines and ranking algorithms, a game that interests me not at all.

While I wrote a lot of stuff because kuro5hin existed as a place to put it, in a sense I still wrote it all for myself.

If I wanted to game a system, I would have made a very different decision in 1994 when I cranked up Bank Street Writer and asked myself how I wanted to record the vision with which I’d awakened of a depraved future where immortality was creating a new and profound type of insanity. Should I write it for the editors, I wondered, bowdlerizing and over-explaining it to make it publishable, or write it as my heart wanted to in all its terrible splendor?

I wrote that first chapter of MOPI, and everything since, the way I would have wanted to read it myself. The fact that others have liked any of it has been nice. In some cases some of the criticism has given me pause to think and humbled me a bit. But it would not be a localroger joint if I dolled it up to make it look profitable for the suits or jacked it with buzzwords and shout-outs to enhance its SEO.

* * * * *