Robert J. Sawyer

Hugo and Nebula Award-Winning Science Fiction Writer

Evo and Rob on the size of online audiences

by Rob - July 21st, 2006.
Filed under: Uncategorized.

In response to another posting of mine on this blog, Evo Terra of and the Dragon Page responded to my comment that online serialization of a first SF novel would probably get the author fewer than 100 readers; in his posting here he says that’s not true — thousands of readers can be reasonably expected. He’s also put a different version of his comment (lowering his claimed number to just 1,000 readers in the process) on the Dragon Page, and there brings up the writers John Salzi and Cory Doctorow, both of whom I admire greatly, as examples to prove his point. Have a look at what Evo said, then come back here for my response to him:

Now you’re saying, well, let’s ask the two most successful examples of online text distribution how well they’re doing as an indication of whether or not the advice I gave to an unknown, first-time novelist was sound or not — which would be not unlike me saying, “Well, let’s check J.K. Rowling’s numbers to see how a first-time fantasy novelist can expect to do.” :) John Scalzi and Cory Doctorow are exceptions, literally; their experiences have been exceptional, and are not the norm.

And even when Cory does talk about this, we get soft numbers from him; Cory usually cites the number of printings his books have gone into — six for his most-successful one to date, all in trade paperback, which, of the three common book formats [hardcover, mass-market, and trade] has the lowest threshold for economical reprinting, instead of the actual number of copies sold.

He does know that figure; he just doesn’t share it. But it’s on his royalty statements — and royalty statements, in fact, don’t list number of printings (because they’re meaningless, since a printing has no fixed size — a trade paperback reprinting could easily be and often is 1,000 copies), so he’s giving us the public number [anybody can see what the printing number is on a book], and is withholding the private number.

There’s an evangelical quality to a lot of what’s said, including your comments here, Evo, about online publishing and podcasting, but the hard numbers that could be disclosed by the evangelists but are not otherwise available to those who might dispute their rosy portrayal often are withheld in favor of soft metrics or vague statements.

Cory and John, incidentally, are excellent writers with very significant, long-established personal online presences. And, yes, if you’re so good that you can be a finalist for the John W. Campbell Award for Best New Writer, as they both were/are, or a winner, as Cory already has been, or a Hugo finalist for your first novel, as John currently is, then maybe you, too, can be an exceptional data point.

But reality-checking demands that if you’re going to shore up your argument with best-case data, you need to also include the worst-case data … or else throw out all exceptional highs and lows, and just look at the middle; the wonderful success of John and Cory needs to be counterbalanced by failures or simply ignored in the equation. As with the testimonials in weight-loss ads, you need the asterisk that says, “Results not typical. Your results may be much less.”

There’s room in the cultural landscape for one, and maybe even a few, wunderkind SF pop-culture prophets who got in on the ground floor a la Cory; there isn’t room for dozens, let alone hundreds — so implying “You can be the next Cory Doctorow!” isn’t what I’d call sound career advice. :)

And one wonders, given the plug-pulling recently done on Tor’s attempts to join in with the Baen ebook distribution system (the plug was pulled by Tor’s parent company, which has, rightly or wrongly, profound concerns about paying to package text in print form for the bookstore trade that is also freely or cheaply available online) if anything like the circumstance of being bought from online serialization will ever happen again. People are quick to turn aberrations (the two cases you cite) into trends, and to cite exceptional data as the norm; that’s no way to build a business or a career.

And you’re still pulling punches, Evo, as far as numbers are concerned. You’ve got 58 podiobooks at your company; you say 17 have a thousand listeners (and ARE you in fact defining that the way I did — at least half the episodes downloaded?), but then, instead of answering my question about how many are SF novels, you go back to vague terms: “the lion’s share” of the 58 (not the 17, at least not on the basis of anything you’ve said here) are, you say, science fiction (and not fantasy).

Surely you know how many of the 58 are SF, and how many of the lucky 17 are; we don’t need to fall back on a generalization. :) Precisely how many data points — how many SF novels serialized at Podibooks that have had at least 2,000 people (to support the claim you made in my blog) or 1,000 people (to support the claim you made on the Dragon Page) download half the chapters? What we know for sure is that there are NO MORE THAN 17 data points, and, for the 2,000 threshold, probably many fewer. If we’re talking about a single-digit number of cases, then let’s directly say that.

And then — and I say again — 1,000 people downloading something for free does not mean 1,000 listeners or readers. Just because they grabbed it doesn’t mean they do anything with it. What percentage of people who grab something for free actually read/listen to it?

Well, we can get some figures, I suppose: on, readers are asked to tip the author/publisher if they like what they’ve heard. Of the subset of the 17 works that have had 1,000 downloads of 1/2 of their installments (if indeed that’s what you mean when you say 1,000), how many of those 1,000 people have actually left a payment?

Here’s a hard number of my own. My novella “Identity Theft” has been online for free at for four months now (since March 20, 2006). In that time, it’s been “read” 5,258 times, which sounds quite impressive. (“Identity Theft” has always been free at Fictionwise because, during its entire life there to date, it’s been an award finalist, first for the Nebula, and now for the Hugo.) Now, first, we should look at the verbs here: “Read” or “listened to” is what you’d say, apparently, but, in fact, we don’t know that.

Over at Fictionwise, the number (which Fictionwise terms “Units sold” on the royalty statement, but that’s a misnomer too, since with a freebie nothing has been sold) actually only measures how many people have clicked a couple of online buttons, adding the item to their online bookshelves — that’s step one of a three-part process in going from Fictionwise customer to actual reader.

Step two, how many of those 5,258 customers have actually downloaded the work from their online bookshelf to their PC or PDA or eBook reader for reading, is a figure I don’t have access to (although I presume does).

And then there’s step three: actually reading the work (or actually listening to it). Granted, lots of paper books sell, too, that no one ever gets around to reading — but at least the author got paid for those; they don’t for freebies at Fictionwise or for anything at Podiobooks (unless the customer elects to offer a tip, something presumably quite rare for works that haven’t been listened to).

So how many actual readers are there of “Identity Theft” through Fictionwise? Who knows? But it’s a subset of a subset of 5,258.

And how many actual listeners are their of works at You say the fact that requires registration is significant; so does, though. When you say 1,000 listeners, you can’t possibly know that. What you mean (I presume) is 1,000 people who’ve hit the free “Subscribe” button on, so the real number of listeners, again, is a subset of a subset of that figure.

And even if you mean 1,000 people went to step two — actually downloading, not just accumulating an OPML list at of things they might someday download (the equivalent of Fictionwise’s server-side bookshelf), there’s still a further reduction: how many people actually listened to it? And that figure you just don’t know, but it surely isn’t 100%.

Since this all started when I was giving career advice, as posted on my blog, I should point out what I and the person I was corresponding with knew: that the serialization he’d been offered was one in which he was going to be paid an advance (of a couple of hundred dollars) by the site in question.

That means, one presumes, that the site in question was going to be SELLING, not giving away, the online text. How big a drop off in “readers” or “listeners” can one expect when one starts actually charging for online material? Many people will grab any freebie they see, but may not actually do anything with it. If they pay for it, I suspect the actual usage rates — the percentages that go to stage two, and then stage three — are much higher.

Well, here’s some more hard data: in the same period — the last four months — during which 5,258 people added my “Identity Theft” to their online bookshelves at Fictionwise — my bestSELLING (as in actually being sold) title at Fictionwise racked up just 23 sales — that’s one-half of one percent of the number who grabbed the freebie.

And I do want to bring this back to where it started by reminding you that, in any event, you’ve done nothing now to disprove the figure of fewer than a hundred readers I suggested; all the data you’ve provided (which amounts to one hard number now: 17) has been related to audio podcasts given away for free.

I stand by my advice, and I’ll add one more piece: when deciding what to do to build a career in science fiction or fantasy, be conservative in your financial predictions. You’re less likely to end up disappointed.

Leave a Reply