Is Safari Books Online a good thing for authors?
by Robbie, May 19, 2006
[UPDATE: I’ve modified this post toward the end in an effort to clarify my questions and concerns about Safari.]
As an author or co-author of ten books, I pay close attention to my quarterly royalty statements. One line-item I’ve always questioned concerns the sales from Safari Books Online. Safari offers an online subscription service that gives individuals access to a limited number of books or corporations and academic institutions unlimited access. Flat-fee subscription services like this are not ideal for authors that are accustom to royalties. There is a fixed pool of money to spread around a variable number of authors.
The money I’ve earned from Safari has always been considerably less than my other royalties. And by a long shot. This concerns me because at least according to Tim O’Reilly, Safari is now O’Reilly’s #3 reseller:
Part of what we’re building with Safari is a channel. We have resellers in libraries, universities, and corporate settings, as well as a strong base of direct subscribers. And that channel is becoming more and more significant. In the most recent 12 months, sales of O’Reilly books through Safari exceeded sales from Borders, making it our #3 reseller behind Amazon and Barnes & Noble, with about 5 times the revenue of our direct sales from oreilly.com. Assuming that we had the same results as Pragmatic, adding PDFs to oreilly.com would more than double our direct sales, but that would still leave it far behind the level of sales that we get through the distribution channel that we’ve built with Safari.
I decided to gather some data to see if this really was an issue or just my perception. I took three of my O’Reilly quarterly statements from the past year and compared the money I earned from royalties versus what I received from Safari. Each line represents a different book:
% from % from Royalties Safari ----------- --------- 99.28 0.712 84.03 15.96 89.93 10.06 94.24 5.755 94.81 5.183 100.0 0.00 92.89 7.10 94.74 5.25 93.56 6.43 98.78 1.21
If you take the average, it comes to 94% of earnings from royalties and 6% from Safari. Does it seem right that the #3 reseller for O’Reilly would only contribute 6% of my earnings? (According to Tim O’Reilly in the comments, this is reasonable)
I’ve heard several other authors complain about this too. Part of the problem is that Safari is not transparent. How exactly are they calculating how much money should go to each author? How much are publishers getting? I can understand how they’d compute earnings under the individual subscription programs, but how does Safari compute earnings for the unrestricted corporate and academic subscriptions?
Flat-fee subscription models encourage readers to access as many books as they want for one price (which is a great thing for customers). For example, instead of reading all of one O’Reilly title on HTML, a customer of the corporate or academic Safari subscription could choose to access a small subset of ALL O’Reilly HTML titles. This results in less revenue for each author, but the same revenue for O’Reilly.
I want to make it clear that I’m not bashing Safari or O’Reilly or anyone else here. I’m just trying to understand something that isn’t very transparent today AND I’m trying to figure out how big of a deal it is to authors. The easiest internet business model (and most attractive to customers) to implement is a flat-fee subscription. However, there isn’t much incentive for authors with this model and I question if a newcomer could pull it off like Safari can with the O’Reilly and Pearson names behind it.
Robbie — A couple of answers to your question. First, you need to realize that Safari is the only reseller that is broken out separately, so of course it looks small compared to the whole. Even our largest reseller represents less than 20% of worldwide sales. Safari is under 10% of our worldwide sales, so your 6% isn’t unreasonable, especially considering the information I posted in the Radar entry Long Tail Evidence from Safari and Google Book Search. A disproportionate percentage of access comes from books that are no longer selling in print. There are many books for which Safari royalties are the only royalties.
And I’ll note that just as in the physical book marketplace, not all books are equal. Some get more uage than others. I note that in your list above, your Safari royalties range from nearly 16% of the total down to 0, with your average being 6%.
As to how Safari royalties are calculated for unrestricted subscriptions, it is based on usage: the revenue divided by the share of total page views. It’s how revenue is allocated to the publishers, and from there to the authors.
Tim O'Reilly May 20th, 2006
Tim, thanks for the response. There are a couple of problems. First, the fact that Safari enables sales for old titles doesn’t really help authors as you imply. If Safari is the only source of royalties, it is unlikely that the author could even buy lunch with the earnings (per quarter!). However, if enough customers access these older titles, it is a boon to the publisher (you).
Second, there is still a transparency problem with Safari. Let’s look at the Safari Basic program for individuals ($14.99/month). Say a customer uses one of the 10 “slots” for one of my books. That generates $1.49 in royalties. Out of that, how much goes to Safari, how much goes to O’Reilly, and how much goes to me? For most of my books, I earn over $2.25 per book through the standard print channels. Isn’t it a problem that the price of the book has no bearing in Safari? I bet I earn much less than half my print royalty through Safari per book.
Mind you, the Safari individual program is the BEST CASE for authors. For the Safari corporate/academic solutions, I bet my take per book is much less (given the unlimited access) and I’d also bet that Safari and the publisher’s take is much higher. Any details would be appreciated.
Robbie May 21st, 2006
Very interesting. I am an avid fan of Safari Bookshelf and have often wondered how they deal with royalties.
Tormod Guldvog May 22nd, 2006
Robbie — As stated in your contract, you get EXACTLY the same deal with Safari as you get in print. You get your royalty rate times the amount of money we receive.
And you talk about the “lack of transparency” regarding the details of each transaction as if you have more visibility into print sales. The number that is reported for print sales is a summary of thousands of individual transactions, which might include direct sales at full list price, sales to individual bookstores at a single-copy discount (I don’t remember exactly, but it’s around 48%) versus sales to the big guys, fully loaded with all their discounts, at maybe 55-57% discount. I don’t hear you saying that sales through Barnes & Noble are unfair because they take a bigger slice than a direct sale.
Yes, there are variations in the amount you might receive from a Safari B2C subscriber who keeps your book on his bookshelf for a year versus the amount you receive from one-time access by a B2B subscriber who has access to the whole library, but in each case, you get exactly the same share of the amount that we, the publisher, receive.
And I’m sorry if you’re unhappy that authors of older books, or books on technologies that are more heavily adopted by the safari user base, get a bigger piece of the pie. But that seems like an odd argument, since exactly the same thing happens in print. Some authors do better than others. This isn’t unfair. This is the market.
Tim O'Reilly June 13th, 2006
P.S. I don’t know why you keep saying “I bet my take per book is much less (given the unlimited access) and I’d also bet that Safari and the publisher’s take is much higher,” despite being told repeatedly that you get the same share regardless of how much the customer pays.
When a B2C account pays hundreds of thousands of dollars for access to a full library, that’s a good thing for authors, because that money is allocated to publishers (and to authors) in proportion to usage — NOT per title. If someone accesses your book, you get a share of that month’s revenue in proportion to your share of all the books accessed.
Tim O'Reilly June 13th, 2006
Tim,
You are glossing over the main issues, so let me try again:
1) Safari is not transparent, which you still haven’t addressed. You say Safari royalties are calculated based on the revenue you receive. Why not show the details of that to authors? Show how much comes from B2C and B2B. This might help alleviate some of the concerns of authors. And you are right, the print world isn’t terribly transparent either, but do you really want to use that as a defense? At least with print, authors have other relative measures of performance such as Amazon/B&N/Bookpool rankings. With Safari, all we have is an unqualified royalty figure in a quarterly printed statement. It is frustrating that publishers think the “trust us” model of transparency is the default way to do business.
2) The Safari model (especially in the B2B case) encourages customers to move freely between books and access only what they need. This is GREAT for customers for sure. Print is very inefficient. When customers buy a computer book, it is unlikely that they will read it cover to cover. It is much more likely that they are interested in only a subset of the information. Authors have benefited from this inefficiency (customer has to buy the whole book). Safari eliminates this inefficiency at the expense of authors (but not publishers). The net-net is that Safari customers will access less of more O’Reilly titles. For example, instead of reading all of one O’Reilly title on HTML, a customer of the B2B Safari subscription could choose to access a small subset of ALL O’Reilly HTML titles. This results in less revenue for each author, but the O’Reilly revenue stays the same. That is the problem with flat-fee subscription models like Safari B2B.
BTW, I’m not sure why you are getting defensive about this discussion. I NEVER said I was “unhappy” about other authors getting a bigger piece of the pie or that it is “unfair”. You just wrongly suggested that Safari enabling more sales in the Long Tail was a good thing for authors, which is not true. It is a good thing for O’Reilly, but not really for authors (at least from a monetary perspective).
Robbie June 13th, 2006
rrc> 2) The Safari subscription model (especially in the B2B case) encourages
rrc> customers to move freely between books. This is GREAT for customers for
rrc> sure. Print is very inefficient. When customers buy a computer book, it is
rrc> unlikely that they will read it cover to cover. It is much more likely that
rrc> they are interested in only a subset of the information. Authors have
rrc> benefited from this inefficiency (customers have to buy the whole print
rrc> book).
This has been my point for a long time. And a change in
revenue model should eventually lead to a change in how one
thinks about writing a book. Now the trick is to write a
collection of chapters that readers *will* want to read
cover-to-cover, and that they *will* want to keep on their
bookshelf for a long period of time.
Please take note that I am not saying that the Safari model
is bad in any way. I am simply saying that it influences my
own thinking about what to write.
In print, an author benefits from selling a large,
high-priced book even if readers don’t need all the
chapters. In Safari an author benefits from selling a book
that readers will keep on their bookshelf for a long time.
As Ken Hess pointed out, this is actually similar to the
change that MP3 players have brought to the music industry.
Many people buy individual songs now, and not albums.
FWIW, I can envision a point in the computer “book” industry
where the unit of record is something small, analogous to
our current “chapter”. It would be interesting to write my
next book as a series of stand-alone “booklets” on, say,
SQL. Each “booklet” could stand alone, and begin bringing in
revenue, but I could also ensure that the *series* of
booklets flowed well. Some booklets would sell better than
others, and it would be interesting to see how that played
out. Based on commonly-bought chapter groupings, collections
of booklets could be created (we could call these
books) to target different audiences. For example, we
might find that there was a market for a collection of
booklets on the more introdutory aspects of SQL. We might
find there was a market for the small collection of chapters
on weird, vendor-specific extensions to SQL. Etc. Publishers
could add value by collating well-thought-out collections of
booklets.
Getting back to your blog entry, I agree that the Safari
subscription model (and probably any subscription model)
changes the dynamic of revenue flow from reader to author.
And yes, because a reader will still pay Safari their $14.95
monthly fee, it is probably true that the effect from a
reader changing out books averages out for Safari in a way
that it doesn’t for an individual author. But that’s just
the way things are. It would be a mistake, IMHO, to try and
force a new way of doing business to generate the exact same
revenue dynamics as an old way of doing business.
Jonathan Gennick June 14th, 2006
Let me weigh in as the CEO of Safari Books Online to try to clear up one myth that is emerging on this discussion regarding the difference between the B2C & B2B models in Safari. Facts are that there is very little difference between content consumption in the two models. We have seen that users tend to access the same number of books monthly regardless of the model. In the B2C model (limited size bookshelf), users tend to have between 6-10 books on their bookshelf in any given month. The number varies based on subscription size (smaller subscriptions obviously have fewer books on the bookshelf). The subscription fee gets split up amongst the books on the bookshelf. In the B2B model (whole library access with no bookshelf restriction), users tend to look at the same number of books each month (between 6-10). The monthly subscription fee gets apportioned to these books, just like in the bookshelf model. The high degree of correlation here has to do more with how people use the service than the model (bookshelf vs. library). Users of the service are generally very project & reference oriented (as opposed to “readers”). These users tend to look at content that is specifically related to the project at hand and stay within a very limited subset of the overall corpus. Users that traverse the site and look an a significant number of books are few and far between. Moreover, we charge more for the whole library access than the bookshelf access because it offers flexibility. So, the actual result is that authors earn more off the B2B subscription that the B2C subscription because a higher monthly fee is spread across the same number (6-10) of books.
We started the business with the bookshelf model but then found that the whole library model resulted in an ability to create higher revenues without diluting payments to publishers and authors. In fact, we will be introducing the library model to our B2C users in September. Our users have told us that they value a model that has the bookshelf restriction removed and that they will pay to have that type of access. If our experience with B2B is a predictor, this will result in more revenues for both publisher and authors.
Sean Devine August 29th, 2006