Re: The True Cost of the Essentials (Implementing Peer Review)

From: Arthur Smith <apsmith_at_APS.ORG>
Date: Thu, 13 Dec 2001 13:57:02 -0500

David Goodman wrote:
>
> It may have cost $1000 to prepare an article for press in 1977.
> Given that the editors and the reviewers services are free, and the
> material is normally submitted in electronic form, how much should it cost
> now?

First note $1000 in 1977 is about $2850 in 2001, according to standard
CPI tables. Some publishers (at least us) pay the scientific editors,
but that only amounts to perhaps $200/article (per published article -
the actual cost is more proportional to total number of articles, so the
number is subject to rejection-rate effects etc.). More publishers pay
for secretarial help and office overhead for the editors, or have
centralized communications offices throughout the review process,
typically another $250-$300 (per published article). Most publishers who
don't just print camera-ready copy have not yet seen ANY net savings
from receiving electronic formats - these can speed up the process and
may reduce error rates, but the multitude of electronic formats,
language issues, font issues, figure/image problems, etc. generally
results in processing expenses at the level of about $50/article JUST to
get and maintain an accurate PDF file for review purposes; further
processing and copyediting typically comes in at somewhere around
$500/article (depending on length), with perhaps a $30-40 savings for
electronic (but note the $50 expense getting the electronic copy
properly prepared in the first place). At least for us copyediting now
includes a significant amount of work in "tagging" the article to an
SGML or XML format, particularly reference sections, allowing robust
inter-article linking. On top of this all is the overhead for
information services pieces (software, hardware, networking to support
the publishing work), managing relations with other primary and
secondary publishers (for interlinking for example now), managing the
money (even if you're "free" you still have to write grant proposals,
manage a budget, justify yourself to whoever is paying those little
expenses you do have), which is probably highly variable depending on
volume etc., but 10% on top of the rest is a minimal estimate, or
perhaps $100/article given the above numbers.

So not counting anything directly associated with print distribution,
subscription management, marketing, or profit, a publisher can probably
expect to be spending $800-$1500 in 2001, of which perhaps $300-$1000+
is directly associated with the copy-editing piece, for every article
they publish. Compared with the $2850 the 1977 number would suggest, we
seem to be getting more efficient over the years. Though the 1977 number
undoubtedly included the things I've discounted (print distribution in
particular, at the level we had in 1977, would come to around
$500/article).

> What does the publisher need to do in pre-press besides copy-edit?

Well, they don't really have to do anything, which could save the
publisher perhaps $500 or so per article. But I doubt there are many
publishers actually spending nothing on the pre-press side: I believe
even Elsevier pays journal editors an amount per article that
corresponds to this - the amount may be more like $150 or $200 and not
cover their costs, but it's something at least.

And we'd love for copy-editing to be cheaper. It may be starting to
happen... but it's definitely NOT an easy transition for an established
publisher, particularly with all the new tagging requirements.

We could save a lot of money if we forced authors to meet much stricter
requirements on what they send us (as far as formats etc. go) and not
give them any of the hand-holding they seem to need on the matter. We
would probably lose a lot of authors (to publish in journals that are
much more expensive for libraries) if we did that, but it would
significantly cut our expenses. We do have some hopes that authors can
be persuaded to shape up in a gentler fashion, but it may take a long
time, and may never be complete.

                        Arthur Smith (apsmith_at_aps.org)
Received on Thu Dec 13 2001 - 19:28:41 GMT