The Perils of Oversimplification: What are the Real Costs of Electronic Journals?

John T. Scott

Paper presented to the ICSU Press Workshop, Keble College, Oxford, UK, 31 March to 2 April 1998


I am retired now and have been out of the field we are here to talk about for all of eight weeks. That means I am eight weeks out of date, and in this business eight weeks is a long time. Some of what I will tell you today may no longer apply. None of it necessarily represents the official view of the American Institute of Physics, although that is where I got my experience during a thirty-year career.

By "oversimplification" in my title I refer to the unfounded assumptions we often meet in discussions of electronic publishing, assumptions that suggest that major savings in the cost of journals can be quickly achieved while at the same time additional features are added. In what follows I will try to show that the situation is more complicated than is sometimes assumed.

The Relentless March of Progress

At this Workshop we are discussing changes being adopted by the scientific publishing business because of new technological developments. I am old enough, however, to remember that we have always been in the midst of new technological developments. Recently I came across an apologia written in 1977 by Bill Koch, a former Director of AIP, for the requirements of the then-new US Copyright Law, which, for the first time, required authors to transfer their copyrights specifically to publishers. (That 20-year-old copyright law is now itself sadly in need of further revision.) "Only with these written transfers of copyright," he wrote, "can AIP and its Member Societies....continue to experiment with new ways of 'publishing' in this world of rapidly changing technology." Yes, that was in 1977. To put it in perspective, recall that the now all-pervasive IBM PC would not be invented until four years later, in 1981. There is no reason to suppose that the march of progress during the intervening twenty years has now reached a pinnacle and will stop. Whatever conclusions we come to this week, however we word our resolutions and shape our agreements, let us remember that they will be out of date very soon, just as I am "out of date" after eight weeks off the job. This is my first "complication."

What is a "journal"?

Note that Bill Koch put the word "publishing" in quotes, implying that he did not know exactly how the word was to be defined under the new developments he foresaw. We still do not know. "Electronic publishing" can mean anything from the so-called "e-print servers" that operate successfully in certain branches of physics and astrophysics and elsewhere, through newly-launched peer-reviewed, quality-controlled journals available only online, to online versions of well established major journals that continue to be available in their print-on-paper editions. The second of the "oversimplifications" of my title, the second complicating factor, refers to the danger of comparing dissimilar products. More specifically, I point to the lack of merit in the argument being posed by some scientists to publishers: "If I can mount a Website containing all the latest work in my field and make it available free of charge to all interested persons, why do your online journals in the same field cost so much?" I hope to show today that what the publishers are offering is much more than you would get from a free Website.

By a "journal" (in any medium) I mean a form of communication among scientists and other scholars that meets the requirements listed in Table 1. Note that I say nothing about medium of distribution, or about the regularity or irregularity of issue. Communications that do not meet these criteria may be fine and necessary things; they have always existed and will continue to exist. But they are not journals.

Table 1
A "Journal" is a form of communication among scientists and other scholars that:

n Is Peer Reviewed

n Is Archived (i.e. readily available in unimpeachable copies for ever)

n Establishes Priority (contributions appear with a "received" date, and are not changed after publication; there is a formal "publication" date, for patent purposes)

n Contains Contributions that can be Readily Cited and Retrieved

n Is Readily Accessible anywhere in the World

For my purposes here I will be discussing "electronic" or "online" journals that are outgrowths of pre-existing print journals, and that continue to appear in print alongside the newer online versions. I am aware that many new "all-electronic" journals are springing up that have no print antecedent. To me it is more significant that these are "new journals" than that they are electronic. I know from experience what it is like trying to start a new journal, regardless of the medium; authors will not send their best work to an unknown, untested new publication whose longevity is still in doubt. For the foreseeable future all the important journals will be the same ones that have dominated each field of study for the last few decades; only the media will change.

Remember that all the journals in a given field compete with each other for authors as well as readers. The best papers go to the "best" journals.

"Cost" versus "Price"

Can we define what we mean by "cost"? Specifically, can we examine the terms "cost" and "price" and agree that their meanings are subjective? The seller's price is the buyer's cost. I believe that what was meant by the planners of this Workshop by the term "cost" is the expense incurred by the users of the journals—librarians or individual readers. This "cost" could be a combination of subscription purchase price and other ancillary expense incurred in presenting the required information to the user. We have librarians among us who can speak to the costs they incur beyond the purchase of subscriptions: staffing, shelving, and cataloguing for print journals, plus hardware, software and technical support for electronic journals.

I, as a publisher, think of subscriptions in terms of their price. The publisher incurs costs, for acquisition, editorial, production, manufacture and distribution, and then sets a price. He strives to keep his costs as low as possible, subject to the need to provide features that will ensure the publications remain competitive in terms of their quality. He sets a price based upon his knowledge of the market and his expectations of return. There need be no direct link between publishers' costs and subscription prices. While we can analyze the components of the publisher's cost (and I am about to do just that) the subscription price is whatever the publisher says it is. This is "complication number 3."

Let's look first at society publishers. Societies come in all sizes, from very large to very small. These days, the large societies have their own in-house publishing programs covering all aspects of the business from acquisition of manuscripts through editorial and composition, to preparation of camera-ready pages for printing and also of electronic files for online distribution, plus the processing of subscription orders. The techniques and technologies used for all of this are the same as those available to commercial publishers, and the unit costs are likely to be very similar. Very small societies cannot afford to set up complete publishing units, and would not find them cost-effective if they did. (I speak here of publications that are meant to be distributed beyond the society membership, to libraries and other institutions.) Small societies often contract with a commercial publisher who agrees to publish their journal and provide free or low-cost copies to the Society's membership in exchange for the right to market the publication to institutions. Again, the production costs will be similar to those of other journals.

What, then, can we say about society-journal prices? First, let us remember the old truism that "not-for-profit" does not mean "for loss." Societies and other associations are free either to operate their journals at or close to "break-even," or to use their publication programs to generate revenue that can support other activities of the organization. The only restriction introduced by the "not-for-profit" designation is that any overall net revenue cannot wind up in the pockets of individuals such as private owners or holders of publicly-traded shares. The societies therefore set prices that will, they hope, produce the desired outcome, which is likely to combine some desire for circulation growth with an expectation of net income.

Commercial publishers set prices according to the demands of their owners or shareholders, who naturally require a maximum return for their investment. These publishers are not philanthropists. They operate according to normal business principles. If their prices are set "too high," whatever that means, they will lose market share and suffer accordingly. The evidence is that they are not suffering. Those scientists who complain that journal prices of some publishers are too high can find the remedy in their own hands; stop submitting manuscripts to these publishers, and stop reading their journals. Meanwhile, the current business model appears to be working well from the publishers' point of view. Manuscripts seem to be in good supply, and circulation numbers (a closely guarded secret) must be high enough to generate net income.

These things being said, it is true that any increase in operating costs that cannot be paid for without threatening the required net revenue must be covered by some increase in income—if not by increased sales, or new revenue streams, then by increased prices. The move to online delivery is just such an event that is likely to put a strain on the relation between publishers' costs and subscription prices.

Additional costs of producing electronic journals

Table 2 shows some of the features of electronic publication that could cause publishers' costs to rise. There are the value-added features that everyone agrees are worthwhile, such as the hypertext links; the need to establish a digital archive, which the publisher will have to do regardless of whether others also maintain archives; additional labor costs due to high-salaried employees in the information-technology area, and the R&D costs of planning for future products. All these features of the new online journals have the effect of increasing publishers' costs. Subscription prices have not yet reflected the full impact, partly because publishers have no idea yet how the new paradigm is going to affect them. Experimentation with different ways of charging for access to the journals is in full flow—there is detail enough in that subject for another Workshop as big as this one.

Table 2

Causes of added cost in electronic publishing

Value-added features:

Hypertext links (including reference linking)
Full-text searching
Mulltimedia, and illustrations in color


Expensive to maintain
Growing over time

Labor costs

IT personnel are expensive
Help Desk required
User Agreements to be negotiated

Research and Development

Technology is not standing still
Future releases and new products being planned

Are there not big savings from customers switching from print to online? Not really, not yet. Libraries are wary of giving up their printed journals, because they are not convinced of the stability of the technology. Individual subscribers (society members, for example) are converting more readily, but their subscriptions are not as important to the total income as those from institutions. The presence of added costs from the electronic aspects, plus the absence of major savings from reduction in print expenses, add up to "complication number 4" for those who would look for big savings with online journals.

Some details

So, what exactly can we say about the details of publishing costs of STM journals, in their online and print versions? Nothing in accurate detail, unfortunately. All publishers resist discussion of their costs, circulation figures, and other relevant business information. I cannot give you this kind of information about any real journal. But I will share with you the complete revenue and expense statement of a completely fictitious journal. The data that will follow correspond to no particular journal, although each line item could perhaps be duplicated in one or another. There may be a lack of consistency between what I have assumed in different parts of these tables. The numbers are not as important as the concepts. Do not take the data literally; rather, let me use the items to point out various details in the journal-publication business as it currently exists. And you can regard these details as making up "complication number 5."

Welcome to the Journal of Applied Kinetic Energy. See Table 3 for a description of this make-believe journal and its publisher. After many years of traditional print publication, the publisher has now added an online version which was at first offered free of charge to current subscribers. Now there is an arcane pricing model for different combinations of print, online and CD-ROM that no-one fully understands. Let's assume that the online subscription income shown is representative.

Table 3

Journal of Applied Kinetic Energy

Published by the Institute for Kinetic Energy Applications
a large organization with inhouse publishing facilities

8,500 pages per year; approx. 350 per semi-monthly issue
1,000 papers published per year; 50% acceptance rate

Member Subscribers: 1,000 @ $100 per year
Institutional Subscribers: 2,500 @ $650 per year
Increasing international readers and authors

100 pages of paid advertising per year

Voluntary page charges requested

Recently added online version

CD ROM available

Tables 4 and 5 show, respectively, the annual revenue and expense statements for this journal.

Table 4
Journal of Applied Kinetic Energy
Advertising Sales $50,000 100 pages @ $500
Page Charges 240,000 4,000 pages @ $60
Reprint Sales 200,000 600 orders @ $300 each
Royalties 50,000 Copyright payments increasing
Single Copies and Back Volumes 38,000  
Current Subscriptions: Member 100,000 1,000 @ $100
Current Subscriptions: Institutional 1,625,000 2,500 @ $650
Online Journal Subscriptions 100,000 500 @ $50 plus 250 @ $300
CD ROM sales 50,000 Mostly to members




Table 5

Journal of Applied Kinetic Energy


Advertising Expense


Inhouse plus P&B; half of income



$50 per page



Major increase with move to online

Editorial Mechanics


Copy editing plus production

Editorial Management


Review of 2,000 mss + inhouse mgmt.

Online Journal Service


I.T. Support Group

Internet Operations


Website maintenance plus Help Desk

Reprint Printing and Mailing


50% of revenue

Journal Printing and Binding


Little savings from move to Online
Paper 150,000
Mailing, Postage 250,000

Billing (page charges and reprints)


80/20 rule applies 
Subscription Fulfillment 35,000



Includes current-customer support

CD ROM production


Equal to income



Net revenue


On the revenue side, note first that advertising does not bring in much income. Generally, journal circulations are too small to attract many advertisers. Experimentation is continuing with advertising on journal Web sites, but it is still too soon for all parties to be convinced of its value. Page charges are sometimes said to be the answer to future STM journal publishing, but we see in this example that they are responsible for only 10% of total revenue. (If there is a strongly competing journal in the same field that does not request page charges, the page-charge revenue is likely to be even lower, and unstable.) "Royalties" includes fees for re-publishing elsewhere, such as in collected works, and also the rapidly-growing payments returned by copyright-clearinghouse services. CD-ROM sales have so far been relatively small, with very few libraries opting for them.

All the inhouse expenses reported in Table 5 include an element of overhead appropriate for each area. There is no separate "below-the-line" overhead. Thus, the charge for composition, for example, includes overhead incurred by that activity.

Advertising expense comes to half of the gross advertising revenue, and is a combination of the cost of solicitation (by sales agents) and the production costs for the 100 pages published per year. This activity is only possible because the publisher has other publications, some of them supported largely by paid advertising, and these 100 pages are a fringe benefit. "Composition" includes not only the traditional task of typesetting pages for print publication but also the preparation of digitized files for online transmission. Very little savings can be realized by accepting authors’ digitized composition files. "Editorial mechanics" covers copy editing and production, plus quality control steps for the online journal that include testing of all hypertext links. "Editorial management" covers the costs of peer review ( two professionals in the field and their clerical support) and also the inhouse management team. The costs of the online-journal service itself and the maintenance of the publisher's Website could easily be larger than is shown here were it not that this publisher can spread the expense over the many publications that share the service. Note that the provision of a "help desk" is a feature that was not needed in the days of traditional print publication.

Printing and binding the print journal, buying paper, and mailing to subscribers all represent areas where savings are possible if readers select "online only" options with no printed journal. In this example the potential savings are perhaps 25% of the total expense, but we should remember that these savings can only be realized if all subscribers agree to do without the printed book. So long as even one printed copy is required, substantial setup costs will be incurred. Also, that potential 25% is calculated on a total that is already increased because of the additional costs of electronic publishing.

Billing and fulfillment, the two areas where a publisher meets his subscribers face-to-face, can be expensive because of the 20% of all transactions that take 80% of the time. Marketing, when an electronic product is added, includes much customer education to make sure that all the options are adequately explained. CD-ROM production costs approximately equal the revenue from sales, so this feature does not contribute to the net income.



The lesson I would like to draw from this example is that the addition of an electronic version to an established print journal is certain to add to the publisher's costs. A solely electronic journal might be produced for about the same cost as the equivalent print product, but there is not yet sufficient evidence to show that subscribers (especially librarians) are willing to give up their printed journals.

The price of a journal subscription is another matter; again, we do not have sufficient evidence yet to show us that publishers will pass along all their increased costs to subscribers or not. Prices could go either way.

One thing we can be sure of is that tomorrow's journals will have features we have not yet imagined, and the effect on their price is equally unfathomable.

Last updated : 06 July 1998
Copyright 1998 ICSU Press and individual authors. All rights reserved