How much does it cost to typeset a scholarly article? PDF, HTML, JATS XML and EPUB
When a scholarly journal draws up its budget, editorial production usually appears as a single line: ‘typesetting’. Yet behind that line sit several distinct tasks — PDF, HTML, JATS XML, EPUB and rounds of corrections — paid for in staff hours or supplier invoices. Understanding that breakdown is the first step towards knowing what it really costs to typeset a scholarly article, and whether your journal is overspending.
In this article we explain where the cost of producing an article comes from, with reference figures you can adapt to your own situation, and how to calculate your journal’s annual cost.
What producing an article involves
Once the manuscript has been accepted, the article has to be converted into the formats that are published and deposited in databases:
- PDF: the typeset version in the journal’s design. It is still the format authors read most and the one readers expect.
- HTML: the web reading version, which works better on mobile devices and makes indexing by academic search engines easier.
- JATS XML: the structured format required by databases such as PubMed Central, SciELO and Redalyc. It is the most technical format and the one that accumulates the most errors when produced by hand.
- EPUB: the e-book format, increasingly important for accessibility.
On top of this come the rounds of corrections: every change requested by the author or editor must be applied to every format, and keeping them all in sync is what takes up the most time.
How many hours each format takes
Timings vary widely depending on the complexity of the article (tables, equations, figures) and the team’s experience. For a standard article of 8 to 12 pages, these are the reference figures we use in our savings calculator:
| Task | Approximate time per article |
|---|---|
| Typeset PDF | 3 h |
| HTML | 1.5 h |
| JATS XML | 2 h (or around €2.50 per page if outsourced) |
| EPUB | 1 h |
| Each round of corrections | around 15 minutes per format |
With two rounds of corrections, an article published in all four formats adds up to roughly 9.5 hours of production. If a journal publishes 100 articles a year, that is around 950 hours: more than half a year of one person’s full-time work.
How to calculate your journal’s annual cost
The basic formula is simple:
Annual cost = articles per year × hours per article × hourly cost
At €25 per hour, the 950 hours in the example amount to around €23,750 a year on production alone. If the JATS XML is outsourced to an external supplier, replace those hours with the per-page rate.
Some costs do not show up in this formula but are worth bearing in mind:
- Delays: every week an accepted article waits to be typeset is a week less of visibility and citations.
- XML errors: badly formed XML can be rejected by databases, forcing the work to be redone.
- Reliance on one person: if only one person knows how to typeset, any absence brings publication to a halt.
- Formats that never get published: many journals give up on HTML or EPUB for lack of time, which limits their visibility and accessibility.
Three production models and their costs
1. In-house typesetter
This gives you control over design and deadlines, but the cost is fixed and depends on one or two people. It is the typical model for university journals with a stable output.
2. External supplier
You pay per article or per page. It is flexible, but every round of corrections means back-and-forth by email, and JATS XML is usually invoiced separately.
3. The editorial team itself
Very common among small journals: the editorial team typesets as best it can, often in Word. The cost never appears on an invoice, but it is paid in hours not spent on scholarly editing.
Where the savings are
Most of the time goes on repetitive tasks: applying the design, tagging the XML and carrying each correction over to every format. That is precisely what can be automated.
With Index Production, the team uploads the author’s Word file and gets the PDF, HTML, JATS XML and EPUB at the same time, in the journal’s design and in sync with one another. Working time becomes the time spent reviewing and approving: around 25 minutes per article, plus about 5 minutes per round of corrections.
In the example of 100 articles a year in four formats, that means going from around 950 hours to around 58: an estimated saving of almost 900 hours.
Calculate your journal’s costs and savings
Every journal is different. In our savings calculator you can enter how many articles you publish, how you currently handle production and which formats you publish, and adjust the hourly cost and the number of correction rounds. In 30 seconds you will see an indicative estimate of the hours and money you could free up.