Growth

Growth

Many of the large, publicly traded commercial publishers recently released first-half (H1) results. Despite investor fears regarding AI and ongoing geopolitical tumult, the picture presented by these reports continues to be one of a vibrant and growing market.  

RELX reported revenue growth of 6% and adjusted operating profit growth of 8%, as compared to the same period last year for its scientific, technical, and medical group (Elsevier). The publisher reports submission growth to the portfolio of a staggering 20% and published output growth of 7%.  

Springer Nature meanwhile reports revenue growth of 6.2% (factoring out Scientific American, which they divested earlier this year) and adjusted operating profit growth of 7.7%. They reported even more stunning submission growth of 30% and published article growth of 13%. (They offer an estimate of 8% published article growth for the market overall.)  

Informa also reported positive (though more tempered) results for its Taylor & Francis group. The publishing unit reports revenue growth of 5.4% as compared with 2025—if one-time data licensing revenues are excluded (if these one-time sales are not excluded, revenues are down 4% year-over-year). The British publisher reports submission growth of over 20%.  

Wolters Kluwer reported 5% revenue growth and 9% adjusted operating profit growth (on a constant currency basis) for its Health division. Its Medical Research group (where its journal portfolio is managed) reported 6% organic growth (no further breakout of the division is available).  

(Note: Wiley operates on a different fiscal calendar, so we have not included them in this roundup). 

While there are some caveats around one-time windfalls and currency effects, large professional and academic publishers are reporting 5%–7% revenue growth for the first half of the year. That is pretty good, all things considered! An erosion in revenue due to AI appears unfounded, at least so far. That said, investors remain cautious. Share prices of these companies continue to diverge based less on financial performance and more on perceived AI risk. Such risk is presently concentrated less in content and events and more in software and analytics. The equity prices of both Informa and Springer Nature’s are up 2% year-to-date. Meanwhile, stock prices for RELX and Wolters Kluwer remain down 10% and 20%, respectively, this year.  

There are signs, perhaps, of a thawing in investor sentiment related to software businesses. Shares in Salesforce, a bellwether of sorts for software-as-a-service, have rebounded this summer response to strong earnings.  

The other notable point in the H1 reporting was the double-digit submission growth seen by these publishers. If we use Springer Nature’s 8% estimate for industry-wide article growth, it appears that the larger publishers are outpacing their smaller competitors. A question we have, however, is the nature of this submission growth. How much of this growth in submissions is a result of paper mills and other bad actors, abetted by the ease of generating papers with AI tools? Compared with other publishers, are these large publishers receiving a proportional amount of these papers or a disproportionate one? The growth in published papers from these publishers is much lower than the growth in submissions. This could mean they are becoming more selective, or it could mean they are receiving (proportionately) more bad papers.  

The proportions aside, it is almost certain the largest publishers are receiving, in absolute numbers, the largest volume of problematic papers (simply due to their scale). Which is all to say, AI’s largest effect on the industry so far appears to be in the submission queue rather than the income statement. This is a problem the largest players are best equipped to afford, however. They can spread the costs of research integrity screening tools over a larger revenue base. They can also better support research integrity teams of trained professionals to handle investigations. AI, in other words, may be (further) reinforcing the advantages of scale. The real AI risk to the industry may not be to the largest publishers, but to those without the resources to manage AI’s growing burden on the publishing process.

Behind the Mark

The rollout of ACS’s rebrand has sparked passionate conversation and a Change.org petition. The petition is at 1,412 verified signatures as of today (September 15). This is still only about 1% of ACS’s 150,000+ membership—growing, but not exactly a mandate. But, as we know, sometimes the internet rewards the loudest voices.  

It is natural to have a connection to the logo of an organization in which we are a member. For some chemists, the phoenix and Liebig bulb were the visual shorthand for their professional identity. Changing an iconic symbol can feel like a personal affront (which is the sign of a great brand—no one complains about changes to brands they don’t care about).  

What often gets lost in these conversations is that a lot of thought and expertise go into these branding decisions. Branding, like many areas of marketing, is an area where there are professionals with a high level of expertise.  

No organization undertakes a rebrand lightly. It is a massive amount of work to develop, and often even more to roll out. ACS describes an extensive two-year process in which “ACS members, customers, employees, and partners participated in market research, workshops, listening sessions, and feedback opportunities.” That process matters. The people that ACS consulted would not all have shared the same needs, experiences, or preferences. The role of the branding team was to interpret and synthesize that input—not simply follow the strongest reaction from any one group. ACS must also think about its future and how its identity will connect with the next generation of chemists. 

Creating a logo today also requires an understanding of digital environments—considerations that did not exist when ACS’s previous logo was created. A modern visual identity must work across multiple websites, apps, social platforms, and other digital channels, as well as print publications, signage, and merchandise—in varied contexts and at dramatically different sizes. Favicons are another important consideration. A favicon is the small icon that appears in browser tabs, search results, and, increasingly, the source cards accompanying AI-generated answers. A favicon typically isolates one recognizable element from the larger logo. That is much easier to do when the logo contains a simple, distinctive symbol, as the new ACS logo does. The ubiquity of favicons is one reason organizations are moving toward simpler logos.  

There are also user experience considerations. Cleaner, higher-contrast logos can be more accessible for people with low vision. ACS has cited accessibility directly as a consideration behind the new identity.  

Another reason brands are moving toward simpler logos is processing fluency. Decades of psychological research, including a landmark 1998 study in Psychological Science by Reber, Winkielman, and Schwarz, has found that people tend to respond more favorably to information that is easier to process. This does not mean a simpler logo is automatically better. But it does mean that ease of recognition is a legitimate design consideration, particularly when viewers may have only a fraction of a second to register a mark. It is another example of the expertise behind what can appear to be a simple visual decision. 

Branding is hard. The result may look simple, but behind it is research, expertise, judgment, and a great deal of hard work. We just wanted the ACS team—and any other marketing team embarking on a redesign—to know: we know that.

Briefly Noted

Dolly Parton’s death was announced this past month. Her contributions to music and culture are widely recognized. Less well-known, perhaps, are her contributions to publishing and science. Parton founded Imagination Library, a not-for-profit that has distributed over 300 million books to children. Working with local community partners, Imagination Library ships an age-appropriate book each month to enrolled children. The program operates in the US, Canada, Australia, Great Britain, and Ireland. Parton’s contributions to science, cataloged in this Nature article, include a $1 million donation to Vanderbilt University Medical Center in March 2020 that supported early development of the COVID-19 vaccine.  

Members of the American Diabetes Association—including Steven Kahn, editor-in-chief of Diabetes Care—have taken the apparently unprecedented step of launching a boycott of a major scholarly society that extends to meetings, grant and manuscript review, journal submissions, and journal management. The dispute began after five researchers, including Kahn, were removed from ADA’s June Scientific Sessions while distributing copies of an editorial published in Diabetes Care criticizing cuts to NIH programs. ADA said its review found that the attendees were removed for violating the ADA Attendee Code of Conduct during the meeting rather than the content they were distributing, then later said additional findings to be presented to the Board might “clarify or correct” information previously provided. ADA also said that its programs and journals continue without interruption and that manuscripts are being accepted and reviewed. The developments raise unusual questions about how a dispute between a society and members of its scientific community can affect not only its publishing program but also the wider field: because Diabetes Care is a major journal, any sustained disruption could affect the communication of diabetes research. 

In “AI and lawsuits” news, textbook authors have filed a lawsuit against OpenAI and Microsoft. The suit claims OpenAI trained on pirated textbooks and stripped out the copyright info. The suit notes that the threat to textbooks is different (worse) than that of trade books. That is because the person who picks the book (the instructor) isn’t the person who pays (the student or school). If a free AI version is “good enough,” the payer (the student) has every incentive to take it. In other words, the threat goes to the heart of the course adoption model fundamental to the textbook business.  

Google launched Expert Intelligence in Gemini Notebook. Readers can now drop a purchased ebook into a Notebook and interact with it directly (e.g., ask questions grounded in the text, and generate quizzes, audio overviews, and infographics). The new service launches with over 100,000 titles from major publishers (Bloomsbury, Macmillan, Penguin Random House, O’Reilly, and others). The launch also included content from 15+ bestselling authors who curated custom “Featured Notebooks” around their own books and additional custom content (e.g., intro letters). This is a rare AI announcement that is actually welcome news for publishers. Expert Intelligence requires users to (gasp!) purchase a book before using the service. While a positive development for book publishers, Google’s relationship with the publishing industry remains as much Mr. Hyde as Dr. Jekyll. As we reported last month in The Brief, a group of publishers and an author filed a class action lawsuit against Google in July (Hachette, Cengage, Elsevier, Scott Turow v. Google) alleging the tech giant trained Gemini on books that that the publisher supplied for other purposes (e.g., participation in Google Books, Google Play Books, and Google Scholar). 

Speaking of Google lawsuits, Judge Amit Mehta (who in a separate case ruled Google holds an illegal search monopoly) heard arguments in August on Google’s motion to dismiss the Penske Media (the publisher of Rolling Stone and Variety) lawsuit over AI Overviews (Penske v. Google). Google called AI Overviews a “product improvement”; Mehta wasn’t sold, saying the situation “seems really unfair” and noting product improvements aren’t automatically immune from antitrust scrutiny. While there is no ruling yet in the case, the tone is an encouraging sign for publishers arguing Google’s market power lets it take publisher content for AI without paying. In addition to being unfair to copyright holders, it is unfair to competitors like OpenAI and Perplexity who are licensing publisher content. 

London startup Inherent, founded by DeepMind alumni, released an AI agent called Faraday that claims to have outperformed Anthropic’s Claude Opus 4.8 and OpenAI’s GPT-5.5 at one specific task: independently rerunning the code or analysis behind a published ML/AI research result and reproducing its reported figure, under a time/compute budget, without seeing the original figure. Notably, Faraday runs on a much smaller model (27B parameters) than the frontier systems it beat.  

Ex-OpenAI chief product officer Kevin Weil is seeking to raise $150M at a $750M+ valuation for a stealth startup focused on gathering scientific data to train AI models (a similar idea to Periodic Labs, which raised $300M in September 2025). Another example of capital and top AI talent rotating out of general-purpose chatbots and into science-specific AI. 

***
Let’s share this dream that all children should grow up in a home full of books. —Dolly Parton, official welcome letter for Imagination Library