I’ve just had three very busy days with friends, colleagues and clients at the London Book Fair. I spent time with every company in my portfolio, and had a great discussion with my publisher about my forthcoming book (watch this space for more). While the show is still fresh in the mind, here are a few reflections—less a comprehensive account of the Fair than a set of personal observations, raw signals and conversations that struck me over the last few days.

1. The Book Fair is simultaneously very large and impossibly small: for a major event, there is a tremendous amount of serendipity. A couple of the best conversations I had came about through cancellations and spur-of-the-moment conversations. After decades of this, I am still guilty of overscheduling my time and not allowing enough time for the unplanned.
2. In the same way that London is both one city and a collection of villages, the Book Fair is a single event and at least six linked constituencies: trade, academic and children’s publishing, rights, authors and technology. Many attendees stay in one of those lanes for the whole of the show. One of the joys of a portfolio career is that I am involved in all of them. I can’t recall an LBF where I met such a broad range of people, and some of the most interesting ideas emerged where the streams crossed.
3. The sheer number of publishers and volume of new titles is almost beyond comprehension. It’s never been easier to produce a book or harder for it to find an audience. And yet production and distribution factors outweighed discoverability in many conversations. Incidentally, this isn’t a book-specific issue: I saw the same dynamics at games events such as UKGE and Dragonmeet last year.
4. William Goldman wrote of Hollywood in the 1970s: “not one person in the entire motion picture field knows for a certainty what’s going to work. Every time out it’s a guess and, if you’re lucky, an educated one that no-one knows anything.” The same seems true for much of publishing today. There are far more lagging than leading indicators.
5. Per recent press coverage, trade non-fiction remains particularly challenging. I heard about gratifying exceptions to that rule, but the common denominator was that the author brought the audience to the project, not the publisher. That’s a shift for many authors who would historically have relied on their publisher for publicity and visibility. Not everyone can make that shift successfully. And if the publisher’s role is concentrated on the (more commodified) production side of the process, they are more vulnerable to being disintermediated by platforms or smart self-publishers.
6. The publishing industry remains staggeringly opaque to even smart, generally well-informed outsiders. I spoke to someone from another industry visiting for the first time, and an overseas publisher visiting for the first time in nearly fifty years, and both were bemused by aspects of what they had seen.
7. To that point, there is a considerable level of procedural debt in the business: processes that haven’t changed in decades, or in the context of the Fair, meetings that happen for no other reason than that they have always done so. The introduction of AI to workflows will probably expose this more clearly than any previous technological change.
8. On the other hand, I was really encouraged by the people I met who are thinking differently: setting up new publishing businesses, or publishing information in different formats.
9. Talking to several of those entrepreneurs about their experiences, we all agreed that the Book Fair takes on a completely different character when you’re there for your own business rather than a corporate. There are no performative meetings: every half hour slot has to earn its keep. But, related to point 1 above, the shape of the curve can turn on a single meeting.
10. The conversation around AI in publishing is changing. Three years ago, the prevailing mood seemed to me to be indifference. Two years ago, it was concern or outright hostility. A year ago, that had evolved into curiosity. The conversations I had this year were strikingly practical and business-like. It’s less about whether to use AI, and more about how to, taking account of the very real issues.
11. Notwithstanding that, publishers’ experience of AI is, like William Gibson’s view of the future, unevenly distributed. There were significant differences between academic and trade publishers, and between younger, nimbler businesses and (with a couple of honourable exceptions) some of the bigger publishers.
12. Many publishers are actively licensing content to AI companies, and it was enormously encouraging to see the launch of the PLS opt-in AI licencing scheme.
13. The ability to talk a good game about AI and answer specific questions about experience and use cases are not well correlated. In conversations, I kept coming back to Seth Godin’s fabulously blunt and revealing question: yes, but what have you shipped?
14. For a lot of publishers, the choice of AI tools is less informed by specific capabilities and more by alignment with existing enterprise IT. There’s an undeniable administrative expedience to using the default AI model that is aligned with your email and office software, but that doesn’t make it the strongest or smartest choice.
15. The qualitative improvement in AI-voiced audio is unreal, but the potential of the technology to increase the proportion of titles available in audio is being held back by the largest audiobook platform, Audible, imposing restrictions on publishers. In my opinion, this is less about good consumer experience, and more about preserving a dominant competitive position. Cory Doctorow has a word for this.
16. From a conversation with a very smart tech company CEO about the experience of our university-age children: the first courses any further education institution should be offering are Metacognition and Critical Thinking 101. Those would be more durable lessons than any technology-specific training.
17. Interestingly, several people noted a renaissance in blogging and newsletters (owned rather than borrowed audiences, of which my writing here is a very modest example). But in many cases that hasn’t been matched by renewed interest in commenting—with the exception of LinkedIn and Substack, which seem to overindex for audience reactions.
18. This was an end of an era for London Book Fair in West London, with the show moving eastward to Excel next year. Only one company that I spoke to was strongly opposed to the move. Most people seemed positive. Roll on 2027.

You can prove me wrong on point 17 by sharing any experience, reflections and challenges below.