The first two things to understand about optimizing book copy for SEO that I’ve learned from Logical Marketing partner Pete McCarthy are:
1. Copy always used to be written based on “knowledge of the book”. It should now be written based on “research into the audiences”.
2. Copy from publishers was almost always B2B, intended for intermediaries in the marketing and supply chains. Now all copy ends up being B2C, important for consumers and search.
These are tough hurdles for the most established publishers to jump, because executing on the audience research piece not only requires a change in the workflow and distribution of work among staff, doing that also requires additional effort that employs skills that may be in short supply, if they exist in the publishing house at all.
With all that additional work in front of them, any automated solution that can be offered to publishers to diminish the pain of this transition is attractive. (We’re working on a few of those ourselves.) I’ve recently become aware of the new technology offered by Trajectory, which examines a book’s text for words that should be used for SEO.
Automation is good, but this is still coming from within the book. So I asked my Logical Marketing team to help me understand: is this helpful?
The answer seems to be, “it is a positive step but a very partial one. And it does not help anybody avoid the research we think is fundamental.”
The article by Jim Bryant of Trajectory on the Publishing Perspectives blog demonstrating the value of their capability is very clear. He used as one example a book called “The Mayo Clinic Diet”. A word cloud Trajectory created from the book showed clearly that there were two big words, “calorie” and “exercise”, which appeared frequently in the book, were important to its thesis, and were not in the copy the publisher created to describe it. That’s the positive step.
But Logical Marketing methodology is to find out what words the consumer uses to describe what is in the book, particularly in search. With a backlist book, this can be easily researched in Library Thing. The word cloud at Library Thing for this very same book says the terms that the consumers use are “diet” and “dieting” (which, being in the title, were already in the descriptive copy), “cookbook” (not in), “health” and “health and fitness” (not in), “nutrition” (not in), “weight loss” (which was in the original copy; hey! it is a diet book!), and “exercise” (in because it was identified by Trajectory.) “Calorie” apparently didn’t figure as important to the consumers on Library Thing as an associated term.
But going beyond the important terms not in the book and therefore never to be found by the Trajectory methodology, were those other prime terms — “cookbook”, “health”, “health and fitness”, and “nutrition”. And a bit more work on our part also identifies longer-tail terms that will be more important for discovery than what Trajectory found: “healthy diet”, and “lifestyle changes” among them. You get that very clearly from Library Thing.
Pete McCarthy did a post for the DBW blog recently that used “To Kill a Mockingbird” to show why the words that are in the book are not sufficient for first-rate SEO. (Pete used a fiction title because it is often said that fiction is harder to SEO than non-fiction. The point is that the methodology still works.) What is that literary classic about that a semantic examination of the text probably won’t tell you? Racism. Civil rights. It is also important Southern Literature. As an aside, you can combine certain of those tags to find comparable other authors for specific aspects of the book that will resonate with certain customers. You know another author who jumps out with that analysis? John Grisham. Will Trajectory’s method show you that? We don’t think so. They have built sophisticated technology to analyze writing style and story structure, but whether that would connect Harper Lee and Grisham is very doubtful. Using the characterizations of consumers connects them very clearly.
We don’t mean to dismiss Trajectory. We might well learn to incorporate it into our methodology in circumstances where it is available to us (if a publisher we’re working with has it), but almost certainly for new titles only. It is not worthless to examine the text of a book looking for clues to good keywords. But it is a mistake to ignore McCarthy’s first rule: that the descriptive copy that serves you best for SEO requires research into the audience, even gaining knowledge as rudimentary as in the example. Semantic examination of the text automates what we’ve always tried to employ manually: knowledge of the book. That’s useful, but it is really the smallest part of a much larger job to create descriptive copy optimized for search.
To modify an old aphorism, knowledge gained from the text is often unnecessary, but never sufficient.
As for backlist, the important terms that the Trajectory examination uncovered are found much more easily and quickly at Library Thing. The tool that was well-built to begin with, has existed and been iterated upon and had data added to it for years, and been underutilized for a long time because it was seen as a “niche consumer” capability, is still the best one. It is populated with terms used by real people not employed by the author or the publishing house. Its data is not based on transactional history, but on reading history. That’s what you really want for SEO. New titles are a bit harder, and Trajectory over time will likely solve pieces of that puzzle, but audience research is still a lot more important than an examination of the book’s text to achieve the desired result of having readers who are unaware of a book but would be likely to want it have it put in front of them for consideration.
We recollect that Amazon had the full text for many books and used to do a semantic analysis of it. (They certainly still do X-ray in Kindle, which is a version of it.) They haven’t emphasized it and they haven’t built on it. That suggests that the commercial value of it is questionable. And Google also identifies “common words and phrases” and provides a word cloud, which you can see here for “The Mayo Clinic Diet” if you scroll down.
And, indeed, we’re dealing with Trajectory 1.0. Over time they may make their analysis increasingly relevant and useful. But no matter how smart and sophisticated you are, you can’t examine the book itself to find out how people reading it react to it and think about it and that is the information that tells you most about who might read it next.