Fewer people read longer topics. But it’s not something to lose sleep over, and certainly not something to shorten topics over.
Tom Johnson has a recent series of posts on topic length (Does DITA Encourage Authors to Fragment Information into a Million Little Pieces?, Do Short Topics Make Information More Findable?, and Why Long Topics Are Better for the User). The discussion around these posts dwells, as all such discussions seem to do, on the question of whether fewer people will read longer posts/topics/articles/etc.
I believe this concern is misplaced. Here’s why.
First, yes, it is absolutely true that the longer a post/topic/article/etc. is, the fewer people will read it.
The reason some topics are longer than others is that some tasks and some subjects are more complex and require more explanation and exposition than simpler topics. A properly constructed topic on these more complex subjects will necessarily be longer. The more there is to say, the more words it takes to say it.
The more difficult a task is, the fewer people will attempt it. The more complex a subject is, the fewer people will study it. The pyramid narrows towards the top. That is just life. Fewer people read longer topics because fewer people attempt the more difficult tasks and more complex subjects that those topics deal with.
Splitting up the topics that deal with more difficult tasks and more complex subjects into smaller topics will not make the tasks easier or the subjects simpler. Indeed, by fragmenting the information that the reader needs, it will make them harder.
It is, of course, true, for just this reason, that a long topic will suggest to the reader that the task it describes is difficult to do or the subject hard to learn. This could make the reader abandon the attempt before they even begin. An unnecessarily long topic is always a bad thing, since it scares people off even attempting to read it. A topic should never be longer than it needs to be to fulfill its purpose.
But reducing a topic on a complex subject below its proper length is worse, because even though the reader may initially attempt the task, they will get stuck, and then blame either the docs, the product, or themselves, none of which are desirable outcomes. A topic should never be shorter than it needs to be to fulfill its purpose.
This is not to say that the business of right-sizing a topic is easy. Minimalism tells us that, where it is possible to learn by exploring and doing, that is most peoples’ preferred way to learn, and the way they will approach the learning task regardless of the style of documentation you give them. A minimalist style that supports discovery, rather than attempting to be systematic, is therefore preferable in many circumstances.
But whatever the appropriate level of guidance is for a particular audience, the guidance required for longer and more difficult tasks and more complex subjects will still be longer than for simpler tasks and subjects. More people will attempt the easier tasks and study the simpler subjects, and therefore more people will read shorter topics.
Unless you are being paid by ad impressions, however, getting the most people to read each topic is not your goal. Rather, your goal is to make your products more valuable to your customer base, and that includes the people at the top of the pyramid who do the hard tasks and study the difficult subjects as well as the people at the bottom who do the easy stuff.
If your product is used in a business setting, it is quite likely that the people at the top of the pyramid are responsible for facilitating and making more efficient the work of the people at the bottom of the pyramid. Because those people are fewer in number, the longer topics that they rely on will be read less often, but those topics are actually delivering more value to their organization than the short topics that are read more frequently at the bottom of the pyramid.
I’ve noted before that documentation analytics may mislead. Here is another case of the same problem. You cannot measure the importance of a piece of content by the number of times it is read. Sometimes a topic is read infrequently because it is read only by mavens who pass on its information through their networks. Sometimes a topic is read infrequently because it covers an issue that occurs rarely — but may be very serious and costly to fix when it does occur. And sometimes a topic is read rarely because it covers a task that few people attempt or a subject that few people learn — but those who do may add great value to their organization, and may be the ones who enable everyone else to be productive.
In short, the number of times a piece of content is read tells you exactly nothing about its value or its importance. Some content is read often, but delivers little or no value on each reading. Some content is read occasionally, but delivers huge value. Unless you multiply the times a topic is read by the value it delivers, simple readership numbers are telling you nothing but a dangerous lie. (Yes, this means you have to multiply a data point that is easy to get by a data point that is hard to get before you have a meaningful metric. But no one said good metrics were easy. Or if they did, they were mistaken.)
A good Every Page is Page One topic should cover a well defined subject or task for a well defined reader, and it should be exactly as long as it needs to be to do its job. If a topic does it job, and only its job, it is the right length, regardless of word count. Harder tasks and more difficult subjects demand longer topics. Fewer people will read those topics. And that is as it should be.