Split a blog post when it is trying to fulfill two promises that a reader could reasonably search for, understand, and act on separately. Keep it together when the second part is necessary to complete the first answer.
That is the two-promise test. It is more useful than choosing a maximum word count because length does not tell you whether an article is focused. A short post can still pull in two directions, and a long post can remain one complete, coherent answer.
Find the promise hiding inside each section
Read your working headings and finish this sentence for each one:
> After this section, the reader will be able to…
If every answer moves the same reader toward the same outcome, you probably have one article. If the answers divide into two independent outcomes, you may have two.
Imagine a draft called “How to Start a Blog and Build an Email List.” The setup sections help a reader choose a platform and publish. The email sections help the reader choose an opt-in and follow-up path. Those subjects can work together in a broad guide, but they also make two separate promises. A person who only needs the email-list answer should not have to travel through the entire setup article to reach it.
The issue isn't that the draft is long. The issue is that it has two front doors.
Ask whether the second answer depends on the first
Before splitting, test the dependency.
- Does the reader need the first part to understand the second?
- Would separating them create repeated setup or missing context?
- Does the title promise both outcomes, or did the second one sneak in while drafting?
- Can each proposed article end with a complete, useful next step?
If the second answer cannot stand without the first, keep them together and improve the transition. If it can stand alone and serves a distinct search reason, separate it.
This protects you from the opposite mistake: turning one connected explanation into a trail of thin posts that make the reader click repeatedly for the basic answer.
Use the arrival test
Picture two people arriving from search.
Reader A asks the question in the first half. Reader B asks the question in the second half.
Would either person feel that the title brought them to the wrong article? Would one have to skip half the page before the answer begins? Would the article's introduction have to speak vaguely so it can accommodate both readers?
If yes, the draft is probably hiding separate promises.
Choosing one reader before you outline prevents much of this problem. If the audience is still blurry, use this guide to choose one reader and one situation before you decide whether to split the draft.
Split without breaking the thinking
When you do split, don't just cut the document in half.
Give each new post:
- its own direct-answer opening;
- only the context that reader needs;
- a complete method or decision path;
- an ending that resolves its promise; and
- a link to the other post only when it answers the reader's next question.
That last step is a bridge, not a repair. The first article must still work if the reader does not click.
You may not be able to add the link immediately if the companion article is still a draft. Keep a note of the intended relationship, publish each post through its own review process, and add the link after the destination is live. My guide to internal linking for small blogs explains why the reader's next question is a better linking rule than a generic “related posts” list.
Know when one article should stay whole
Keep the material together when removing a section would leave the title promise incomplete.
A post about choosing a blog platform may need a short section on cost, ownership, and portability. Those are not automatically three more posts. They are decision criteria inside the same choice.
Likewise, an instructional article may need setup, steps, success signals, and troubleshooting. Those sections do different jobs, but they serve one promise: helping the reader complete the task and know whether it worked.
The cleanest test is the ending. If each half can answer a distinct title and finish with a useful outcome, you likely found two posts. If one half ends with “now go read the rest before this makes sense,” you probably cut through one answer.
Split by promise, not by length. The goal is not more posts. The goal is fewer wrong doors between the reader and the answer.
