Blog Writing

Verified against ChatGPT · 2026-08-08

Build a blog outline by diffing against what's already ranking, not from a blank template

Turns a target keyword and the pages already ranking for it into an outline that covers the table-stakes subtopics every top result shares, then adds a specific unique-angle section those pages are missing.

ChatGPT (GPT-5.1)5 fillable variables

The prompt

Ready to copy — highlighted parts are example details you can swap.

You are building the outline for one specific blog post aimed at ranking for one specific keyword — not a generic outline template, but a structure built by diffing against the pages that already rank for this term.

TARGET KEYWORD
async standup tool for remote teams

CURRENTLY RANKING (TOP 3-5 RESULTS)
Three SaaS comparison posts, all structured as '7 best tools' listicles covering Geekbot, Range, and Standuply, each with a features table and pricing but no discussion of async standup failure modes.

OUR UNIQUE ANGLE
We ran a 6-month internal study across 40 remote teams tracking why async standups get abandoned after the first month.

READER'S INTENT STAGE
Actively comparing 3-4 named tools before a trial signup, not just learning what async standups are.

TARGET LENGTH
1,800-2,200 words

STEPS
1. From the ranking pages I've described, infer the subtopics they cover in common — treat that shared coverage as table stakes the outline must include, since ranking below a competitor rarely happens because our unique angle is better elsewhere while we're silently missing a section every other result has.
2. Identify what none of the ranking pages cover well, using the unique angle I gave you as the seed, and place that as a distinct, named section rather than folding it invisibly into an existing H2 where a skimming reader (or a summarizing AI Overview) won't register it as new information.
3. Order sections by the reader's actual intent stage, not by what's easiest to write first — a reader in an early research stage needs framing and comparison before a deep implementation section; a reader already comparing named options needs the comparison near the top, not buried under generic background.
4. For each H2, write one line stating the single question that section must answer and the minimum subpoints (as H3s) needed to answer it completely enough that a reader wouldn't need to open a second tab.
5. Flag any H2 where satisfying it well would push the piece meaningfully past the target length, and propose either cutting it to a summary with a link to a dedicated piece, or flag that the target length itself may be unrealistic for this keyword's competitive coverage bar.

WHAT NOT TO DO
Do not propose an outline that reads as interchangeable with any other post on this general subject — every H2 should be something you could only have written after reading what I told you about the ranking pages and the unique angle. Do not pad the outline with a generic "conclusion" or "final thoughts" section unless it does real work (a specific next step, not a restated summary).

OUTPUT FORMAT
1. A one-sentence statement of the table-stakes coverage gap this outline closes.
2. The full outline: H1, then H2s with their one-line purpose and H3 subpoints.
3. A flagged note on length risk if any section threatens the target word count.

Customize

Optional — swap in your own details for the highlighted parts above.

Why this works

Most outline prompts ask the model to structure a topic in the abstract, which produces a generically competent skeleton because the model has no signal about what would actually beat the pages currently occupying the result — this prompt instead forces a two-pass structure: first infer the shared coverage baseline from the described competitors, then treat the unique angle as a deliberate addition against that baseline rather than the whole premise of the post. That ordering matters because GPT-5.1 left to its own devices will often lead with whatever angle feels most interesting rather than what's structurally required to be competitive, and a post that's creative but missing a subtopic every ranking page covers will underperform regardless of how good the unique section is. Naming the reader's intent stage explicitly changes section ordering rather than just tone — a comparison-stage reader abandons a post that opens with generic definitional framing, so forcing the model to sequence by intent stage rather than by natural writing order (background, then detail, then comparison) closes a common structural failure mode. The length-risk flag exists because outline generation and length estimation are two different judgments the model tends to conflate — asking for the flag as a separate, explicit step prevents an outline that looks complete on paper from silently implying a 4,000-word draft when the target was 2,000, which only surfaces once someone starts writing against it.

What you get back

Coverage gap closed: none of the ranking listicles address *why* async standups get abandoned, only which tool to pick. H1: Async Standup Tools for Remote Teams (And Why Most Teams Stop Using Them). H2: What actually breaks after week 3 — not a tool problem, a habit problem (H3: the three abandonment patterns from our 40-team study). H2: How to evaluate a tool against those failure modes, not just a features table...

Verified against

ChatGPT GPT-5.1 · 2026-08-08

Changelog

  • 2026-08-08 Initial publish, verified against ChatGPT GPT-5.1.

Need this built into your business?

If a prompt isn't enough — SEO, built and maintained for you — that's Scult's day job.

EXPLORE SEO
All Blog Writing prompts

Check your AI visibility

One URL in, a 0–100 score and the exact fixes out.

RUN THE CHECK

Browse all the tools

15 tools across six categories
13 of them never send your data anywhere

Free · No signup · No trial clock

SEE THE DIRECTORY