How to Do SERP Analysis for a Content Brief (Without Copying Page One)

Most briefs claim to be based on SERP analysis. Ask the strategist what they found, though, and the answer is usually a list of the top five URLs pasted into a doc, maybe with word counts next to them. The writer opens each one, notices they all have an H2 called "What is X," and writes a sixth article with an H2 called "What is X." Everyone did their job. The piece still lands on page three.

Pasting URLs takes a minute. Reading them properly takes thirty.

This guide walks through a method that takes about thirty minutes per keyword and produces findings a writer can act on. It sits between keyword research (which tells you the topic is worth covering) and the brief itself (which tells the writer what to build). If you skip it, the brief is a guess dressed up with a keyword.

Search like the reader, then write down what you see

Open a private window, set the location to your main market, and search the exact keyword. Before clicking anything, scroll the entire first page and note what Google chose to put there. You are not reading articles yet.

Record the result types in order. An AI overview at the top? A featured snippet, and if so, whether it is a paragraph, a numbered list, or a table? A People Also Ask box, a video carousel, a row of product listings, a local map, a Reddit thread sitting at position four? Each of these is Google telling you what format it thinks satisfies the query, and the format matters more to your brief than anything inside the ranking articles.

A page one made of software vendor homepages means the searcher wants to buy, and your 2,500-word guide will not break in no matter how good it is. A page one full of forum threads usually means the existing articles are thin or untrustworthy, which is an opening. Mixed results (two guides, two tools, a YouTube video, a Wikipedia entry) mean Google has not settled on an intent, and you have to choose one and say so in the brief.

Classify the intent in one sentence

After the scroll, write a single sentence describing what the searcher is trying to get done. Skip the old four-bucket labels. "Commercial investigation" tells the writer almost nothing. Try something closer to this: "A marketing manager who has been told to set up a content calendar and wants a template they can copy today, plus enough reasoning to defend the structure to their boss."

That sentence goes straight into the search intent field of the brief. If you cannot write it, you have not finished looking, and the brief is not ready.

Open the top results and read for patterns

Now click through the organic results, usually positions one through seven. Skip anything that is obviously a different intent (a dictionary definition when you are writing a how-to, say). For each page, spend three or four minutes. You are looking for a short list of things, and most of them are not word count.

The shared outline. Skim the H2s of each page and mark the sections that appear on nearly all of them. These are table stakes. A reader who does not find them will assume your article is incomplete, and Google has clearly learned to expect them too. Usually this is three to six sections.

The angle each page takes. Two articles can cover identical sections and be completely different pieces. One is written for beginners, one for a practitioner with five years in the job. One is vendor-neutral and one is quietly a sales page. Note the angle in a few words per result. When every result shares the same angle, you have found the easiest gap to exploit.

What is missing or stale. This is the part of the analysis that actually earns rankings. Look for screenshots of interfaces that have since been redesigned, statistics from 2019, tools that no longer exist, advice that contradicts what practitioners now say in forums. Look for questions from the People Also Ask box that none of the top pages answer directly. Every item on this list becomes a reason for your article to exist.

Proof. Note what kinds of evidence the ranking pages use. Original data, named experts, worked examples, templates the reader can download. If the top three all include a downloadable template, yours needs one. If none of them has original data and you have access to some, that is your differentiator, and it should be the first line of the brief.

Word count goes last, and it gets a range rather than a number. A spread of 1,400 to 2,200 across the top five tells the writer roughly how much ground to cover. Averaging those numbers and telling a writer to hit 1,812 words produces padding.

Mine the questions

Expand the People Also Ask box a few times. Each click loads more questions, and after four or five expansions you will have fifteen or twenty. Copy them all. Then check related searches at the bottom of the page, and if you have time, the autocomplete suggestions for the keyword followed by "how," "why," and "vs."

Then sort them. Some deserve their own H2 because the answer takes several paragraphs. Some get answered in a sentence or two inside an existing section. The rest belong to a different article entirely, and those go to whoever owns your content cluster plan so they can become future briefs.

Question phrasing is worth keeping verbatim. Answer engines and AI overviews lift passages that answer a question in the terms the question used, so a heading that matches "how long should a content brief be" word for word has a better chance of being cited than a clever heading that circles the same idea.

Check whether AI answers change the job

For many informational queries, an AI overview now answers the basic question before anyone scrolls. Read it closely. Whatever it covers well, your article has to go past. Whatever it gets wrong or leaves vague is where a writer with real experience can add value, and those gaps deserve a named section in the brief.

Note which pages the overview cites, too. They are telling you what Google considers extractable, and it is often not the page ranking first.

Turning findings into brief fields

Here is where most SERP analysis dies. The strategist fills a doc with observations, pastes a link to it into the brief, and the writer never opens it. Findings have to be translated into instructions, field by field.

• Search intent: the one-sentence description of what the reader is trying to get done.
• Format: guide, listicle, template page, or comparison, based on what page one rewards, plus any SERP feature you are targeting (a numbered list for a list snippet, a table for a table snippet).
• Required sections: the table-stakes H2s, labeled as required, so the writer knows these are not optional.
• Angle and differentiator: what this piece does that the current top results do not. One or two sentences, and the most important field in the brief.
• Questions to answer: the PAA questions sorted into "own heading" and "answer inline."
• Evidence requirements: the proof the piece has to include to compete, whether that means a template, a data point, or an interview quote.
• Competitor URLs: two or three, with a note on why each is listed. "Good structure, outdated tools" is useful. A bare URL invites the writer to paraphrase it.
• Length range: the spread from the top results, with a reminder that coverage matters more than hitting a number.

The content brief template already has a slot for each of these. The analysis just gives you something honest to put in them.

Mistakes that make the analysis worse than skipping it

The first is treating page one as a blueprint. If the brief says "cover everything the top five cover," the writer builds a composite of five existing articles, and Google has no reason to rank a sixth copy above the originals. Table stakes are the floor. The differentiator is the reason to publish.

The second is analyzing the wrong results. Rankings shift by location and device, and a strategist searching from a logged-in browser in one city sees a different page from a customer in another. Use a private window or a rank-tracking tool that shows the neutral results for your target market.

The third is stale analysis. A SERP read in March and briefed in August may describe a page that no longer exists. If a brief has sat in the queue for more than a month or two, spend five minutes re-scrolling before the writer starts. When the analysis is for a page you already have, the refresh brief process covers how to compare the current results against what was ranking when you first published.

And the fourth is quieter. Strategists sometimes see a page one they cannot compete with (government sites, massive publishers, a brand that owns the category) and brief the article anyway because it was on the calendar. The SERP analysis is allowed to say no. A brief that ends with "recommend targeting the long-tail variant instead" saves a writer two days of work that was never going to rank.

How long this should take

Thirty minutes for a typical keyword. An hour for a high-stakes page with a messy, mixed-intent SERP. If it regularly takes longer than that, the analysis has drifted into reading every competing article in full, and that reading belongs to the writer once research starts.

Tools can compress most of this. A good generator pulls the ranking pages, extracts their headings, collects the PAA questions, and drafts the table-stakes list in seconds, which is what ContentBrief.io does. What a tool cannot do is look at page one and decide that every result is aimed at beginners and your readers are not. That call is still yours.

Write it down in the brief. Writers can follow a sentence like that.