SERP Analysis: How to Read a SERP Before You Write

You picked a keyword. The tool says 320 searches a month, difficulty 22, green light. You brief the article, you pay for it, you publish it, and four months later it sits at position 34 with nine impressions.

The keyword was not the problem. The SERP was. And you never looked at it.

SERP analysis is the five-minute check that saves a $2,000 article. It is the cheapest step in the entire content process and the one almost every agency skips, because a spreadsheet of volume numbers looks more like work than actually opening Google.

What is SERP analysis and why does it come before writing?

SERP analysis is opening the live search results page for your target keyword and answering three questions before a single word gets written: what format is ranking, who is ranking, and how much of the page is left for organic results at all.

Keyword research tells you a query exists. SERP analysis tells you whether Google wants what you are planning to publish. Those are different questions, and only one of them predicts whether the page earns anything.

Do it after your keyword research, before your brief. The order matters: a keyword that survives volume and difficulty screens can still be a dead page once you look at the actual results.

What format is Google actually rewarding for this query?

Look at the top 10 results and count formats. If seven of them are listicles, Google has decided this query wants a list — and your 3,000-word ultimate guide is competing in a category the SERP already rejected.

This is the single most common way a good keyword produces a bad page. The writer defaults to the format they always write, not the format the SERP is paying for.

Sort the top 10 into buckets and write down which one wins. The buckets that actually show up on service-business queries:

If a format is not represented in the top 10 at all, that is not a gap you found. That is usually a format Google already tested and buried.

Who ranks — and are you even invited to this SERP?

Scan the ten domains. If eight of them are directories (Yelp, Angi, Thumbtack), plus a Reddit thread and one national brand, you are not invited to this SERP — Google has decided this query gets answered by aggregators, and one more agency blog post will not change its mind.

You are looking for at least two or three results that look like you: a business your size, a site your age, a page with roughly your link profile. Those are the seats you can plausibly take. If every seat is held by a national publisher or a directory, the honest answer is to skip the topic.

Reddit and forum threads deserve their own note. Google's own visual elements gallery is an illustrated guide to what each result type looks like, what it is called, and — in Google's words — “whether you can optimize your website for each element.” Use it to name what you are looking at. Our own read on a discussion-heavy SERP is blunter: a sales-adjacent blog post rarely displaces a community thread, so pick a different query.

This is also where a fresh domain has to be honest with itself. A page on a two-month-old site is not beating a 15-year-old directory on a head term, and no amount of on-page work changes that. We would rather you write four pages on winnable SERPs than one heroic page on a SERP that was never available.

How much of the page is left after the SERP features take their cut?

This is the question nobody asks, and it is the one that decides whether the traffic exists at all. On a SERP with an AI Overview, four ad slots, a map pack, and a People Also Ask box, organic position 1 is below the fold — and the click math collapses accordingly.

The numbers are not subtle. Ahrefs compared 300,000 keywords using Google Search Console data and found that when an AI Overview is present, the #1 organic result sees a 58% lower clickthrough rate. Position 3 loses 46.4% and even position 10 loses 19.4% (Ahrefs, December 2025 data).

Behavioral data says the same thing from the user side. Pew Research Center tracked 900 US adults across 68,879 real Google searches and found that when an AI summary appeared, users clicked a search result on just 8% of visits — versus 15% when no AI summary was present. They clicked a link inside the summary itself on only 1% of visits (Pew Research Center, March 2025 data).

So before you commit, scroll the SERP and note what sits above the first organic result. Then discount the traffic estimate accordingly. A keyword with 500 searches and a stacked SERP is worth less than a keyword with 120 searches and ten clean blue links.

Why is the search volume number lying to you on local queries?

On a local query, the map pack takes the clicks and the volume number in your tool does not know that. A term like plumber near me can show 1,900 searches a month and deliver almost nothing to a blog post, because the intent is served entirely above the organic results.

The AI Overview data actually cuts the other way here, and it is worth knowing. Ahrefs analyzed 146 million SERPs and found AI Overviews appear on 20.5% of keywords overall, but on only 7.9% of local searches — versus 57.9% of question queries and 21.4% of informational ones (Ahrefs, September 2025 data).

Read that carefully, because it is the useful part. Local SERPs are mostly not being eaten by AI Overviews — they are being eaten by the map pack and Local Services Ads. Informational and question-shaped queries are where the AI Overview shows up, which is exactly where most blog content lives.

Practical consequence: if the SERP shows a map pack, your lever is Google Business Profile optimization, not another article. If the SERP shows an AI Overview and ten organic links, the article can still work — but you write it to be quotable, which is a different job. We cover what changed in how AI Overviews rewrote the playbook.

How do you decide in five minutes whether to write the page at all?

Open the SERP in an incognito window, set to the location you actually sell in, and run six checks. Total time: five minutes. If it fails two of them, drop the keyword and move on — that decision just saved you a month.

• Format check. Count the top 10 by content type. Can you credibly produce the winning format? If not, stop.
• Peer check. Are there at least two results from sites roughly your size and age? If it is all directories and national brands, stop.
• Feature check. List everything above the first organic result. AI Overview, ads, map pack, LSAs, snippet, PAA.
• Discount check. Reduce the tool's volume estimate by what the features take. On a map-pack SERP for a blog post, treat it as near zero.
• Angle check. Read the top three pages. Can you say something specific they do not — a number, a process, a threshold? If your only edge is being longer, stop.
• Money check. Does this query, if it converts, connect to something you sell? If it cannot be linked to a service page in one hop, it is a hobby post.

Two failures is a stop. One failure with a strong angle is a maybe. Zero failures is a write.

What does a winnable SERP look like for a fresh domain?

A winnable SERP for a new site has at least three of these: no AI Overview, no map pack, no more than two ads, at least two small-site results in the top 10, and top results that are visibly thin — a 500-word page with no specifics, or an outdated post with a 2022 date.

That is the profile of a long-tail, question-shaped, low-competition query. It will have less volume than the head term you wanted. That is fine. A page that ranks 4th on a 90-search keyword and books one job beats a page that ranks 40th on a 2,000-search keyword and books nothing.

The corollary, and it is uncomfortable: some of your best keywords do not deserve a blog post. They deserve a service page. Buyer terms belong on money pages — a query like emergency plumber pricing is not an article, it is a section of your SEO service page. Blog posts take the informational and problem-aware queries and funnel them there.

Anyone who promises rankings without looking at the SERP is guessing with your money. If you want us to run this check across the keywords you are already targeting — and tell you honestly which ones to abandon — get my free audit. No contract, no pitch deck, just the list.