How to Fix Pages Stuck on Page Two in Search Console

Pages that already rank somewhere between position 8 and 25 are the cheapest traffic you’ll ever win. Google has already decided the page is relevant. You’re not asking it to trust you, you’re asking it to move you up a few slots.

I ran this routine on my own site this week. Over the 30 days to October 10, Search Console shows alejandrorioja.com at 603,486 impressions, 6,207 clicks, and a 1.03% click-through rate, with an average position of 12.3. That gap between impressions and clicks is the whole opportunity.

Table of contents

Open Table of contents

• Step 1: Pull the right 30 days
• Step 2: Keep only pages worth the time
• Step 3: Put each page in one of three buckets
• Step 4: Let Claude do the comparison
• Step 5: Make the change, then wait 28 days
• A realistic weekly pace

Step 1: Pull the right 30 days

1. Open Search Console and go to Performance > Search results.
2. Set the date range to the last 28 or 30 days. Shorter windows are noisy, longer ones hide recent changes.
3. Turn on Impressions and Position next to Clicks and CTR.
4. Click the Pages tab first, not Queries. You’re fixing pages, so start from them.

Then click any page and switch to Queries to see what it actually ranks for. Pages without a query list are usually too new or too thin to bother with.

Step 2: Keep only pages worth the time

Filter out the noise before you do any thinking:

• Impressions above 500 in the window. Below that you’re guessing from a handful of data points.
• Average position between 8 and 25. Position 8–10 is the bottom of page one. 11–25 is page two and the top of page three.
• Skip anything branded, anything that’s a tool or product page, and anything you’d be embarrassed to rank for.

On my own data that leaves a short, readable list. A few real rows from the same 30 days:

Page	Impressions	Clicks	CTR	Avg position
Reverse video search explained (EN)	3,837	125	3.26%	20.7
High-CPC AdSense keywords	13,508	22	0.16%	14.2
How to download a YouTube playlist	1,975	24	1.22%	24.3
GoDaddy email login	44,398	44	0.10%	8.8

Same filter, four very different problems. That’s why you sort before you fix.

Step 3: Put each page in one of three buckets

Bucket A: ranking fine, nobody clicks. Position 8–10 with a CTR under about 1% means you’re visible and unchosen. The GoDaddy page sits at 8.8 with 44,398 impressions and a 0.1% CTR. More content won’t fix that. A sharper title and description might, and so might accepting that an AI answer or a brand result above you is taking the click. Rewrite the title to promise the specific outcome, put the number or the method in it, and keep it under 60 characters.

Bucket B: ranking low because the page is thin or off-target. Position 15–25 with real impressions, like the YouTube playlist page at 24.3, means Google sees relevance but not enough depth. Open the top three results for the main query and write down what they cover that you don’t. Add those sections. Don’t pad. Add the answer they have and you lack.

Bucket C: leave it alone. If the page converts, earns links, or serves a query you don’t want more of, moving it up the rankings buys you nothing. Skipping is a valid decision.

Step 4: Let Claude do the comparison

The tedious part is reading competing pages and spotting the gaps. This is the prompt I paste into Claude or ChatGPT for a Bucket B page:

Here is my page: [paste the full text]. My target query is “[query]”. Search Console shows I average position [X] with [Y] impressions a month. Below are the top three pages that outrank me: [paste their text or URLs]. List every subtopic, question, or example they cover that my page doesn’t. Then tell me which three additions would matter most for someone trying to do this task today, and draft those sections in my voice: first person, direct, numbered steps where they fit. Don’t invent statistics or experience I haven’t given you.

Check two things before you publish what it gives you. First, any number it adds needs a source you’ve opened yourself. Second, any “I did X” sentence has to be something you actually did. These models will happily supply both.

Step 5: Make the change, then wait 28 days

Edit the page, update the updatedDate if your site shows one, and request indexing for the URL in Search Console’s URL Inspection tool. Write the date and the old average position in a spreadsheet.

Then leave it for 28 days. Rankings move in the week after a recrawl and keep settling for a while. Judging a page after four days is how people end up rewriting the same article three times.

Compare clicks and average position over the same 28-day window before and after. Keep the edits that moved something and note what you changed, because that’s your playbook for the next ten pages.

A realistic weekly pace

Five pages a week is plenty. One hour to pull and sort the list on Monday, then a short rewrite per page. Within a month you’ll know which bucket your site is mostly made of, and that tells you whether you have a title problem or a depth problem.

If you want the wider keyword-research setup around this, my keyword research guide and on-page SEO checklist cover the parts that happen before and after this routine.