Find the pages worth improving in Google Search Console

Search Console lists every page on your site that Google has shown in results. Sorted the default way, by clicks, the top of that list is mostly pages that already rank well and blog posts that answer questions. Neither is where an update pays most.

The pages worth updating are usually further down: pages that show up for searches from people ready to buy, but sit just too low to get many clicks. They are in the report. They just aren't sorted to the top.

This guide goes step by step through the Search Console performance report to find those pages, separate buying searches from learning ones, and put a rough monthly value on moving each page up. You need Search Console access and, for the last step, your analytics.

What the four numbers mean

The performance report shows four metrics:

• Impressions: how many times a page from your site was shown in results.
• Clicks: how many times someone clicked through to it.
• CTR (click-through rate): clicks divided by impressions.
• Average position: where your result appeared, averaged across all its impressions. Lower is better; 1 is the top.

Impressions are the important one here. They are measured by Google, not estimated by a keyword tool, so they show real demand for the searches you already appear for.

Step 1: Open the report and set the view

1. In Search Console, open Performance, then Search results.
2. Leave the date range at the last three months. That is long enough to smooth out weekly swings and short enough to reflect where the page is now.
3. At the top, click Average CTR and Average position so all four metrics show in the chart and the table below it.

If most of your customers are in one country, add a Country filter now. A page can rank very differently from one country to the next, and an average across all of them can mislead.

Step 2: List the pages with room to climb

Switch the table to the Pages tab. Sort by impressions, highest first.

Now narrow it to pages at average positions 4 to 20. You can do this with the filter icon above the table, but it is easier to Export the table to a spreadsheet and filter there, because you will add columns later.

Why 4 to 20? Pages in the top three have little left to gain and a lot to lose. Pages below 20 are a long way from page one and usually need more than an update. Between those, typical click-through rates change fast with each position, so a modest move up adds a lot of clicks.

Also set an impressions floor so that you aren't chasing a handful of clicks. There is no single right number: a few hundred a month might be meaningful on a small site, a few thousand on a large one.

Step 3: Look at each page's searches

A page's average position blends every search it appeared for, so the page total can hide what is going on. Click a page in the table: Search Console adds a filter for that page. Then switch to the Queries tab to see the searches behind it, each with its own impressions and position.

Remove brand searches (people typing your company or product name). They were already looking for you, and moving up for them adds little. Click Add filter, choose Query, then Custom (regex), set it to Doesn't match regex and enter your brand and its common misspellings, separated by a vertical bar:

acmedesk|acme desk|acmedesks

What is left are the searches the page competes for.

Step 4: Separate buying searches from learning ones

The same page can appear for someone comparing products and for someone reading about a topic. They are worth very different amounts. A simple way to sort them is by the words in the search:

Two rules help. Buying and comparing words beat question words: "how much does a standing desk cost" is someone pricing a purchase. And a bare product phrase such as "standing desk", with no other words, is usually someone shopping. The guide to high-traffic keywords that deserve no link budget explains the reasoning in more detail.

There is a quicker way to see which pages carry buying searches. Remove the page filter, add a query filter with Matches regex, for example:

price|cost|buy|cheap|sale|quote|best|top|vs|review|alternative

Then switch to the Pages tab. The impressions now count only searches that match, so the pages at the top are the ones with the most buying and comparing demand. Add a bare product phrase or two to the pattern if your market has them.

Step 5: Put a value on a click

Open your analytics (Google Analytics or similar) and find, for the same three months, visits from organic search that landed on each page or type of page, and the orders or leads they produced. Then:

value per click = conversion rate × value per conversion

If you don't track revenue, use a value you are comfortable defending, such as the average first-year value of a customer multiplied by the share of leads that become customers. Keep product and category pages separate from blog posts: they usually convert very differently.

Step 6: Estimate what each page is worth

For each page's main non-brand search:

extra revenue a month = impressions a month × (CTR at target position − CTR now) × value per click

Pick the target as the next realistic step, not #1. Roughly halving the position number is a sensible default. For the CTRs, use typical figures. These vary by results page (ads, shopping results and other features all take clicks), so treat them as a guide:

A worked example

Take an online shop selling standing desks. The figures are illustrative. After steps 1 to 4, five pages remain, each with its main non-brand search. Impressions are the three-month total divided by three.

Analytics for the same period shows that 1.2% of organic visits landing on product and category pages led to an order, and the average order was $500. A click there is worth 1.2% × $500 = $6. Only 0.08% of organic visits to blog posts led to an order: 0.08% × $500 = $0.40 a click.

Rounding each position to the nearest one in the CTR table, and roughly halving it:

Sorted by extra clicks, the height guide wins by a distance. Sorted by revenue, it is third, and the two pages to update first are the category page and the electric desks page, together worth about $5,190 a month ($3,240 + $1,950), modelled.

One correction is worth making. /desk-converters sits on page two, where impressions undercount demand (the next section explains why). If a keyword tool puts "standing desk converter" at about 8,000 searches a month, the estimate becomes 8,000 × 1.9% = 152 extra clicks, worth 152 × $6 = $912. That moves it above the height guide. Treat the keyword tool figure as an estimate, not a measurement.

The learning pages still have a job. The height guide gets 22,000 impressions a month; a link from it to /standing-desks/electric, placed in a sentence about desk height ranges, costs nothing.

Ship/Scale does this ranking for every page with impressions in Search Console, using each search's intent and monthly volume, ignoring brand searches, and taking the value of a click from your connected analytics or figures you set.

Where this goes wrong

• Page two undercounts demand. An impression only counts when your result is on a results page someone actually loads, and most searchers never go past the first page. Impressions for pages at positions 11 to 20 understate how many people search. Use a monthly search volume alongside them where you have one.
• Some searches are hidden. Search Console leaves out rare queries to protect searchers' privacy, and they disappear entirely once you add a query filter. A page's query totals can be well below its page total. Long-tail pages lose the most.
• Averages blend devices and countries. A page can be #3 on mobile and #15 on desktop. If the average looks odd, filter by device or country before you trust it.
• Word rules misread some searches. "Standing desk for small spaces" has no buying word and is still someone shopping. Check the searches carrying the most value by hand.
• Visits and clicks aren't the same count. Analytics visits and Search Console clicks are measured differently and rarely match. That is fine for a value per click, which is a ratio, but don't expect the totals to agree.
• Last-click value undervalues learning pages. Someone who reads the height guide and buys a week later via a brand search won't show up against the guide. If your analytics can show that path, weight those pages up. Don't assume it.
• The CTR table is typical, not yours. Compare each page's own CTR with the table for its position. If it is far below, the page may have a title problem rather than a ranking problem; the guide to content decay or a conversion problem covers that.

The key points

In the performance report, filter out brand searches, keep pages at positions 4 to 20 with real impressions, and use a regex filter on the query to see which pages carry buying and comparing searches. Value a click from your analytics, then estimate each page's worth as impressions × the CTR gain from a realistic move × value per click. Update the pages at the top of that list first, and link to them from the learning pages that get the traffic.