AI SEO In Detail

AI SEO In Detail › Module B · Search Traffic Analysis › Episode 10

Episode 10 · 6:03 · free

Find Pages Stuck on Page 2 with Search Console

Filter Search Console queries at average position 8–20, pick the page worth working on, and learn what to change on it. Real data from one site, step by step.

Playing the video loads it from YouTube (privacy-enhanced mode).

Free script + worksheet

Striking Distance Finder + Worksheet

striking_distance.py pulls every visible query and query + page row from the Search Console API, keeps the queries at average position 8–20 above an impressions floor, maps them to the page Google shows and ranks the pages to work on first. The worksheet tracks what you changed and when.

  • Filter: position 8–20, impressions floor — Run the script on the last four weeks. Start with a floor of 50 impressions on a small site and raise it on a big one. Average position is weighted, so these are candidates, not verdicts.
  • Map queries to pages — You edit pages, not queries. The script ranks pages by the striking-distance impressions they collect and flags shared queries (cannibalization, episode 12) and pages that already win another query.
  • Change, log, wait — Check the results page first, then title and heading, the answer up top, missing coverage, internal links. Log the change and the date in the worksheet and re-measure after four weeks.
Download the script + worksheet (zip)

Python 3.9+ · Excel worksheet · no keys included · no sign-up

Full transcript

1,089 words · the narration as spoken in the episode.

Intro

You have pages that Google already likes. Not enough to put them on page one, but enough to show them, hundreds of times a month, on page two.

Almost nobody clicks there. So in your reports these pages look like failures. They're not. They're the cheapest traffic you'll ever get, because Google has already done half the work.

In this episode I'll show you how to find them in Search Console, using one of my own sites as the example all the way through, how to decide which page to work on, and what to actually change once you've found it.

What "striking distance" means

Striking distance is a simple filter. Take every search query where your average position is between eight and twenty. That's the bottom of page one and the whole of page two.

Why those numbers? Above eight, you're already getting most of the clicks a page can get. Below twenty, you're usually missing something big, and a small fix won't move you. Between them, the page is relevant enough to rank, and close enough that a better title, a clearer answer, or a few internal links can push it up.

The second filter is impressions. A query at position twelve with three impressions a month isn't an opportunity, it's noise. Set a floor that fits your site. On a small site, maybe fifty impressions over four weeks. On a big site, much higher.

Remember from episode eight: average position is weighted by impressions. A query at "position twelve" can be number five on some days and number thirty on others. So treat the filter as a way to find candidates, not as the truth about each one.

Real data: one site's striking distance

Before we go on, a note about the data you're about to see. It's real. It comes from one of my own sites, straight from the Search Console API. But I've anonymized it: the domain, the page URLs, and the search terms are all replaced. That way we stay focused on the principles, and not on the keywords themselves. The numbers are untouched.

Each dot is one search query. Left to right is average position. Up is impressions, on a log scale. The amber band is what I call striking distance: queries where the site ranks somewhere between eighth and twentieth. Right at the edge of page one.

In this one site, over four weeks, forty-three queries sit in that band. Together they got two and a half thousand impressions, and only a hundred and eight clicks. Those are people who saw the site and didn't click, because it was too far down.

These labels are just placeholders, so I can't tell you what the biggest one is. What matters is the shape: a few queries with real demand, parked in the zone where a small improvement can move them onto page one. That's where we'll look first.

From query to page

A list of queries isn't a to-do list yet. You can't edit a query. You edit a page. So the next step is to pull the same data with two dimensions, query and page together, and see which page Google shows for each query.

Same site, same four weeks. In the top two hundred query-and-page rows, twenty-five sit in the striking-distance band. And they point to just twelve pages. Usually you'll see one of three patterns.

One page gets almost all the impressions for the query. Good. That's your target page.

Several striking-distance queries point to the same page. Even better. On my site, one page collects seven of them. One round of changes there can lift all seven at once. So I sort pages by the total striking-distance impressions they collect, and start at the top.

Two or more of your pages share the same query, and neither ranks well. That can be a sign that they compete with each other. Don't fix that here. That's keyword cannibalization, and it gets its own episode, number twelve.

What to change on the page

Before you touch the page, search the query yourself, or pull the results page through the API, and look at what ranks above you. We'll go deep on that in the SERP module, but even a quick look answers the key question: are the pages above you the same kind of page as yours? If they're all interactive tools and yours is a plain table, no title change will fix that.

If the type matches, work through four things, in this order.

First, the title and the main heading. Does the wording the searcher uses appear in them? Not stuffed, just present. Pages often rank on page two for a phrasing they never actually say.

Second, the answer. Does the page answer the query in the first screen, or is the answer buried halfway down? Move it up.

Third, coverage. Look at what the pages above you cover that you don't. A missing section, a missing table, outdated numbers. Add what's genuinely missing, not filler.

Fourth, internal links. Link to the page from your stronger, related pages, with anchor text that describes what it's about. That's often the quickest change, and it's the one people forget.

Then write it down: the page, the queries, what you changed, and the date. Without that, you can't measure anything later.

The traps

Three traps to watch for.

One. Changing a page that's already doing well for something else. On my site, the page with seven striking-distance queries also ranks around third for another query, the one that brings more clicks than any other in the export. Before you rewrite a title, look at every query that page gets clicks from. Don't trade a winner for a maybe.

Two. Row limits. If you pull only the top few hundred queries, you'll miss most of the long tail, and that's often where striking-distance queries live. In this example I kept the top two hundred to make the chart readable. For the real work, page through the full result, as we did in episode four.

Three. Measuring too early. After a change, Google has to crawl the page again and re-evaluate it. Give it at least four weeks before you judge. And compare the same queries, over periods of equal length. How to do that properly is the topic of episode forty-six.

Wrap-up

So: filter positions eight to twenty, set an impressions floor, map queries to pages, and fix the page, not the keyword. The striking-distance script and worksheet are linked in the description. Next episode: pages that already rank well but don't get the clicks they should, and how to decide which titles to rewrite.