There is a specific kind of frustration that comes from opening Search Console and finding a query sitting at an average position of 7 with a handful of impressions, zero clicks, and a 0% CTR. You are ranking — technically. Google considers your page relevant enough to show, but not compelling enough to choose. That limbo between indexed and earning traffic is where most real SEO work happens, and it is nearly impossible to escape without an honest, granular view of your keyword ranking in Google.

This guide covers how to read ranking data without fooling yourself, which data sources to trust, and a repeatable workflow for turning middling positions into clicks. It also covers the layer most rank trackers still ignore: whether you appear in AI-generated answers at all.

What a keyword ranking in Google actually measures

A position is not a single, stable fact. It is an average computed across every search that matched your page during a reporting window, weighted by impressions. That means a reported position of 7 can hide a page that ranks 2 for one variant of a query and 40 for another. It can also hide a page that ranks well on desktop in one country and nowhere else.

Several factors move that number around without anything changing on your site:

  • Personalization and location. A searcher in Manchester and a searcher in Manila can see very different results for the same query.
  • Device. Mobile and desktop results frequently diverge, especially for local and transactional queries.
  • SERP features. AI Overviews, featured snippets, local packs, and shopping carousels push organic results down the page. Position 3 behind an AI Overview can receive fewer clicks than position 6 on a plain results page.
  • Query variants. Search Console groups similar queries, so a single row may represent dozens of near-identical searches.

The practical takeaway: treat your keyword ranking in Google as a directional signal, not a precise measurement. What matters is the trend over weeks and the distribution of positions across a cluster of related queries — not whether you moved from 7.2 to 6.9 overnight.

Use three data sources, not one

1. Google Search Console

GSC is the only source of first-party data about how Google actually shows your pages. It gives you impressions, clicks, CTR, and average position at the query, page, country, and device level. Its weaknesses matter just as much: it samples data, hides low-volume queries behind the (other) row, and reports nothing at all for queries where you received no impressions.

Because GSC only shows what already happened, it is excellent for finding opportunity and useless for spotting pages that are invisible. Pair it with something that tracks positions proactively.

2. A dedicated rank tracker

Rank trackers query Google on a fixed schedule from fixed locations, which makes them good for spotting movement and comparing against competitors. A tool that connects directly to Search Console and merges both datasets — GSC's impression-weighted averages alongside daily tracked positions — gives you the best of both: the reality of what Google served, and the stability of a controlled measurement.

3. AI visibility monitoring

Ask ChatGPT, Perplexity, or Gemini a buying-intent question in your category and you will get a short list of recommended brands. If you are not in it, you have a visibility problem that no traditional rank tracker will ever surface. Tracking these mentions — which prompts you appear in, which competitors dominate, and how often — is now as important as tracking the ten blue links.

How to read ranking data without fooling yourself

Look at impressions before position. A query with five impressions tells you almost nothing about performance. It is a hint, not a verdict. Filter for queries with meaningful impression volume before drawing conclusions; a 0% CTR on five impressions is statistically meaningless.

Compare position against a CTR benchmark. Roughly speaking, position 1 earns 25–30% CTR, position 3 earns 10–12%, position 5 earns 5–6%, and position 7–10 earns 2–3%. If a query sits at position 4 with a 0.5% CTR, the ranking is fine and the snippet is the problem — title, meta description, or the SERP features around you.

Mine the position 5–15 band. This is where SEO returns are highest. Pages already ranking there have proven relevance; small improvements — better intent match, added depth, stronger internal links — can move them into the top three and multiply clicks. Pages at position 60 need a completely different strategy.

Separate brand from non-brand. Branded queries inflate your average position and CTR. Segment them out so you can see how you rank for the terms you actually want to win.

Watch the direction, not the decimal. A keyword moving from 12 to 8 over six weeks is a real win. A keyword bouncing between 7.1 and 6.8 is noise.

A repeatable workflow for improving your keyword ranking in Google

  1. Export every query with impressions but a weak CTR. In Search Console, filter for queries with at least a few hundred impressions and a CTR well below the benchmark for their position. These are your highest-leverage targets.
  2. Cluster queries by intent. Group the export into topics — a single page should usually target one intent, not one exact-match phrase. Cluster analysis reveals which pages are over-stretched and which are under-built.
  3. Diagnose the gap. Open your ranking page and the current top three results side by side. Are you answering the question the searcher actually asked? Is your page thinner? Is the format wrong — a listicle competing against a product page? Most ranking problems are intent problems.
  4. Improve, do not just republish. Rewrite the title and meta description to earn the click, add the sub-questions the top results cover, add internal links from pages with authority, and tighten the page structure with clear headings. Update the content date only if the content genuinely changed.
  5. Re-measure at 2, 4, and 8 weeks. Rankings rarely respond instantly. Track the same query set against a baseline so you can attribute movement to the change you made rather than to seasonality or an algorithm update.

Rankings no longer live only in the ten blue links

Generative engine optimization — GEO — is the practice of making your brand appear in AI-generated answers. It works differently from classic SEO. AI systems favor content that is clearly structured, factually specific, and easy to quote: direct answers near the top, explicit definitions, comparisons in tables, and consistent entity information across your site.

You cannot measure GEO the way you measure a keyword ranking in Google, because there is no fixed result page. What you can measure is presence: how often your brand is mentioned when a set of recurring prompts is run, which competitors appear alongside you, and whether the mentions are accurate. Treat it as a second scoreboard that sits next to your rankings, not a replacement for them.

Three mistakes that stall rankings

  • Chasing position instead of clicks. A page that moves from 8 to 5 but gains no traffic has not solved anything. Always connect rank movement to sessions and conversions.
  • Optimizing pages that cannot win. If the results page is dominated by a government site, a Wikipedia entry, and three retailers, a blog post will not break in. Pick fights you can win.
  • Ignoring the SERP layout. Before you write anything, look at what Google actually shows. If AI Overviews and a local pack fill the first screen, position 3 may be worth far less than position 3 on a clean results page.

Frequently asked questions

Why does Search Console show position 7 while my rank tracker says 12?

They measure different things. Search Console averages positions across every impression you received, weighted by impression volume and including long-tail variants of the query. A rank tracker checks one exact query from one location at one moment. Neither is wrong — they are answering slightly different questions. Use GSC for opportunity and a tracker for stability.

How long does it take to move a keyword ranking in Google?

For an existing page with established relevance, meaningful movement often appears in two to six weeks. New pages typically take three to six months to rank competitively in a non-trivial niche. If nothing has moved after eight weeks and the page has been recrawled, the problem is usually intent or authority, not patience.

Does a 0% CTR mean my ranking is bad?

Not necessarily. With very low impressions, a 0% CTR just means nobody has clicked yet — five impressions is not enough data to judge. For high-impression queries, though, a near-zero CTR at a top-10 position usually points at the snippet, the SERP layout, or a mismatch between the query and the page.

Does Search Console measure AI Overviews and AI Mode?

Impressions from AI surfaces are folded into Search Console totals, but you cannot isolate them as a separate dimension. That is why AI visibility tracking has become its own discipline — you need to check AI answers directly rather than infer them from click data.

The bottom line

Your keyword ranking in Google is a starting point, not a scoreboard. The useful questions are always one level deeper: which queries have real demand, which pages are close enough to push, and what specifically is stopping the click. Answer those, re-measure on a schedule, and add an AI visibility check to the routine — because a growing share of your future customers will never see a ranked list at all.

Start with the queries you already have. Five impressions at position 7 is not a failure; it is a page Google is willing to consider. Give it a reason to be chosen.