← Back to Blog
AI SEO

How to Run a Search Visibility Analysis (Free, in About 90 Minutes)

A step-by-step guide to measuring your search visibility on Google and in AI answers, using Search Console and a fixed set of test prompts.

By Paal Soberg
How to Run a Search Visibility Analysis (Free, in About 90 Minutes)

What you'll achieve

By the end of this guide you will know how often your website appears in Google and in AI answers such as ChatGPT, which pages are carrying the load, and which are slipping. It is written for owners and solo marketers at UK small businesses with no in-house SEO. Expect to spend about 90 minutes the first time, and 30 minutes on each repeat.

Search visibility is how often, and how prominently, your site appears when people search for what you offer. That now covers two places: Google's results, and the answers generated by AI tools. A search visibility analysis is the routine of checking both, so you can see what is working and what needs attention.

What you need

  • Access to your site's Google Search Console (free; ask whoever built your site if you are unsure)
  • A spreadsheet
  • Free accounts for ChatGPT and, ideally, Perplexity or Gemini
  • A private browsing window
  • A list of what you sell and the towns or regions you serve

Step 1: Define what you want to be found for

Write down 15 to 20 searches a real customer might type or ask. Mix four kinds: brand ("your company name"), service ("emergency boiler repair"), local ("accountant in Leeds") and question ("how much does a loft conversion cost in the UK"). Use customer wording, not your own jargon. Keep the list fixed, because changing it each time makes comparisons meaningless.

Step 2: Open the Search Console performance report

In Search Console, choose Performance, then Search results. Set the date range to the last three months and tick the options for total clicks, impressions, click-through rate and average position. Impressions show how often your pages appeared; clicks show how often someone chose them. Common mistake: judging only by clicks. A page with rising impressions and no clicks is visible but unconvincing.

Step 3: Record your baseline in a spreadsheet

Copy the headline figures for clicks, impressions, click-through rate and average position into a dated row. Then add a second sheet for your fixed query list. This baseline is the whole point. A single reading tells you little, but the same reading taken every month shows direction. Date every entry in DD/MM/YYYY so future you can follow it.

Step 4: Find the queries and pages that matter

Switch to the Queries tab and sort by impressions. Note any query where you appear a lot but rarely get clicked. Then open the Pages tab and compare the last three months with the previous three. Pages with falling clicks are the ones decaying. Search Console shows average position as a rounded, averaged figure that varies by device and location, so treat it as a guide, not a fixed rank.

Tip: Brand-new sites may show very little data. Check back in a few weeks rather than reading too much into sparse numbers.

Step 5: Spot the gaps between your list and your data

Compare your Step 1 list with the queries Search Console reports. Anything you care about that never appears is a gap: either you have no page for it, or the page is not answering the question well. Mark each gap as "no page" or "weak page". This tells you whether you need a new article or a rewrite.

Step 6: Test your list in AI answers

Open a private browsing window and ask ChatGPT, then Perplexity or Gemini, your service, local and question prompts exactly as written. Logged-in accounts can skew answers by your history. For Google's AI Overviews, search the same phrases and note whether an overview appears and which sources it names. This step matters because AI answers pull from live web pages, so the same clear, specific content tends to help in both places.

Step 7: Score each AI result with a simple scale

For every prompt and tool, log one of four outcomes: not mentioned, mentioned by name, cited with a link, or recommended. Keep the scale this simple so you actually use it. Do not treat one run as a verdict. AI answers vary from one attempt to the next, so look for patterns across the whole list, not a single win or miss.

Step 8: Check that AI crawlers can reach your site

Open yourdomain.co.uk/robots.txt in a browser. Look for rules that block crawlers such as GPTBot, ClaudeBot or PerplexityBot, or a blanket block on everything. An accidental block means those tools may never read your pages. Crawler access is a prerequisite, not a visibility measure: being allowed in does not mean you are being used in answers.

Step 9: Check your business details are consistent

Compare your business name, address and phone number on your website, Google Business Profile and main directories. Inconsistent details make it harder for search engines and AI tools to tie everything to one business. Fix any mismatches at source. It is dull work, and it pays off in clarity.

Step 10: Turn findings into a short action list

Write no more than five actions, each tied to a finding: refresh these two declining pages, write an answer for this missing question, fix this robots.txt rule. Data without action is homework, so give each action an owner and a date. If you do nothing else, schedule the next analysis in your diary for four weeks' time.

How often should you repeat it?

Monthly is enough for most small businesses. Search Console data moves slowly, and AI answers are noisy enough that checking daily mostly produces anxiety. Keep the same query list and the same scoring scale so each month is comparable with the last. Nobody can honestly promise a ranking by a particular date, so judge progress by trend over several months.

Where does this analysis fall short?

It is manual and sampled. Paid tools can run far more prompts than you can by hand, but they are also approximating a system that gives different answers each time, so treat their numbers as directional too. For a small business, a consistent manual routine is a sound starting point.

If you want a wider view of the topic, what search visibility is and why it matters for a UK business covers the basics, and the search visibility audit in under 30 minutes is a quicker version of this routine. For how SEO, GEO (generative engine optimisation, meaning being cited in AI answers) and AEO (answer engine optimisation) fit together, see the search visibility strategy guide.

You've finished

You now have a baseline from Search Console, a record of how you appear in AI answers, a list of gaps, and up to five actions. That is a working search visibility analysis, and it cost nothing but time.

Your next move is to act on the gaps. Most come down to publishing useful, specific articles and keeping existing ones fresh, and how to improve search visibility in the UK sets out a practical way to do it.

If you would rather not write those articles yourself, ClimbrIQ crawls your site, builds a content calendar from what is actually on it, writes and publishes articles, refreshes pages that slip using your Search Console data, and emails a weekly report. Every article waits 24 hours before going live, so you can edit, hold or veto it. It costs £59 a month for one site.

See your first article free

Does AI know your business?

Our free check asks a large language model the questions your buyers ask and shows you whether your brand comes up. It reads the model's trained knowledge, not a live search — no account needed either way.

Try it free