Claude Rank Tracking: My 60-Day Test and How to Track Your Brand in Claude
Updated 13 October 2026. I first published this 60-day test in April 2026. Since then, "Claude rank tracking" has come to mean two different jobs, so this version answers both. Tool prices below were checked on each vendor's own pricing page on 24 September 2026.
Can Claude track rankings? It depends which rankings you mean
If you mean your Google positions, no. Claude can't see Google's live results, but it is very good at analyzing the position data Search Console already gives you for free. If you mean whether Claude recommends your brand, that is trackable, but only by running the same prompts repeatedly and logging mentions and citations.
| What you want to track | Can Claude do it alone? | What works |
|---|---|---|
| Your Google positions for target keywords | No live positions. It analyzes the data you give it. | A Search Console export or connector plus Claude (free). A paid rank tracker if you need daily positions and alerts. |
| Whether Claude mentions or cites your brand | No. One chat is one sample. | A prompt-tracking tool, a free one-off checker, or your own API script (1.5 to 5.5 cents per Claude answer in my test). |
My 60-day test: results at a glance
I ran Claude as the rank-tracking layer for 12 sites for 60 days: three of my own (this site, seosmohub.com and topinlists.com) and nine client sites, five of them under NDA. I split rank tracking into the eight jobs I actually do each month and scored each one against Semrush, my paid benchmark at the time, and against Search Console on its own as the free benchmark.
| Task | Claude + GSC | Semrush | GSC alone | Winner |
|---|---|---|---|---|
| Live position monitoring | 0/10 | 10/10 | 6/10 | Semrush |
| Position change alerts | 0/10 | 10/10 | 3/10 | Semrush |
| Competitor ranking surveillance | 0/10 | 10/10 | 0/10 | Semrush |
| Quick-win opportunity identification | 9/10 | 7/10 | 5/10 | Claude + GSC |
| CTR analysis and optimization | 9/10 | 7/10 | 5/10 | Claude + GSC |
| Keyword opportunity identification | 8/10 | 9/10 | 4/10 | Tie (Semrush slight edge) |
| SERP feature tracking (AI Mode, snippets) | 7/10 | 8/10 | 5/10 | Tie |
| Ranking trend reporting | 8/10 | 9/10 | 6/10 | Tie (Semrush slight edge) |
Scores are my rating of output quality, speed and how actionable the result was, on a 1 to 10 scale. Two measurements from the test carry most of the weight:
- A quick-win analysis of the same Search Console export took 35 minutes in Semrush's Position Tracking report and 8 minutes with Claude, and Claude's version came with a suggested fix for each page.
- When I pasted Search Console data from before and after a ranking drop, Claude's most likely cause was right in 8 of the 12 cases I could verify. The 4 misses were ones it had marked as low confidence.
The pattern is simple. The three tasks that need a live results page scored zero. Everything that is analysis of data you already own scored well.
Can Claude check your live Google rankings?
No. Claude has no view of Google's results page for your keyword, location and device. Its web search, available on every Claude plan since May 2025 (Anthropic's announcement), looks pages up to write an answer; it does not report where your URL ranks. Any position Claude states without data from you is a guess.
What failed in month one
Month one was mostly failures, and they were useful ones:
- Asking for current rankings. Claude said plainly that it couldn't see live Google data. That is the correct answer. A chat that gives you a position here is inventing it.
- Asking it to act as a rank tracker. It designed a sensible monitoring workflow, but the key step, fetching positions, needed a search API it didn't have.
- Asking where competitors rank. It could only offer estimates from training data that was months old. Out-of-date competitive data is worse than none when you're making decisions with it.
What worked in month two
The fix was to stop asking Claude to fetch rankings and give it the rankings Google already reports. Search Console shows average position, impressions, clicks and CTR for your queries and pages (it hides very rare queries for privacy). Google keeps 16 months of it, and new data arrives about 48 hours after collection (Google's help page).
- In Search Console, open Performance, set the date range (I use the last three months), and export with queries and pages.
- Paste or attach the CSV in Claude. A Claude Project keeps your site's context from one week to the next.
- Ask for one analysis at a time, using the prompts below.
If you'd rather not export files, you can connect Search Console to Claude directly. I walk through that in how to connect Google Search Console to Claude.
Three prompts for the free Search Console + Claude workflow
These three come straight from the test. Each expects a Search Console export pasted or attached in the chat.
Prompt 1: find the quick wins
Here is my Google Search Console performance data for the last
90 days [paste CSV].
Identify the top 10 quick-win opportunities: pages ranking between
position 8 and 20 with more than 50 impressions. For each page, give me:
1. The current average position
2. The query driving most impressions
3. The single change most likely to move it to page one (title tag,
H2 rewrite, FAQ addition, schema fix, or content expansion)
4. Your confidence: High, Medium or Low
Rank the 10 by expected impact.
Prompt 2: diagnose low CTR
Here is my GSC performance data [paste data].
Find every page whose CTR is well below what its average position
would suggest. For each one, work out the likely cause from the query
data: is the title misaligned with the ranking query, is the meta
description missing a reason to click, or is a brand name taking the
first words instead of the keyword?
Write a new title tag and meta description for the 5 biggest
underperformers.
On one client export, this flagged 12 of 47 pages whose titles opened with the brand name while the query driving impressions didn't include it. Rewriting those 12 titles lifted their average CTR by 23% within 30 days.
Prompt 3: compare two periods
Here is my GSC data for two periods:
Period 1 (the previous 90 days): [paste data]
Period 2 (the latest 90 days): [paste data]
Tell me:
1. The 10 pages that gained the most positions
2. The 10 pages that lost the most positions
3. Pages where impressions grew but clicks did not
4. Pages where both impressions and clicks grew
5. The one action with the highest priority, based on the data
Point 3 is where AI Overviews and AI Mode tend to show up: impressions hold while clicks fall. More Search Console prompts are in the GSC section of my 47 Claude SEO prompts, and reusable versions are in the Claude SEO skills library.
When do you still need a paid rank tracker?
When you need positions that Search Console can't give you: daily checks with alerts, your competitors' rankings, map-pack results, or history older than 16 months. For a single site reviewed once a month, Search Console plus Claude covers it. Past three client domains with daily monitoring, the manual exporting costs more time than a tool costs money.
| Your situation | Why the free workflow falls short |
|---|---|
| Daily positions and alerts across several domains | Search Console lags about two days, works one property at a time, and Claude only runs when you run it. |
| You need competitors' rankings | Search Console only covers sites you have verified. |
| Local map-pack visibility matters | Search Console doesn't break out map-pack results, and Claude can't see them. |
| You need history older than 16 months | That is the limit of Search Console's data. |
For scale: Semrush's cheapest SEO plan ($139 a month) tracks 500 keywords daily, and Ahrefs includes its Rank Tracker from the Lite plan ($129 a month), per their pricing pages on 24 Sep 2026. I compared six paid trackers in best SEO rank tracking software in 2026, and ran a separate zero-budget experiment in 60 days on free AI tools only.
How do you track your brand inside Claude's answers?
Pick the questions your buyers ask, run each one in Claude on a schedule with web search on, and record three things: whether you're mentioned, where you appear in the list, and whether your pages are cited. Answers change between runs, so judge rates across many runs, never one screenshot.
That is what the tools on page one for "claude rank tracking" sell. It measures something different from Google rankings, so it needs its own numbers.
What my own baseline shows
On 24 September 2026 I ran 10 buyer questions through ChatGPT, Claude, Gemini and Perplexity using DataForSEO's LLM Responses API, with web search on. Out of 40 answers, this site was cited in 3 (7.5%): once each by ChatGPT, Claude and Gemini, and never by Perplexity. Claude cited it for "best SEO consultant in India for SaaS and software companies". Three of Claude's ten answers came back with no citations at all and cost about a third as much as the others, which is what an answer written without a web search looks like.
That matters for any tracker you use. A run without a web search tells you what the model remembers, not what it would cite today.
Tools that track Claude answers, with prices
When I asked all four engines which tools track brand visibility in AI answers, Semrush came up in every answer. ChatGPT and Perplexity also named Profound, Peec AI and Otterly.AI, and Gemini and Perplexity added Ahrefs Brand Radar. Here is what those tools, plus Rankscale, charge and whether they cover Claude at all:
| Tool | Claude covered? | Cheapest way in (24 Sep 2026) | Notes |
|---|---|---|---|
| Otterly.AI | Paid add-on, $29/month on Lite | Lite, $29/month, 15 prompts | ChatGPT, Google AI Overviews, Perplexity and Copilot included. Free trial. |
| Rankscale | Yes, on every plan | Pro, $99/month, 1,200 credits | Also covers Gemini, AI Mode, Copilot and Grok. 7-day free trial. |
| Ahrefs Brand Radar | Yes, for your own prompts (each Claude update uses 8 checks) | Custom prompts included with paid Ahrefs plans; Lite ($129/month) comes with 150 checks a month | Worth it mainly if you already pay for Ahrefs. |
| Semrush AI Visibility Toolkit | Not listed | $99/month, 25 prompts | Names ChatGPT and Google AI Mode. No free trial. |
| Profound | Enterprise plan only | Price not shown; free 7-day trial | The pricing page lists only a trial and a custom-priced Enterprise plan. |
Peec AI's pricing page showed plan names but no prices when I checked, so it isn't in the table. Two free tools will check Claude once with no account: Sellm's free Claude rank tracker (you enter the question, your brand and a country, 3 checks a day) and Geoptie's Claude checker (you enter your domain and it infers the brand and query). One run is a sample, not a trend, but it's a fair first look. If ChatGPT matters more to you than Claude, I compare those tools in best ChatGPT SEO tracking tools.
Do it yourself with DataForSEO's API
If you're comfortable with a short script, you can run your own prompt panel. DataForSEO's LLM Responses API sends a prompt to Claude, ChatGPT, Gemini or Perplexity with web search on, and returns the answer plus the URLs it cited. It costs $0.0006 per live request plus whatever the model provider charges, on a prepaid balance with a $50 minimum payment (DataForSEO pricing).
- Write 10 to 25 prompts the way buyers phrase questions, not the way you'd phrase keywords.
- Send each one to
/v3/ai_optimization/claude/llm_responses/livewithweb_searchset to true. The same pattern works forchat_gpt,geminiandperplexity(endpoint docs). - For every response, log the date, prompt, model, whether your brand appears in the text, and each cited URL.
- Run it weekly and chart the citation rate per engine rather than reading single answers.
This is what each answer cost me on 24 September 2026, with web search on and a 1,200-token output cap:
| Engine (model) | Per answer that cited sources | Per answer with no citations |
|---|---|---|
| ChatGPT (gpt-5.5) | $0.087 to $0.110 (8 answers) | $0.059 (2 answers) |
| Claude (claude-sonnet-5) | $0.051 to $0.055 (7 answers) | $0.015 to $0.018 (3 answers) |
| Gemini (gemini-3.5-flash) | $0.039 to $0.085 (9 answers) | $0.011 (1 answer) |
| Perplexity (sonar-pro) | $0.012 to $0.021 (10 answers) | none |
All 40 answers cost $1.97 together. At those averages, 25 prompts a week on Claude alone comes to about $4.50 a month, and the same 25 prompts on all four engines about $21. Daily tracking changes the sums: 15 prompts a day on all four engines is about $90 a month, more than Otterly's Lite plan with its Claude add-on ($58 a month for 15 prompts, with a different engine mix). The API wins for weekly panels and single engines. For daily multi-engine tracking, a tool can cost less than the tokens.
Two caveats. API answers come from the model with web search switched on, not from the claude.ai app with a user's history and settings, so treat them as a consistent proxy rather than a copy of what any one person sees. And web_search allows a search without guaranteeing one; the docs also list a force_web_search option, which I didn't use in this run.
A free manual check in Claude
No budget at all? Run this in Claude with web search on, in three fresh chats, once a month, and keep the results in a spreadsheet:
I'm researching [category] for a [type of company] in [country].
Which [products / consultants / tools] would you recommend, and why?
Please cite the sources you used.
Record whether you're named, your place in the list, and which URLs it cites. Three chats is a small sample, but a competitor who shows up in all three is a real signal. For what to do about it, see how to get recommended by ChatGPT; the same levers apply to Claude.
What I told the three clients who asked
This test started because three clients asked me in the same week whether Claude could replace the rank tracker they were paying for. Each got a different answer:
- One site, 80 keywords, monthly reviews. We cancelled Semrush and moved to Search Console plus Claude, saving $129 a month. Their monthly rank review went from 45 minutes to 25, and after 60 days they hadn't missed the tool.
- Five client domains that needed daily alerts. They kept Semrush. Nothing free replaces daily monitoring and alerts across five domains, and doing it by hand would have cost more than the subscription.
- A competitive niche where rivals' rankings moved daily. They kept Ahrefs, on a cheaper plan, for competitor rank tracking only. Content briefs, quick-win analysis and CTR work moved to Claude.
Claude isn't a rank tracker. It's the best analysis layer I've used for ranking data you already have, and for most people that is the part of a rank tracker they actually use. For AI answers, measure rates across prompts and engines, and expect them to move slowly.
Sources and how I checked
- 60-day test: my own work across 12 sites, scored against Semrush and Search Console.
- AI answer baseline: 10 prompts across 4 engines on 24 Sep 2026 through DataForSEO's LLM Responses API, web search on, US location for ChatGPT, Claude and Perplexity. Costs are the per-task charges DataForSEO reported.
- Prices: each vendor's pricing page, checked 24 Sep 2026 (linked in the tables above). Prices change, so check before you buy.
- Search Console retention and delay: Google Analytics Help, Connect Search Console.
- Claude web search: Anthropic, launched 20 March 2025, all plans from 27 May 2025.
Leave a Comment
Your email address will not be published. Comments are moderated before appearing.