01What is LLM ranking strategy?

LLM ranking strategy is how you get your brand to appear in the answers of AI chatbots like ChatGPT, Google Gemini, and Claude. It is the new SEO. Instead of just ranking web pages, you aim to be the source that the LLM uses to generate its answer.

In this episode, Jabez Reuben explains his method. He and his team scrape data directly from LLMs. They see which brands show up, what categories appear, and who wins.

Then they plan content to get their clients into those answers.

The goal is to be a known entity that LLMs trust and cite. It is not about tricking the system. It is about giving the AI real reasons to mention you.

02Key tactics for LLM rankings

Jabez shared several tactics. Here are the main ones:

  • Publish listicles that are upgraded. Not the old-style stuffed list. LLMs now filter out biased listicles. You need better ones.
  • Create comparison articles with brands that top the LLM results. If a brand keeps appearing, write a comparison with them.
  • Write user guides and review articles. These build trust and consensus.
  • Use satellite sites or EMDs. These are small websites in the same niche but targeting different sub-niches. For example, one site for Hermes agent, another for Claude code.
  • Do off-page topical authority. Publish content on third-party platforms like Reddit and LinkedIn. Build your brand's authority everywhere.

These tactics work together. Listicles get quick wins, while guides and reviews build long-term trust.

03How to start with LLM ranking

To start, you need to see what LLMs are currently answering. Here is the setup:

  1. Take your target keywords. The keywords you want to rank for in LLM answers.
  2. Build an agent. Use a simple cloud agent. Give it Chrome access. Tell it to run each query in ChatGPT, Gemini, Claude, and other models.
  3. Let the agent scrape data. It should capture fan-out queries, reasoning, verbatim text, sources, and snippets from the backend.
  4. Extract the winners. Find which brands show up and in which categories. See who is the top answer.
  5. Plan your content. Decide which listicles, comparisons, or guides to write based on that data.
  6. Publish and repeat. Do this monthly to keep up with changes.

Jabez uses this exact method with his clients. It gives a clear picture of what to create.

04How to use listicles for LLM rankings

Listicles still work, but you must update them. Jabez explains the new way.

First, do not put your brand as the number one winner. Put it in positions two, three, or four. LLMs flag articles that feel sponsored or biased.

By placing your brand lower, you avoid that flag.

Second, rotate which sub-niche you win. For example, if you are an SEO agency in London, you might win 'best for SaaS' in one article, 'best for local' in another, and 'best for enterprise' in a third. Never give those votes to your competitors in other articles.

Third, link to your own review articles. Do not link directly to your homepage. Instead, daisy chain.

Publish a review article first. Then, in listicles and comparisons, link to that review. This builds more power to your articles.

05Results and comparisons in the episode

Jabez shared several findings. Here are the key results:

  • ChatGPT models differ. In free ChatGPT, 5.5 instant has less reasoning and uses fresh listicles faster. In 5.6 sole, the thinking model prefers bigger publications.
  • Claude flags biased content. Since June, Claude has been adding warnings to listicles that seem sponsored or affiliate. The way around it is to avoid putting your brand at number one.
  • LLMs filter biased listicles. Early in the year, simple listicles worked. Now LLMs are smarter and filter them out.
  • Tracking is necessary. Jabez uses Radar Kit for location-based tracking. But he also tests manually with VPNs, changing locations to see different answers.

These results show that LLM rankings change constantly. You must keep testing.

06Tips, limits and mistakes to avoid

Here are important tips and warnings from the episode:

  • Do not fake brands. Some people create fake brands to avoid giving votes to real competitors. This backfires. LLMs and Google can track fake entities. Use sister brands if you must, but establish them first.
  • Do not ignore off-page topical authority. You need to build trust beyond your own site. Publish on Reddit, LinkedIn, and other platforms. This helps you get into training data.
  • Do not rely only on listicles. LLMs may flag them. Mix in guides, reviews, and comparisons.
  • Do not expect instant results. The process takes months. You must keep publishing and tracking.
  • Do not ignore manual testing. Tools help, but varied locations and user history change answers. Test manually to see real results.

Following these tips helps you avoid common failures.

07Who this LLM ranking strategy is for

This strategy is for SEO agencies, freelancers, and brands that want to be visible in AI answers. It is especially useful if you work in competitive niches.

If you are an agency, you can offer LLM ranking as a service. Jabez does this. It is still new and not saturated, so it is a chance to lead.

If you are a brand, you can implement these tactics yourself. You will need time and some technical skill. But the payoff is being cited by every major LLM.

In the episode, Jabez also talks about building a 'company brain' using Obsidian. This stores all your research and learnings. It helps your agents get smarter over time.

That is valuable for anyone doing serious LLM optimization.