# How to rank in ChatGPT: what decides which companies it names? Source: https://outname.now/guides/how-to-rank-in-chatgpt Updated: 2026-10-07 By: Outnamed (https://outname.now) You cannot rank in ChatGPT the way you rank on Google, because it has no fixed list and no published ranking factors. Its answers change from one ask to the next, so the goal is to be named more often across many asks. You can measure that frequency and improve the pages and facts the assistant can draw on, but nobody can guarantee a mention. ## Key points - ChatGPT has no fixed ranking, so the useful number is how often you are named across many answers to the same question. - OpenAI has not published the factors behind its recommendations, so anyone quoting exact weights is guessing. - One screenshot proves nothing, because the same question often returns a different list the next time. - What you control is the quality, accuracy and reach of the facts about your company on your own site and on public review and community sites. - No tactic guarantees a mention, and nobody can say how long a change takes to show up. ## Is there a ranking in ChatGPT? No. When a buyer asks ChatGPT for a recommendation, it writes a new answer each time. That answer may name several companies in some order, but there is no fixed list behind it that you can move up. Ask the same question twice and you will often get two different lists. The same is true of Claude, Gemini and Perplexity. So a screenshot of you in first place tells you very little, and a screenshot of you missing tells you just as little. A better way to think about rank is frequency. Out of many answers to the same buying question, how often are you named, and where in the answer do you appear when you are? ## What is publicly known about what decides which companies ChatGPT names? Less than most guides suggest. OpenAI has not published a list of factors that decide which companies appear in a recommendation. Any article that gives exact weights for those factors is guessing. In general terms, an assistant draws on two things: what it learned during training, and, when web search is used, pages it retrieves at the time of the question. Whether it searches depends on the question, the product settings and the assistant, and that behavior changes over time. When it does search, it reads what is on the web about the topic. That includes your own site, review sites such as G2, Capterra, Trustpilot, Yelp and Google reviews, and community threads such as Reddit. Which of those sources an assistant leans on differs per assistant and per question, and it is not public. ## How do you measure whether ChatGPT names your company? Measure it as a count, not as a single answer. The method below works by hand, and it is slow but honest. Here is an example with made-up numbers. For the question "best project tool for small agencies", Your company is named in 6 of 30 answers and Competitor B in 21 of 30. That gap is the thing to work on, and next month's count shows whether it moved. 1. Write 10 to 20 buying questions the way your customers type them, such as "best accounting software for freelancers" or "which dentist near me takes new patients". 2. Ask each question 30 times, each in a fresh chat, with web search on if the assistant offers it. 3. Write down every company named in each answer and where in the answer it appears. 4. For each question, work out the share of answers that name you and the share that name each competitor. 5. Note the sources the assistant shows, when it shows any, and which ones come up repeatedly. 6. Repeat the same questions every month so you can compare like with like. ## What can you change to be named more often? No step below guarantees a mention. They are the things within your reach, and each one makes the facts about you easier to find, read and trust. Outnamed does the measuring above at scale, across ChatGPT, Claude, Gemini and Perplexity, through their official programming interfaces with web search on. It also writes answer pages for your own site, one per buying question, which you approve before they go live. You can do the same work yourself. - Write one page per buying question on your own site, and put the direct answer in the first two sentences. - Use your own real facts: what you do, who it is for, where you work, what makes you different, and numbers you can stand behind. - Keep those pages current. Update them on a schedule, and fix facts that have gone out of date. - Read what customers say about you in public reviews. Fix the complaints that repeat, and ask happy customers to leave an honest review. - Take part honestly in the community threads where your category is discussed, and never pretend to be a customer. - Keep your company details consistent across your site and every profile page. - Check that your site does not block the crawlers of the assistants you want to be read by. ## Which tactics should you avoid? Do not buy or write fake reviews. Review sites work to remove them, and they distort the very evidence the assistants may read. Do not trust anyone who guarantees a mention or a position in ChatGPT. The assistant decides, not the vendor, and its behavior changes without notice. Do not judge your progress from one answer. Judge it from a count over many answers, taken the same way each month. Do not copy a generic list of tricks. Without public ranking factors, the safest work is accurate, specific, well-sourced information about your company. ## Is it worth the effort? Buyers are already asking. In the G2 Buyer Behavior Report from June 2026, covering 1,000+ software buyers, 8 in 10 software buyers used an AI chatbot to find recommendations in the last two years. In the same G2 report, 37% say chatbots shape their shortlist most, level with review sites at 38%. Volume is a separate question. Ahrefs reported that, in its own traffic over 30 days, visitors from AI search were 0.5% of traffic and 12.1% of sign-ups. That is one company's data and not a benchmark for yours, but it shows that a small share of visits can matter more than its size suggests. How long changes take to show up is not known. It depends on the question, the assistant and what it retrieves. Measure monthly and judge the trend. ## Questions people also ask ### Can you pay to rank in ChatGPT? Nobody can sell you a guaranteed place in an organic answer, and you should be wary of anyone who tries. OpenAI may offer advertising products, and those can change, so check its current terms. Paid placement is a different thing from being named in an ordinary answer. ### How long does it take to show up in ChatGPT? Nobody knows. It depends on the question, the assistant, whether it uses web search for that question, and what it finds. Track the same questions every month and look at the trend. ### Does ChatGPT give everyone the same answer? Not necessarily. Answers vary between asks, and wording, settings and conversation history can change them. That is why you ask each question many times and count, instead of relying on one answer. ### Do I need a different approach for Claude, Gemini and Perplexity? The foundation is the same: accurate, specific pages and good public reviews. How each assistant picks and weighs sources differs and is not public, so measure each one separately. ### How many buying questions should I track? Start with 10 to 20 that real customers would type. Pick the ones closest to a purchase, such as comparisons and "best for" questions. Add more once you have a monthly routine. ### Do Reddit and review sites matter? They can. Assistants that search the web may read them, along with your own site. How much weight any assistant gives them is not public and may change, so treat them as evidence you can improve, not as a lever with a known effect. Sources for the figures on this page: G2 Buyer Behavior Report, June 2026 (1,000+ software buyers); Ahrefs, own traffic over 30 days. These are their figures, not a forecast for your company.