GEO Score: what it is, and why you are flying blind in Alice without it
GEO Score is a number that shows how ready your site is to land in AI answers. Not a Google rank. Not a design grade. A summary of technical access, content, trust, and whether your listings agree with each other. Without it you argue with a competitor in the dark.
Below: what the score is made of, how to read it, what to do when it is low, and why "we are fine" does not mean Alice will name you.
What GEO Score is made of
In Leadsy the express audit looks at 56 parameters: robots.txt, llms.txt, schema, FAQ, NAP, Yandex and GPT bots, response time, the mobile version, outside signals. GEO Score is the top indicator. A low score almost always means serious gaps, not "small stuff for the programmer".
A rough scale for an owner: below 40, you are out of the game on most platforms. 40 to 65, you have a chance, but a leader at 75 or above takes the recommendation. Above 65 is good, and without upkeep the market will catch up.
In AIO the absolute score matters less than the gap. You at 62, the leader at 81: the recommendations will name them, even if you feel "fine enough". Same point as in the Alice article.
How GEO differs from PageSpeed
PageSpeed 95 with a closed robots.txt: the site is fast, and for AI assistants it is dead. Slow, but open, with an FAQ and NAP: GEO is higher, and leads from Alice are possible.
Google ranks do not replace a mention in a Yandex Neuro answer. Two different channels. An SEO agency report will not show whether ChatGPT names you.
GEO Score is about readiness for AI answers: bots, content, trust, listings. It is not about how pretty the storefront is.
Why you look once a month
Competitors edit their sites. Models update. You change prices. One snapshot decides nothing. The rhythm: audit, fixes by priority, check again in 3 to 4 weeks.
Save a screenshot of the score today. Compare it a month after you fix robots, the FAQ, and the listings. A rise of 5 to 10 points is a normal result. If it did not rise, look at which red flags are still there.
What not to confuse with GEO Score
It is not PageSpeed and not a Google rank. A fast site with a closed robots.txt scores low. A slow site that is open and has an FAQ scores higher.
"One number replaces a test in Alice" is wrong. It adds to the test. Always run three voice queries after the numbers. A number with no check in the assistant is guessing again.
Five signs you are flying blind
- You have never measured GEO Score. Decisions come from a feeling.
- You looked at the score a year ago. The site and the listings have changed since.
- GEO is green. Alice names other brands three times.
- You fix the design and never touch robots.txt. The score drops quietly.
- You do not compare yourself with the niche leader. You only say "we are fine".
Without the number and the voice test, you are arguing with a competitor who has a map of what to fix.
How to read the breakdown in the report
The overall GEO Score is the top number. Under it are blocks: technical access, content, trust, listings, outside signals. Red in technical access hits harder than yellow in "about us". Do not average it by eye. Look at what is critical.
Weak ChatGPT with a strong Yandex: a separate queue. GPTBot, llms.txt. English-language mentions are not what a local business needs. Open access and a price list in HTML are.
Strong Yandex on paper, weak Alice in tests: listings and FAQ, not the overall score. A NAP mismatch between the site and Yandex Business is the classic "GEO 68, and the answer names a competitor".
The gap versus the niche leader is not "they are better than us". It is a list of items. They have an FAQ on service pages, you do not. Their Yandex access is open, you have Disallow on /price. Work the list.
After a redesign the score can fall: schema is duplicated in the footer, the FAQ is hidden in the mobile layout, robots.txt was left alone. A pretty site is not a high GEO Score.
A rise of 5 to 10 points in a month, after you close a critical item, is normal. Zero when "we fixed everything" means something is still red, or the competitor tightened up too.
An online store: look at Product, stock, and delivery in the report. A service business: FAQ, Service, NAP. One score, different priorities in the action plan.
Do not chase 100. Chase a smaller gap with the leader, and your own name in three Alice tests. A number with no voice is guessing again.
The express audit is free. The full audit with the gap is for when you need a month's queue, not a hundred "small things" with no priority. Pricing: the page.
Typical case: the owner looks at GEO once a year, when they change contractors. Over that year the competitor tightened the FAQ, and you changed prices without updating the maps. The score fell quietly. The rhythm is once a month.
A red llms.txt with green content is low priority. A red robots.txt with any content is first. The Leadsy report sets the queue.
Comparison with the niche leader in the gap: they are 78, you are 61. That is not "bad". That is "they are in the answer". Close the strong items from their list, not an abstract "SEO growth".
After a domain change or a move to https, take a new snapshot. Redirects, www duplicates, an old robots.txt on a test subdomain: the score can dip while "the migration went fine".
A short owner checklist: the score bookmarked, three Alice queries on the calendar, red items from the report in the task list, a repeat in 30 days.
GEO does not replace the FAQ or robots.txt. It shows which of those is on fire.
What to do with a low score
Do not fix everything at once. Take the critical items from the report: bots, NAP, FAQ, price list. Then schema and llms.txt. Then outside mentions. Otherwise you drown in cosmetics while Disallow: / is still sitting there.
A red critical item goes into work now. Yellow goes into the plan for the week. Leave green alone unless a competitor is ahead of you in the test queries.
Common mistakes
"Our GEO is high, that is enough forever." It is not. "GEO is SEO." It is not. It is the layer for AI answers. "We will buy ads, so GEO is unnecessary." Yandex Direct does not replace a recommendation in the answer.
Pricing and the full audit
The express check with GEO Score is free. The full audit with gap analysis and an action plan is paid, see pricing. Done-for-you implementation is for when you have no team.
The gap versus a competitor shows where the leader is stronger, item by item, not by feeling. On the losses: clients leaving for competitors.
Where to start today
GEO Score is not a separate service. It is a way to see what to fix first. Measure, close the critical items, measure again in a month.
Without the number you decide by feel. A competitor with a report decides by a map. You can guess who reaches the Alice client first.
A low score next to a strong competitor is a reason for a review, not for panic. Snapshot in the form below first, then priorities.
A green GEO Score while other brands show up in three Alice tests means look at the gap with the leader. The number and the voice test complete each other.
After a large rebuild, a new snapshot is required. A redesign often breaks schema in the footer, cuts the FAQ in the mobile layout, or duplicates Organization. The score can fall on a "pretty" storefront.
Do not chase a perfect five. Chase a smaller gap with the leader. You at 58, them at 72: they are in the answer until you close the items where they are strong.
A short checklist: a screenshot of the score today, three queries in Alice, red flags first, a repeat in a month. Four steps.
The free express audit is step one. The full audit with the gap is for when you need an action plan and a queue of edits. Pricing: the site.
An online store loses in AI answers a different way: a marketplace is in the answer, and you have no Product schema. A service business loses because the FAQ is empty. One mechanism, different holes.
Do not mix up "few leads" and "we are not recommended". The first is treated with ads. The second is the AIO layer. You can spend the budget and feed a competitor in AI answers at the same time.
A table for the quarter: one Alice query a month, and whether your name is in the answer. Five rows are clearer than "the season seems down".
Leave robots.txt alone until you have GEO Score. Do not write "AI articles" before the price list is in HTML. The number first, then the fixes from the report.
"Our GEO is fine" and "Alice names us" are different things. Check the score against a voice test every month.
A red robots.txt in the report hits GEO harder than a missing blog. A yellow FAQ on the homepage while the service pages are empty is the second most common hole.
Leadsy shows a breakdown by platform. Weak in ChatGPT, strong in Yandex Neuro: different queues of fixes.
A free audit does not force you to buy the full review. It does stop the guessing. In AIO, guessing costs more than a plan.
Questions: FAQ. On Alice: recommendations.
A yellow FAQ with a red robots.txt: robots first. Otherwise the bot never finishes the new FAQ. Take the order from the report, not from the urge to "write the text fast".
The gap with a competitor in the full audit: do not copy their site. Close the items where they are ahead. Three red flags in a month beat a weekend redesign.
A green GEO Score with other names in Alice: look at the gap. The number and the voice test are a pair.