A public record of what insurance searches actually return, rebuilt every month from live captures. Not estimates, not a tool's projection: the same searches run in every US state, every Canadian province and 109 cities, with everything Google returned written down.
The US insurance industry is projected to spend over $14 billion on digital advertising in 2026, according to eMarketer. Insurance consistently ranks among the most expensive categories in paid search, and search is where the buying decision starts: 78% of insurance consumers call a business after running a search.
For twenty years the alternative to paying for those clicks was to rank for them. That worked because the results page was a list of links, and a list of links can be earned.
In 2026 that stopped being the whole picture. Google now writes its own answer above the results on most insurance questions, names a few brands inside it, and credits the sources it used. On local searches it skips the answer entirely and hands the result to three businesses on a map. Ranking first is still valuable, and it is no longer the same thing as being seen.
Almost nobody is measuring the difference. Rank trackers report position in a list that a growing share of searchers never reach. This Index measures the whole page instead, in every US state, every Canadian province and 109 cities, and publishes it free so anyone can check their own position without buying anything.
Position-weighted share of the top ten organic results, indexed so the leader in each geography sits at 100.
Whether the brand appears in the text of Google's AI answer. Counted from the answer itself.
Whether Google lists the brand's page as one of the sources beneath its answer. Different from being named.
Whether the business appears in the three-slot local pack on near-me searches, and in which slot.
These are never added together. A single blended visibility score would hide the only finding that matters, which is where the four disagree.
Four lines, four different competitive fields. Insurers hold the majority everywhere, but who holds the rest changes completely: comparison and lead-generation sites in car insurance, publishers and forums in home, regulators in business insurance, directories and local agencies in local search.
Car insurance is the most contested by intermediaries. Comparison sites, lead-generation sites and publishers hold 28.6% of page one between them, more than in either other line.
Business insurance is the most concentrated. Insurers hold 65.5%, and the second largest group is state regulators and government agencies at 9.3%, which appear on workers compensation questions rather than as competitors.
Local search is the only line where agencies and brokers hold meaningful ground at 12.5%, alongside directories at 12.3%. The unclassified share here is largely single-market agency websites, one per city, which is what the local map data shows too.
Home insurance sits closest to car insurance at 56.8%, but the runner-up is different: publishers at 11.3% and forums at 7.9%. Home is the line where people read reviews and ask other people before buying.
Insurance answers draw on more than insurers. Use the filters to include or exclude participant types and see how the picture changes.
Filtering recalculates each share against the types you have selected, so excluding comparison sites shows the split among everyone else rather than leaving a gap.
Published in full, because a measurement you cannot check is an assertion. These do not change month to month, which is what makes movement mean something. Keywords carry no place names: geography is applied at capture time, so the same strings run everywhere.
of coverage and pricing searches returned an AI answer above the results.
of near-me searches returned a map instead, and an answer only 6.3% of the time.
states and provinces where the brand ranking first is not the brand Google cites.
of the sources Google credits also rank in the top ten. No answer cited nothing that ranks.
of page one on business insurance is held by insurers, the most concentrated of the three lines.
of page one on car insurance is held by comparison, lead-gen and publisher sites.
of the 866 businesses holding an insurance map slot appear in a single city.
is the median rating of a business in an insurance map pack. Half sit at 4.9 or 5.0.
metros where a carrier, not an agency, ranks first for "insurance agency near me".
sources The Hartford supplies to business insurance answers, against times it is named in one.
of Canadian map slots are held by the top two operators, against 14.0% in the US. Same concentration, different businesses.
median rating in Canadian map packs, against 4.9 in the US. A lower bar to clear.
Coverage car, home and business insurance across 56 states and provinces; local search across 109 metros.
Single capture August 2026 is one capture per geography. Three-capture medians with a published spread begin in September.
How a keyword is weighted not every search is worth the same, so visibility scores weight each keyword by its value: twelve-month mean monthly searches multiplied by cost per click. A term with 823,000 searches at $157.72 counts for more than one with 1,600 searches at $20, because it is worth more to win. The twelve-month mean is used rather than a single month because insurance volume swings hard across a year: "auto insurance" ranged from 368,000 to 1,830,000 in the last twelve.
Where weighting is not used feature presence rates are reported unweighted, one search one vote. Weighting those would bias them toward expensive commercial terms and hide the finding that informational searches behave completely differently from near-me ones.
Known limits volume and cost per click are national figures applied to state boards, because the source does not offer them sub-nationally. Roughly a quarter of local organic results are still unclassified, mostly single-market agency sites. Coverage is Google AI Overviews only.
Full method including what we get wrong.