How the leaderboard scores are made.

What this measures — and what it does not

This leaderboard measures how well our automated requests could fetch a real estate agent’s website and find clear facts about that agent. It does not test access from every AI product.

Some crawlers do not execute JavaScript. A site that shows text in a browser can therefore return little readable text to an automated request. Search and AI products handle websites differently; our result describes this test, not universal access.

What a score does not mean

• It is not a measure of an agent’s competence, experience, ethics, sales record or service quality.
• It is not an endorsement, recommendation or referral.
• It is not a prediction that an AI engine will mention this agent. AI engines also draw on reviews, directories, press coverage and other sources we do not measure.
• A low score does not imply a weak business. Many successful agents in Metro Vancouver work primarily through referrals, WeChat, Xiaohongshu or other channels where a website plays little part.

The scores are our opinion, formed by applying the published method below to publicly accessible web pages. We publish the method in full so that any reading can be checked.

Scope

Geography. Metro Vancouver: Vancouver, Surrey, Burnaby, Richmond, Coquitlam, Langley, Delta, New Westminster, Maple Ridge, Port Moody, Port Coquitlam, White Rock, Pitt Meadows, North Vancouver, West Vancouver, Tsawwassen, Ladner, Cloverdale, Fort Langley, Aldergrove, Bowen Island and Lions Bay.

Measurement window. Websites were measured between 21 and 24 September 2026. A website changes whenever its owner changes it; a score describes the site as it stood during that window, not as it stands today.

Requests. Each site received a small number of requests: the home page, robots.txt, sitemap.xml, llms.txt, and where one could be found, an About page. Requests were rate-limited and identified themselves. We did not log in, submit forms, or access anything a visitor could not reach.

Where the information comes from

Every item shown on the leaderboard comes from one of two places:

We assembled a candidate list of real estate professionals working in Metro Vancouver from publicly available sources, then independently located and tested their websites. The names and brokerages published here were read from each agent’s own site, not carried over from the candidate list — an agent is only listed once we have confirmed their name appears on the site we are scoring.

Where an agent’s site does not name their brokerage, that cell is marked “Not provided” rather than filled from elsewhere. On the current leaderboard this applies to 3 of 100 entries.

You can verify any row yourself: open the website linked from that row and look for the name.

We are not affiliated with, endorsed by, or connected to the Canadian Real Estate Association, REALTOR.ca, the BC Financial Services Authority, Greater Vancouver REALTORS, or any brokerage. Brokerage names appear only as statements of fact about where an agent works, as published by that agent.

Score 1 — Technical Accessibility 0–100

Could our automated requests fetch the site and read its raw HTML?

The first two checks carry the most weight in our score. Crawler rules and text available in raw HTML can affect what automated systems can retrieve, but each AI product handles them differently.

A site that returned no usable page to our requests — for example, because of DNS failure, a refused connection, a broken certificate, an error response or a timeout — is reported separately and excluded from the leaderboard. We did not test it in a full browser.

Score 2 — Content Readiness 0–100

Did the pages we could fetch state useful facts clearly? This score is applied to the home page and, where one is found, the About page.

Entity clarity checks whether a page states the person’s name, role and location together. Separate mentions in a header, an About page and a footer are weaker evidence than a clear passage tying those facts together.

Detection terms are maintained in both English and Chinese, so that Chinese-language sites are not scored as though they were empty.

The overall score

A site that returned no usable page is reported separately rather than given a combined score. Of 5,115 reachable sites, 82 could not be scored for content; their content component counts as 0 in the combined score.

The two halves are weighted equally as a judgment about technical access and content. Both component scores are published alongside the total so that readers can weigh them differently.

Ties are broken by technical score, then agent name alphabetically. Many agents share the same overall score; a one-place difference in rank is not meaningful.

For reference, technical and overall statistics cover 5,115 reachable sites. Content statistics cover the 5,033 sites scored for content; the remaining 82 count as 0 in the overall score.

The 100th place on the current leaderboard scores 80, which is above 97.4% of the reachable sites measured.

Who appears on the leaderboard

The leaderboard shows the 100 highest-scoring sites that meet all three conditions:

1. The site responds. Sites that time out, refuse connections, fail DNS, present a broken certificate, or block all non-browser clients are excluded. They score 0.
2. The site belongs to one agent. Team and brokerage sites shared by two or more agents are excluded, so that a shared site’s quality is not credited to one person.
3. The agent’s name appears on the site. This is what allows us to publish the name and brokerage on the strength of the site itself.

Across Metro Vancouver, we found a website for 9,592 of the 17,507 agents in our data. The public leaderboard shows the 100 highest-scoring sites that meet the conditions above.

This is not a ranking of all agents. An agent may be absent because we did not find their website, because their site is shared, because their site does not name them, because it returned no usable page to our requests, or because it scored below the current 100th-place score of 80. Absence carries no negative meaning.

We do not publish low scores against named agents. Only the top 100 are listed publicly.

Known limitations

We would rather state these than have them found.

• Readability is not citation. Our requests do not reproduce every AI product’s access. In a small test of 12 buyer-style questions put to Perplexity, 21 cited agent or team sites had an average combined score of 69.3, against 56.2 across reachable sites. Two cited sites fell in the bottom quarter because our test could not fully measure them; one returned no usable page to our requests. This small, selected sample cannot establish cause or tell us how often high-scoring sites are not cited.
• Home pages and About pages only. Deep blog posts, neighbourhood guides and listing pages are not scored. A site with substantial content elsewhere may be understated here.
• A single moment. Scores describe the site as it responded during the measurement window. A site that was slow, mid-migration or temporarily misconfigured that week is recorded that way.
• Weights are a judgment. The 50/50 split between technical and content, and the points within each, reflect our reading of how current AI engines behave. They are defensible, not objective. Both component scores are published so you can weigh them yourself.
• Automated detection is imperfect. Names rendered only in images or by JavaScript may be missed. A brokerage referred to by an informal short name may not be matched. Both failures cause exclusion, not a lower score.
• Coverage is incomplete. We could not find a website for a substantial share of agents. Some of them do have one, under a name our search did not reach. Not appearing in our data does not mean not having a site.

Corrections, removal and updates

If something is wrong, tell us and we will fix it. Wrong name, wrong brokerage, wrong website, a site that was down when we tested — write to us and we will re-test and correct it.

If you would rather not be listed, say so and we will remove you. No reason required, no questions asked. We act on removal requests within two business days. Removal is permanent unless you ask us to reinstate you.

Re-testing. Any listed agent can ask for a re-test at any time, free. If you have fixed something, we will measure it again and update your entry.

Update cadence. We intend to re-measure quarterly. Each published leaderboard carries the date it was measured; scores are never presented as current beyond that date.

This page describes the method behind a measurement of website characteristics. It is not advice about choosing a real estate professional. Anyone selecting an agent should consider experience, references, local knowledge, fees and fit — none of which this leaderboard measures.