How We Rank Markham Real Estate Agents
Most agent ranking sites describe their factors in a paragraph and stop. This page publishes the actual calculation: every weight, every tuning constant and a worked example using the current top ranked agent. The numbers below are read directly from the same configuration file the site uses to rank, so this page cannot drift out of step with the maths actually being run.
Weights last reviewed September 2026
The formula
Every agent receives a score between 0 and 1. Each of the six components is first normalised to a 0 to 1 value, then multiplied by its weight, then summed. Agents are ordered by that score. The same formula runs on the homepage and on every neighbourhood page. What changes between them is which reviews feed it, not how they are weighted.
score = (review average × 0.3) + (review
volume × 0.2) + (review recency × 0.15) + (profile completeness × 0.15) + (years licensed × 0.1) + (response rate × 0.1)
The six components
| Component | Weight |
|---|---|
| Verified review average | 30% |
| Verified review volume | 20% |
| Review recency | 15% |
| Profile completeness | 15% |
| Years licensed | 10% |
| Response rate to reviews | 10% |
| Total | 100% |
Verified review average · 30%
The mean star rating of every verified review on this site, rescaled so 1 star scores zero and 5 stars scores full marks. Reviews that fail verification are excluded and never counted.
Verified review volume · 20%
How many verified reviews an agent has, on a logarithmic curve. Going from 2 reviews to 12 moves the score far more than going from 90 to 100, so a high volume agent cannot bury a consistently excellent one.
Review recency · 15%
Recent experience counts for more than old experience. Each review contributes half its recency value after one year, so an agent who was excellent in 2019 but inactive since will sit below an equally rated agent working today.
Profile completeness · 15%
The share of profile fields an agent has filled in, including RECO registration number, languages, specialties and service areas. This is entirely within an agent control and rewards being transparent with the public.
Years licensed · 10%
Years registered as a real estate professional in Ontario, on a curve that flattens at 15 years. Experience matters but it is capped deliberately so it can never outweigh how an agent actually treats clients.
Response rate to reviews · 10%
The share of reviews an agent has publicly responded to. Responding to criticism counts exactly the same as responding to praise. Agents cannot delete reviews, only reply to them.
How each component is normalised
| Component | Calculation |
|---|---|
| Verified review average | (mean rating − 1) ÷ 4 |
| Verified review volume | ln(1 + review count) ÷ ln(1 + 40), capped at 1 |
| Review recency | mean of 0.5 (days old ÷ 365) across all reviews |
| Profile completeness | fields completed ÷ 10 |
| Years licensed | ln(1 + years) ÷ ln(1 + 15), capped at 1 |
| Response rate | reviews with a public response ÷ total reviews |
Tuning constants
| Review count for a full volume score | 40 |
|---|---|
| Recency half life | 365 days |
| Years licensed for a full experience score | 15 |
| Minimum verified reviews to hold a rank | 3 |
| Minimum reviews before response rate is scored | 3 |
Profile completeness fields
These 10 fields are counted. This component is entirely within an agent control, which is deliberate: it rewards being transparent with the public rather than rewarding spend.
- photoUrl
- bio
- brokerage
- yearsLicensed
- recoRegistrationNumber
- languages
- specialties
- neighbourhoodsServed
- phone
What we deliberately do not use
- Sales volume and transaction counts. We hold no licensed source for them, and busy is not the same as good.
- Reviews from anywhere else. We do not import or scrape Google, Zillow, Realtor.ca, MLS or TRREB data. Only reviews submitted and verified on this site count.
- Payment of any kind. No advertising relationship, subscription or sponsorship affects ranked position.
- Editorial judgement. No one on our side can move an agent up or down. There is no override in the code.
When weights change
The weights live in a single configuration file. If we change them, the change applies to every agent at once and this page updates automatically with the new numbers, because it reads from that same file. We will note the date of any change here. Weights last reviewed September 2026.
Common questions
Can an agent pay to rank higher?
No. Ranked position is produced entirely by the formula published on this page. There is no paid placement in the ranked list and no manual override anywhere in the code that builds these pages. If sponsored placement is offered in future it will be labelled Sponsored and shown separately from the ranked list.
Why is review volume capped rather than counted directly?
Volume is scored on a logarithmic curve that saturates at 40 reviews. Counting reviews directly would let a high volume agent bury a consistently excellent one with fewer transactions. The curve means moving from 2 reviews to 12 matters far more than moving from 90 to 100.
Does an older review count the same as a recent one?
No. Each review contributes half its recency value every 365 days. An agent who was excellent several years ago but has been inactive since will sit below an equally rated agent working today.
What happens to an agent with very few reviews?
An agent needs at least 3 verified reviews to hold a ranked position, so a single five star review cannot produce a number one spot. Below that threshold an agent still gets a full profile and appears in a separate listed but not ranked section. Response rate is also not scored below 3 reviews, and the remaining weights are renormalised so a new agent is not punished for a rate we cannot measure.
Do you use sales volume or transaction count?
No. We hold no licensed source for sales data and we do not scrape MLS, TRREB, Realtor.ca or Zillow. Ranking is based on verified client reviews, profile completeness and years licensed. Sales volume tells you how busy an agent is, not how well they treated the people they worked with.