How we calculate rents.

1. Where does the data come from?

NinjaRent uses a three-source model. Each source is clearly labelled wherever it appears:

• Tenant submissions (achieved rents). Our intended primary source. Verified renters submit the exact rent they pay. Building-level accuracy, no intermediaries.
• ONS Price Index of Private Rents (achieved rents, aggregate). The official UK rental statistics series published by the Office for National Statistics. Provides mean rent per London borough, per bedroom count, updated monthly. We use this as the authoritative baseline when comparing a property to its borough.
• OpenRent listings (asking rents, v1 only). Individual property-level asking prices scraped from OpenRent's public search results. Used to bootstrap per-property data in v1 until tenant submissions are sufficient. Labelled as "asking" wherever shown, and replaced by tenant submissions as they come in.

Honest disclaimer: asking prices typically run 5–10% above achieved rents. Use the ONS borough median as your reality check.

2. How do we verify submissions?

A crowdsourced database is only as good as its worst fake submission. We use a defence-in-depth approach:

• Contactable Submitter: Every submission requires email verification via a magic link. This significantly raises the cost of bulk fake submissions.
• Address Validation: Submitted addresses must resolve to a real UK postal address.
• Statistical Outlier Detection: Submissions that fall more than 3 standard deviations outside the local distribution are held for manual review.

3. Privacy & Anonymity

Public views must never allow a third party to identify an individual tenant or their exact rent. This is a hard constraint.

• Minimum Cell Size: Single-submission buildings never show the exact rent publicly. We require a minimum of 3 submissions for an exact address before we display a range.
• Aggregation: If we have fewer than 3 submissions for a building, we aggregate the data into a wider street or postcode district ring (e.g., 250m or 500m).
• No Identifiers: Flat numbers and unit identifiers are never shown on the public site.

4. Data Decay

Rent data ages quickly. A submission from 2022 is a very different signal to a submission from 2026. To handle this:

• Submissions older than 24 months are excluded from primary comparable calculations.
• Submissions 12–24 months old are included but weighted lower and clearly labelled with their date.