PlaidCars is built on a simple idea: anyone considering a Tesla in the UAE should be able to see every relevant listing, every price change, and a clear picture of what a fair price looks like, without paying for it or signing up for anything. This page is the documentation of how we do that, what the limits are, and what we will not pretend to know.
What we collect
We track listings of Teslas for sale in the UAE across all major regional marketplaces. The current sources are listed at the bottom of every car detail page. We do not host listings; we record them, and we link out to the source for the actual transaction.
For every listing we observe, we capture the asking price, the mileage shown, the title status if disclosed, the colour, the trim, the year, and the dealer (or private seller) identity. We record a fresh observation each time we crawl, so when a price changes, a listing is reserved, sells, or disappears, we keep the history.
We also build a canonical record per physical car. When the same vehicle appears on more than one marketplace, we merge the listings under a single car so the page you see represents one car, not one ad. When two marketplaces disagree about a car’s identity, we hold the public page back until the conflict is resolved.
Vehicles, listings, and snapshots
We use three words consistently. A vehicle is one physical car. A listing is one ad for that car on one marketplace; a vehicle can have one or several listings live at the same time. A snapshot is one observation of a listing at a moment in time. The price-history chart on each car is built from snapshots.
A vehicle is public on this site only while it has at least one active listing and there is no internal conflict to resolve. Once a vehicle has not been observed for sixty days, we treat the listing as expired and the car drops off the public site.
Market buckets
Not every Tesla on the UAE market is competing in the same pool. We classify each car into one of four market buckets based on its specs region, export status, and showroom location: local retail (intended for sale and use in the UAE), re-export (imported and likely to be re-exported), domestic (locally-titled cars that are not standard local-retail inventory), and unknown (where we cannot tell).
This matters because re-export inventory and locally-priced retail cars sit on different pricing curves. Mixing them in a single comparison would distort the picture. By default the site filters to local retail. You can toggle the others on at any time.
How the deal score works
For every car, we compute a deal score by comparing its asking price against a tight cohort of similar cars. The cohort matches on five things: same model, same trim, year within one year either way, the same market segment (local retail, re-export, or domestic), and mileage within 20% of the target. We only consider currently active and reserved listings observed within the last 90 days, and we exclude the car itself from its own cohort.
When the target car’s mileage is unknown to us, the mileage band is not enforced. When the market segment is unknown, the segment constraint is not enforced. We would rather give a slightly looser cohort than refuse to score a car for a missing data point that the seller did not publish.
The score is the percentage difference between the car’s asking price and the median asking price across its cohort. We label the result:
- Great deal — at least 8% below the cohort median
- Good deal — 2% to 8% below
- Fair price — within 2% either way
- Above market — 2% to 8% above
- Overpriced — more than 8% above
If a car has fewer than five cohort matches at the time of computation, we do not show a deal score. Instead we say: “Not enough comparable listings yet to score this price — we need at least 5 similar Model Ys, and we found {N}.” This is the most important sentence on this page.
We will never invent a deal label from too small a sample. As the UAE market grows and more cars enter our index, the share of cars with a confident score rises. Until then, we would rather say nothing than say something we cannot defend.
Even when no score is shown, we display a cohort context block on the car page: the count of comparable cars, the price range, and the median mileage. The data is the point, regardless of whether we feel comfortable scoring it.
What we deliberately do not do
We do not predict prices. We do not estimate trade-in values. We do not score cars on condition we have not seen. We do not show the VIN of any car we list, anywhere on the public site — VIN is a privacy-sensitive identifier and we keep it internal. We do not republish private contact information from listings; private sellers appear as “Private Seller,” and dealers appear by their cleaned canonical name (for example “Kavak” rather than “kavak”).
We are independent of Tesla, Inc., of every marketplace we crawl, and of every dealer indexed on the site. No party pays us to rank, hide, or boost a listing. There is no premium tier and no advertising on the site.
How fresh is this data
We re-observe active listings on a rolling cadence. The internal dashboard tracks listings that have not been seen in 24 hours and vehicles overdue a check-in. A car’s last-seen timestamp is shown on its detail page. We treat freshness as a continuous discipline rather than a fixed promise: we will be honest about the lag rather than imply real-time accuracy we cannot deliver.
Challenging a label
If you believe a car is mis-labelled, a vehicle has been confused with another, a deal score is unfair to a seller, or a listing has been merged or split incorrectly, send a message via x.com/plaidcars with the URL and a sentence on what looks wrong. We log every report and act on the ones with merit. We do not change a label on request from anyone with a financial interest in the listing.
Versioning
This methodology is versioned. Material changes to the cohort rule, the deal-score thresholds, the market-bucket definitions, or the data sources are noted at the bottom of this page with a date.
Current version: 1.0 — 15 May 2026.