Home / Methodology
Methodology
How rows were collected, what the data looks like, and where it is thin. Data as of 2026-10 (the most recent reported date across both data files).
How rows were collected
- Public sources only. No private data and no personal data.
- Every row has a source URL and a confidence label.
- Values are never invented. A blank field is shown as "not stated". This can mean the source did not publish the value or that it has not been captured yet.
- Published ranges are kept as ranges. No midpoint is invented.
- Annual equivalent is computed as monthly premium x12, 6 month premium x2, yearly premium x1.
What the data looks like
- insurify.com supplies about 38% of insurance rows (140 of 368).
- Coverage level is not stated on most insurance rows: 216 of 368. For Insurify rows the source methodology gives a range of limits, which is in the row notes in insurance.csv.
- State is blank on many insurance rows: 180 of 368. Most are national averages. ValuePenguin rows are Texas statewide averages according to that source's methodology.
- 35 insurance rows have no premium. They are availability or FSD benefit rows.
- Install costs skew toward Florida: 25 of 104 install rows (24%) name Florida, and 25 of the 31 rows that name a state (81%). Counted by source, about 45% of install rows (47 of 104) come from 4 Florida installer pages, including rows shown as "National".
- Install rows are typed: 53 of 104 are project totals. The others are hardware, component, permit, add-on, price difference, hourly rate or floor rows and are not full install costs.
- 65 install rows have a range but no published average.
computed from this dataset, see methodology
Confidence labels
| Label | Meaning | Insurance rows | Install rows |
|---|---|---|---|
| high | Published by the source directly. | 203 | 37 |
| medium | Crowdsourced or secondary table. | 165 | 62 |
| low | Single owner report. | 0 | 5 |
computed from this dataset, see methodology
Insurance sources
| Source URL | Rows | Share of dataset | Source type | Confidence |
|---|---|---|---|---|
| https://www.nerdwallet.com/insurance/auto/learn/tesla-insurance | 102 | 27.7% | rate_study | high |
| https://insurify.com/car-insurance/vehicle/tesla/model-3/ | 70 | 19.0% | rate_study | medium |
| https://insurify.com/car-insurance/vehicle/tesla/tesla-model-y/ | 70 | 19.0% | rate_study | medium |
| https://www.insurance.com/vehicles/tesla-insurance/tesla-model-3-insurance-cost/ | 32 | 8.7% | rate_study | high |
| https://www.valuepenguin.com/tesla-car-insurance | 26 | 7.1% | rate_study | high |
| https://www.denniscw.com/insurance/rates | 25 | 6.8% | sheet | medium |
| https://www.tesla.com/support/insurance/what-is-tesla-insurance | 15 | 4.1% | insurer_page | high |
| https://www.tesla.com/support/insurance/fsd | 14 | 3.8% | insurer_page | high |
| https://www.moneygeek.com/insurance/auto/tesla-car-insurance/ | 8 | 2.2% | rate_study | high |
| https://www.lemonade.com/investor/news/lemonade-autonomous-car-adds-support-for-tesla-fsd-v14-lite | 6 | 1.6% | insurer_page | high |
Install sources
Row counts and shares are computed from this dataset, shown with each source link and confidence label.
Known gaps
- No single owner reports for insurance yet (0 rows); 25 rows come from a public crowdsourced submission table, labeled medium confidence.
- Model S, Model X and Cybertruck are thinly covered: 10 for Model S, 9 for Model X, 3 for Cybertruck, compared with 166 for Model 3 and 145 for Model Y.
- Powershare has few sources: 7 rows from 3 source URLs.
- Install data is concentrated in Florida and the Tampa Bay area: about 45% of install rows come from 4 Florida installer pages, with few quotes from other states.
- Reddit and similar pages could not be read by our tools, so no rows come from them.
computed from this dataset, see methodology
Questions
Why does the home chart use only one source?
Different sources use different driver profiles and coverage. Using one source and one model year makes the bars more comparable with each other than rows from different sources.
What does annual equivalent mean?
It is the published premium converted to a yearly figure: monthly premiums times 12, six month premiums times 2, yearly premiums unchanged. Nothing else is added.
Why do some cells say not stated?
Either the source did not publish that value, or it appears only in the source's methodology notes and has not been added to the row yet. We leave it blank and show not stated instead of guessing.
Related: insurance explorer, install costs, affiliate disclosure.