How the Klunk Score Works
A 0–100 reliability rating for every make, model, and year. Built from NHTSA owner complaints and recalls. Recomputed daily.
The short version
The Klunk Score is a single number from 0 to 100 that estimates how reliable a specific model year is, based on real reports filed by real owners with the National Highway Traffic Safety Administration. More documented problems means a lower score. A perfect 100 means we have zero complaints on record. A 5 means it is a documented klunker.
Every score is recomputed against fresh data every day.
The data behind every score
Klunk does not generate scores from opinions, manufacturer surveys, or AI guesses. Every score draws from authoritative public data:
- NHTSA owner complaints. Hundreds of thousands of free-text reports filed by owners, each with an ODI tracking number, defect category, and severity flag. We pull this dataset daily and store every complaint with its original ODI number so we never double-count.
- NHTSA safety recalls. Every recall campaign for every make, model, and year. Recalls show on the dashboard alongside complaints and inform the score where appropriate.
- NHTSA defect investigations. When NHTSA opens an investigation, that signal is tracked too.
Source links: NHTSA complaints database and NHTSA recalls database.
How the score is computed
The score weighs the volume and pattern of complaints against the model year, alongside recall density. The exact formula is calibrated so small differences are readable at the high end (where you are picking between solid choices) and large differences are readable at the low end (where you are deciding whether to walk away).
Two design principles drive the score:
- Real data only. Every input is government-published and independently auditable. We can show you the underlying complaints on the same page as the score.
- Daily freshness. Owners file new complaints every day. The score reflects what is on file as of today, not a snapshot from last quarter.
What the score does not include: paid placements, manufacturer surveys, dealer reviews, owner-forum scrapes, AI-generated text, or any input that an affiliate partner could buy or influence.
The tiers
Every score also falls into a verbal tier so it is readable at a glance:
| Score range | Tier |
|---|---|
| 80 – 100 | Smooth Ride |
| 60 – 79 | Solid Pick |
| 40 – 59 | Proceed with Caution |
| 20 – 39 | Check Engine |
| 0 – 19 | Total Klunk |
Audit it yourself
Every dashboard page on Klunk shows the underlying complaints by category. The number behind every score is right there to read. If a 2018 Ford Explorer shows a Klunk Score of 38 with hundreds of transmission complaints, you can scroll the same page and see the actual complaint text owners filed. We don't hide the inputs.
Known limitations
- Volume bias. Best-selling models accumulate more complaints than rare models even if they are equally reliable per capita. Treat cross-make comparisons of very different volume tiers with caution.
- Newer model years are undercounted. A 2024 vehicle has had less time on the road than a 2018, so its complaint count is naturally lower. Use the Best Year to Buy page rather than raw cross-year scores when a model is very new.
- NHTSA underreporting. Many owners never file a NHTSA complaint at all. The score reflects what owners chose to report, not the true universe of issues.
Updates and corrections
Complaint data is refreshed daily via NHTSA's public feed. Scores are recomputed against the current dataset on every page load.
If you believe a complaint or recall is misattributed, email [email protected]. When we revise the methodology or add a new input, we note it here with a changelog entry.
Use of this data
The underlying NHTSA dataset is in the public domain. The Klunk Score and the editorial framing are Klunk's. Cite us if you republish — a link back to klunk.app is appreciated.
Last updated: 2026-05-05 — Methodology v1.