How We Build the Dataset

Our multi-stage data pipeline combines OEM documentation, regulatory filings, and proprietary validation to produce the most reliable fitment database available.

The Problem With Fitment Data

Most fitment databases on the market are incomplete, outdated, or riddled with errors. A single wrong tire size recommendation can cause handling issues, speedometer inaccuracies, or even safety hazards.

Existing sources suffer from three problems:

• Fragmentation — OEM data is scattered across manufacturer portals, dealer networks, and regulatory databases, each using different schemas, naming conventions, and identification systems
• Version drift — Mid-cycle facelifts, regional variants, and optional wheel packages create hundreds of edge cases that static databases miss
• Translation errors — Cross-border data requires mapping between TPC codes, ETRTO standards, DOT specifications, and manufacturer-specific part numbering systems

We solve this with a multi-source ingestion pipeline that cross-validates every record before it enters the production dataset.

Our 7-Stage Data Pipeline

Technology

Our pipeline is built on a proprietary data processing stack developed over several years. Due to the competitive nature of this market, we don't disclose specific tooling or implementation details.

The full pipeline involves significant computation time per run. We process a large volume of raw records to produce the final validated dataset — the majority of raw inputs are filtered out through deduplication, validation, and constraint checks.

Update Cadence

The dataset is updated quarterly, timed to capture new model year introductions which typically happen in Q3 (European market) and Q1 (Asian manufacturers). Each update adds approximately 5,000-15,000 new fitment records and corrects any errors identified since the previous release.

API and embed widget subscribers receive updates automatically. CSV/SQLite customers can download the latest version from their account.

Accuracy Commitment

We regularly audit the dataset by sampling records and verifying them against original OEM sources. Accuracy is a core priority — every quarterly update includes corrections from the previous cycle.

If you find an error in the dataset, report it. Confirmed errors are corrected in the next quarterly update, or sooner for safety-critical issues.

Why Not Just Scrape It Yourself?

We get asked this sometimes. The honest answer: you could try, but it's harder than it looks.

The raw data is scattered across dozens of sources in different formats and languages. Normalizing vehicle names alone (handling the fact that a "VW Golf Mk8" is the same car as a "Volkswagen Golf CD1" is the same as a "Golf VIII 5H") requires a mapping table that took us months to build and validate.

Then there's the validation. Without cross-referencing against physical constraints and multiple sources, you end up with a dataset full of impossible fitments that will damage your credibility with customers.

We've been doing this for years. We've built the tooling, refined the validation rules, and established the source relationships. The dataset you're buying represents thousands of engineering hours — compressed into a single file you can start using today.