We tested conversion fidelity across 1,000 public SEC Form 10-K and 10-Q corporate filings containing complex financial statements, nested notes, and mixed typography.
Evaluation Metrics
We measured three core factors: (1) Cell Boundary Alignment (preventing text collisions), (2) Number Parsing Precision (handling brackets, decimals, and international commas), and (3) Formula Preservation.
The Results
PDFparse achieved a 99.8% cell alignment score, outperforming conventional converters that scored below 84% on nested balance sheets. Furthermore, 96.4% of total sum rows were successfully reconstructed into native Excel formulas.
Key Takeaway
High-precision document automation requires domain-aware algorithms designed specifically for corporate numbers, not generic page scraping.