240
Resume Variants
29.0
Point Spread (Same Content)
4
Personas/Industries
9
ATS Profiles Tested
Text Conventions Matter More Than You Think
Across 240 resume variants from 4 personas tested against industry-matched ATS profiles, text convention choices alone created a 29.0 point spread in match scores — without changing a single word of actual content.
Key Finding
The best convention set scored 57.0 while the worst scored 28.0. Same work history, same skills, same achievements — different score simply because of headers, bullets, and section order.
Section Header Naming
Standard headers ("Professional Experience", "Skills", "Education") performed most consistently across all industries. Creative alternatives ("Career History", "Areas of Expertise", "Academic Background") showed more variance — sometimes scoring slightly higher on one ATS profile but lower on others.
Recommendation
Use standard, widely-recognized section headers. ATS parsers are trained on millions of resumes using conventional headers. Creative naming adds risk without measurable benefit.
Bullet Character Choice
The bullet character (•) scored highest. Dashes (-) and asterisks (*) performed similarly. The critical finding: numbered lists (1. 2. 3.) consistently scored 15-24 points lower than other formats across all ATS profiles.
Avoid Numbered Lists
Numbered bullets caused the largest single-variable penalty in this study. Some ATS parsers interpret numbered items as ordered sequences rather than parallel achievements, which disrupts keyword extraction.
Section Order
Section ordering had a measurable but small effect on match scores (2-3 point spread). The skills-first order ("Skills > Experience > Summary > Education") showed a slight edge in technical roles, while the standard order ("Summary > Experience > Skills > Education") was most consistent overall.
Consistency Across Industries
By testing 4 personas across technology, healthcare, finance, and marketing, we confirmed that convention effects are consistent regardless of industry. Standard conventions produced the most predictable scores — which matters because you rarely know which ATS a company uses.
Methodology
We generated 240 resume text variants from 4 industry-specific personas (software engineer, nurse, accountant, marketing manager) by systematically varying 3 conventions: 3 header naming sets, 5 bullet character styles, and 4 section orderings. Each persona was scored against 2-3 industry-matched ATS profiles using our algorithmic scan engine. This tests what text-based ATS parsers can measure — visual formatting (fonts, margins, columns) is invisible after text extraction and was not tested.
Disclosure
Vivid Resume is an AI resume platform. We used our own algorithmic scanner, not real ATS systems like Greenhouse or Lever. Full methodology and raw data available for review.
Limitations
4 personas: results cover technology, healthcare, finance, and marketing but may differ for other industries.
Simulated ATS: our algorithmic scanner approximates ATS behavior but is not a real ATS system.
Text-only scope: visual formatting was intentionally excluded.
Convention interaction effects are not fully analyzed.