Openbound Research / Methodology & Data
Research methodology & data
Clear denominators. Traceable evidence. Honest limits.
This supplements Openbound’s canonical sponsorship methodology, which remains the source of truth for employer matching, posting evidence, confidence and filing history.
Current job postings
We count distinct active posting IDs from active employer sources, eligible for the U.S. feed. Each report declares additional title and occupation rules. This is Openbound’s observed inventory, not a census of U.S. vacancies. One posting may represent multiple vacancies, and separate source IDs can describe similar work.
Employer and source provenance
Postings originate from employer careers sites and supported applicant tracking systems. Employer filing identity is resolved separately from the public brand. Verified matching and historical evidence remain separate dimensions. The report’s downloadable aggregates retain scope and extraction timestamps; its audit artifact retains posting IDs so each count can be reproduced.
H-1B evidence and sponsorship likelihood
Research reuses the product’s saved sponsorship conclusions: Highly likely, Likely, Possible, Unclear and Unlikely. These are evidence classifications, not measured candidate success probabilities. The posting’s explicit restriction or offer outranks historical evidence. Missing history does not mean an employer declined sponsorship. Read the canonical methodology for the current decision rules.
LCA and new employment vs continuation
Certified LCA records are steps in a filing process, not hires, visas or approvals. USCIS employer records distinguish initial-employment approvals from continuing-employment approvals; continuation alone does not establish new-hire demand. Research must not add these measures into an invented hiring count. Fiscal years and partial-year coverage refer to the source evidence, not the article’s publication year.
Role taxonomy and seniority
The design report begins with the existing Product Design / UX title family. Reproducible editorial subtypes distinguish UX Research, Design Systems, Content Design, Product Design and UX / UI / Interaction, in that order. Physical, apparel, footwear, accessories, packaging, mechanical and hardware titles are excluded, as are engineering, architecture, production and product/program/project-management titles that can occur in the broad design family. Titles can be ambiguous; title-based subtypes are not independent verified occupations. No seniority comparison is published until a reliable comparable field is available.
PERM is a separate pathway
PERM relates to permanent residence sponsorship. It is not used as a substitute for H-1B evidence in the design report or its employer ranking. Any future PERM report must state its own fiscal-year coverage, employer matching and case-status definitions.
Location and workplace
Country=US denotes U.S. eligibility and can include multi-country postings. The design report counts each posting once under its stored primary state; missing states remain visible as Unspecified / multi-location. These are not all eligible worksites. Workplace uses the stored Remote, Hybrid or On-site value; a remote flag supplies Remote when the explicit field is absent. Otherwise the field is Unspecified, never assumed On-site.
Freshness and updates
Data through is the UTC date when extraction began, not proof every source was checked that day. The JSON preserves generated_at and extraction_finished_at. Paginated reads occur over a short interval and are not a transactional database snapshot; live inventory may change during extraction. Each monthly or quarterly update regenerates the aggregate artifact and charts, receives editorial review, and updates the same evergreen article URL.
Limitations and responsible comparisons
Coverage depends on the employers and sources Openbound monitors. Classified signals depend on available evidence and its update cycle. Small groups are unstable, employer rankings measure current inventory with evidence rather than hiring performance, and salary coverage alone does not resolve regional ranges or conditional pay. The first report therefore does not publish salary correlations. No hiring outcomes, lottery selection rates or candidate success probabilities are measured.
Read the first design sponsorship report · Request a correction