Free Tools To GoFreeToolsToGo

Prediction Methodology

Last updated: May 8, 2026

All predictions on this site are statistical estimates derived from historical public data. This page explains exactly how each estimate is calculated so you can judge its reliability for your specific situation.

PERM Processing Time Estimates

Data foundation

We compute processing time statistics from DOL OFLC quarterly PERM disclosure files covering the most recent 8 fiscal quarters. Each row in these files represents one PERM case with a known filing date, decision date, and outcome (Certified, Denied, Withdrawn).

Only cases with a known decision date are included in processing time calculations. Cases still under review (no decision date) are excluded from statistics — they are used only for case lookup, not for computing percentiles.

Cohort assignment

Cases are grouped into filing-quarter cohorts by the quarter in which the PERM application was received by DOL (e.g., "2024-Q2" means the application was received between April 1 and June 30, 2024). This is the date on which statistical comparisons are based — not the decision date.

Processing time calculation

For each decided case, processing time = decision date − received date, in calendar days. We compute the following percentiles for each filing-quarter cohort:

  • p10 (best case) — 10% of cases in this cohort resolved within this many days. A case filing in the same quarter has a 10% chance of resolving this quickly.
  • p50 (median) — Half of cases resolved within this many days. This is the "most likely" estimate.
  • p90 (worst case) — 90% of cases resolved within this many days. Only 10% took longer.

Minimum cohort size

Cohorts with fewer than 10 decided cases do not generate processing time estimates — too few data points to produce reliable percentiles. Recent filing quarters often start with few decided cases and gain statistical significance as time passes.

Queue position estimate

The DOL FLAG processing times page shows the filing month currently being adjudicated for each queue (Analyst Review, Audit Review, Reconsideration). We compare your filing month against the current processing month to determine queue position status:

  • Already passed — DOL is processing months after your filing date (you are in the current batch or already decided)
  • Active — DOL is processing your exact filing month
  • Reaching soon — DOL will reach your filing month within the next 3 months at current pace
  • Ahead in queue — Your filing month is more than 3 months ahead of current processing

Audit Risk Estimation

Why this is an estimate

DOL OFLC quarterly disclosure files do not include an explicit audit flag. There is no column in the public data that says "this case was audited." DOL audit selections are not publicly disclosed at the individual case level.

How we estimate it

Audited PERM cases take significantly longer to process than analyst-reviewed cases because the employer must respond to an audit letter within 30 days, and DOL must review the response. We identify likely-audited cases as those whose processing time significantly exceeds the cohort median:

  • Cases with processing time > 1.5× cohort median are classified as likely audited
  • Estimated audit rate = (likely-audited count) / (total decided cases in cohort)

This method produces estimates consistent with DOL's published audit rate statistics for program-wide audit rates, but is not a direct measurement. Individual cases may have processing times above 1.5× median for reasons other than audit (complex legal issues, reconsideration, staffing).

Audit risk scores

For employer-level audit risk, we compute an estimated audit rate for the employer across all their PERM cases in our dataset:

  • LOW — Estimated audit rate < 10%
  • MEDIUM — Estimated audit rate 10–25%
  • HIGH — Estimated audit rate > 25%

Employers with fewer than 10 cases in our dataset do not receive an employer-level audit score — the sample is too small for a meaningful estimate. The combined score falls back to the cohort-level estimate.

Visa Bulletin Predictions

Data foundation

We store the final action dates and dates-for-filing from every monthly Visa Bulletin we have scraped. Predictions are based on the historical movement pattern of each category-country combination.

Prediction algorithm

For each category-country pair (e.g., EB2-IND), we compute a weighted moving average of the number of days each bulletin advanced over the 6 most recent months. More recent months receive higher weight. The predicted next-month date = current date + predicted advance.

Confidence levels

  • High — 3+ consecutive months of forward movement with low variance in advance amount
  • Medium — Some forward movement but with significant month-to-month variation
  • Low — Recent retrogression, stagnation at CURRENT, or insufficient history
  • N/A — Fewer than 3 months of data for this category-country pair

Retrogression

When a cut-off date moves backward (retrogression), confidence drops to LOW and the prediction reflects the retrogression trend rather than assuming forward movement will resume. Retrogression is common and unpredictable — it occurs when USCIS and DOS estimate that demand will exceed annual visa supply.

Limitations and Important Caveats

  • All estimates are based on historical patterns. Future processing times depend on DOL staffing, appropriations, policy changes, and adjudicator discretion — none of which are in our data.
  • PERM processing times increased significantly in 2020–2022 due to pandemic-related staffing. Our 8-quarter window includes only post-pandemic normalization data.
  • Visa Bulletin predictions do not account for USCIS demand forecasts, annual visa supply adjustments, or policy changes — factors that drive retrogression and sudden advancement.
  • Individual case outcomes cannot be predicted from statistical patterns. Your case may resolve faster or slower than any percentile estimate for your cohort.
  • These tools are not a substitute for an immigration attorney who has access to your specific case facts and current DOL/USCIS policy guidance.

Questions or Corrections

If you believe a calculation is incorrect or our methodology description does not match what is actually displayed, please contact us or see our data sources page for the primary sources to verify independently.