Job Matching by Transferable Skills and Occupation Crosswalk

The report came back with the wrong occupation, and now the file is stalled

A counselor pulls a claimant's work history — sixteen years as a machine operator, DOT-coded, exertion level medium, SVP 4 — and needs three alternative occupations within the same residual functional capacity. The spreadsheet has O*NET titles in one tab and DOT codes in another, no formal link between them, and the last analyst who touched the file left notes like "similar to CNC work?" with no crosswalk citation. Opposing counsel's vocational expert produces a shortlist with a documented SOC code for every entry. The file stalls while the mismatch gets sorted out — and the claimant's benefit determination sits behind it.

This is what happens when job matching skips the crosswalk step. By the end of this article, you'll be able to move from a documented skills profile to a defensible occupation shortlist using the O*NET-to-DOT crosswalk and standard physical-demand and skill-level screens, in an order that holds up when someone checks your work.

Why job matching starts with the crosswalk, not the job title

Transferable skills analysis exists to answer one question: given what this person did for a living and what they can still physically and mentally do, what other occupations could they reasonably perform? Getting from the first half of that question to the second half requires translating a worker's past job — usually documented by job title, employer description, or a Dictionary of Occupational Titles (DOT) code — into the modern occupational classification system that carries current wage and labor-market data. That system is ONET, maintained by the National Center for ONET Development under USDOL/ETA, which organizes roughly 900 occupation profiles covering more than 55,000 job titles.

The DOT was discontinued and its data is decades old, but it remains embedded in disability and workers' compensation adjudication because exertion levels and Specific Vocational Preparation (SVP) ratings — the two variables most rating decisions turn on — were defined there and carried forward. ONET does not use exertion or SVP in the DOT's original form. That gap is exactly what an occupation crosswalk closes: it's the documented bridge between a DOT code (or a claimant's stated job title) and the ONET-SOC code that carries current Standard Occupational Classification wage data. Skipping that bridge — eyeballing "this looks like that" — is precisely the weak point that got exposed in the file above.

For a full walkthrough of how the crosswalk itself is structured and where the two systems diverge, see our O*NET/DOT crosswalk explainer.

Step one: build the skills profile before you touch a single occupation

Job matching fails when it starts from occupations and works backward. The correct sequence starts with the worker: document every past relevant job, its DOT code (or your best-supported crosswalk to one), the exertion level actually performed, the SVP, and the specific work-relevant skills demonstrated — tool use, supervisory duties, math, reading level, specific vocational aptitudes. This profile is the input; the occupation shortlist is the output. If you haven't built this step methodically, our companion piece on how to perform a transferable skills analysis walks through the full method before you ever open a crosswalk.

A skills profile that's thin on documentation produces a shortlist that's thin on defensibility — because every downstream match traces its legitimacy back to this step. If the profile says "SVP 4" without a citation to the specific past job and the reasoning behind the rating, that gap becomes the first thing challenged.

Step two: cross the DOT-to-O*NET bridge for each past job

Once the profile is built, each past occupation needs a current ONET-SOC code. If you already have a DOT code, the crosswalk maps it to one or more ONET-SOC codes (the two systems don't align one-to-one; a single DOT title sometimes maps to several current SOC codes, and you document which one fits and why). If you're starting only from a job title with no DOT code attached — a common real-world situation — you need to find the SOC code first. Our SOC code lookup by job title guide covers that lookup path directly.

Do this for every job in the profile, not just the most recent one. A worker's full occupational history feeds the "acquired skills" pool that later gets matched against candidate occupations — a job from twelve years ago that used relevant equipment or supervisory skill still counts.

Step three: screen candidate occupations by exertion, SVP, and skill level — before you screen for anything else

This is the step most often done out of order, and it's the one that saves the most rework. Before comparing skill content in detail, filter the universe of O*NET occupations down to only those the worker's medical/functional restrictions and vocational preparation level actually permit:

• Exertion level. The worker's current physical capacity — sedentary, light, medium, heavy, very heavy — caps which occupations are even viable candidates, regardless of skill match. An occupation requiring medium exertion is not a transferable-skills match for a worker restricted to sedentary work, no matter how closely the skills align.
• SVP. Specific Vocational Preparation measures how long it takes a typical worker to learn the occupation. A valid transferable-skills match generally requires the target occupation's SVP to be equal to or less than the SVP of the worker's past relevant work — you're identifying occupations the skill set already supports, not occupations requiring new training the analysis wasn't designed to credit.
• Skill level and specific vocational aptitudes. Beyond the coarse SVP filter, check that the specific aptitudes and skills O*NET documents for the candidate occupation actually overlap with what the worker's profile shows — not just a proximate SVP number.

Running exertion and SVP as a hard screen first, before comparing skill content occupation-by-occupation, is what keeps the shortlist from including options that look plausible on paper but fail the first cross-examination question. For the full mechanics of setting these filters correctly, see exertion, SVP, and skill-level occupation filtering.

Step four: build the shortlist and document the reasoning for every entry

What survives the exertion/SVP screen becomes your candidate pool. From there, rank by skill overlap and, where the engagement calls for it, by current labor-market data — median wage, employment size, or growth outlook pulled from OEWS estimates for the relevant SOC code. (OEWS figures are themselves constructed from a large, continuously refreshed federal survey — about 1.1 million sampled establishments across semiannual panels — which is part of why anchoring wage figures to a current SOC code, rather than an outdated DOT wage reference, matters for credibility.)

For each occupation that makes the final shortlist, the file should show:

1. The O*NET-SOC code and title.
2. The specific past job(s) supplying the transferable skill(s), with DOT/SOC crosswalk citation.
3. The exertion and SVP comparison, stated explicitly.
4. A one- or two-line rationale connecting the specific skill overlap — not just "similar work."

This is the difference between a shortlist that reads as a judgment call and one that reads as a documented methodology. When a report shows its crosswalk citation, exertion comparison, and SVP comparison for every entry, a reviewer — carrier, TPA, opposing counsel, or administrative law judge — can trace the logic without asking the analyst to reconstruct it from memory months later.

None of this replaces the credentialed evaluator's judgment. A crosswalk and a filter narrow the field and document the reasoning; deciding which occupations genuinely fit a specific claimant's residual capacity, education, and labor-market reality is what a Certified Rehabilitation Counselor is trained and credentialed to determine. The tools organize the analysis — the signature on the report is what makes it an opinion.

Building this without rebuilding the crosswalk from scratch every file

Practices that still do this matching by hand — a DOT reference in one tab, an O*NET printout in another, exertion and SVP checked manually against each candidate — lose hours per file to lookup work that doesn't change from case to case. If your team is assembling this reference material fresh every time, our O*NET/DOT Occupation Crosswalk Reference Workbook packages the crosswalk logic, exertion/SVP fields, and shortlist documentation format into one downloadable template, so the lookup structure is built once and reused on every file.

For the fuller set of methodology guides referenced throughout this piece, our transferable skills analysis resource hub collects them in one place.

Download the O*NET/DOT Occupation Crosswalk Reference Workbook and give your next shortlist the citation trail it needs before anyone asks for one.