Most career switches start as opinions: a friend raves about their field, a podcast makes a role sound exciting, a recruiter dangles a title. Opinions have one fatal flaw — they contain no information about what the job actually requires, what it pays, or where it is heading. Those questions have checkable answers. The U.S. government publishes structured data on nearly every occupation, free of charge, and almost nobody making a career switch ever opens it.
This guide teaches a method that reverses the usual order. Instead of deciding first and researching later, you let the data narrow the field: pick candidate roles, compare their skill requirements in the O*NET occupational database, validate demand and salary ranges with Bureau of Labor Statistics (BLS) figures, and then build a bridge plan for whatever gap remains. The two primary sources — O*NET OnLine and the BLS Occupational Outlook Handbook — are both free. Use them directly; everything else is commentary.
Why labor data beats intuition
Career switching suffers from an information asymmetry. People inside a role know what an ordinary Tuesday actually looks like; people outside the role rely on the job title, salary rumors and highlight reels. Two public datasets close most of that gap:
- O*NET (onetonline.org), maintained by the U.S. Department of Labor, profiles roughly 1,000 occupations. Each profile lists typical tasks, skills ranked by importance, knowledge areas, abilities, technology tools used, and education and training requirements (the Job Zone system), plus links to related occupations. Because the ratings are standardized, you can compare two careers side by side the way a database would.
- The BLS (bls.gov/ooh), the federal labor-market statistical agency, publishes a handbook profile for every occupation with median pay, entry requirements and a 10-year employment projection, along with state and metro wage programs underneath.
Two caveats before you start. First, both datasets are U.S.-centric. If you work elsewhere, use them as a structural model and look for your own country's equivalent — most national statistical offices publish occupation-level data of a similar shape. Second, these are aggregates: national medians and typical requirements. The final validation is always live postings in the region where you plan to work.
Step 1: Generate a shortlist of target roles
Do not start from a blank page. Take a task-level inventory of your current job: not the title, but what you actually do all day — coordinate schedules, analyze spreadsheets, coach colleagues, write reports, troubleshoot equipment, negotiate with vendors. Write the verbs down.
Then use two O*NET tools:
- The skills search on My Next Move (mynextmove.org, O*NET's career-exploration front end). Enter up to six skills you already have and it returns occupations that use them, ranked by how closely they match your profile. It is the fastest way to surface roles you did not know existed.
- Related occupations on any O*NET profile. Every occupation page links to careers with similar requirements, so you can hop in small, safe steps from what you know toward what you want.
AI brainstorming fits here too — asking a chatbot for roles that match your task inventory is a legitimate use — but treat the output as candidate names, not recommendations. Keep three to five occupations, each with its SOC code (for example, 13-1151.00 for Training and Development Specialists). The code keeps you anchored to a specific, data-rich definition rather than a vague job title that means something different at every company.
Step 2: Compare skill requirements across occupations
This is the heart of the method. Open the O*NET profile for your current occupation alongside each candidate and read three tabs:
- Skills — ranked by importance level. Write down the target role's top ten.
- Tasks — the literal day-to-day. If the task list makes you wince, no salary figure will fix that.
- Job Zone and knowledge areas — education, training and subject-matter requirements.
Now build a simple comparison table: for each of the target role's high-importance skills, mark whether you practice it today, practiced it in the past, or have never used it. The ratio tells you what kind of move you are making. If most of the core requirements map to work you have actually done, you are building a bridge. If almost nothing maps, you are restarting — still possible, but the timeline and cost are different, and you want to know that now.
Illustrative scenario: a retail first-line supervisor considers two candidates. Against a marketing manager profile, her experience maps weakly to the digital-campaign and analytics skills that dominate the profile. Against a training and development specialist profile, her coaching, scheduling, curriculum delivery and performance-feedback experience maps to most of the top skills. Same underlying ambition — helping people get better at their work — but the data points to the shorter bridge.
Step 3: Map your transferable skills — and translate the vocabulary
Overlap is not only about what you have; it is about how it is described. Every occupation has its own dialect, and career switchers routinely lose interviews because their resumes describe genuinely relevant experience in the wrong words.
Go through your overlap table from Step 2 and sort each matched skill into three buckets:
- Have and can name. You do it regularly and can describe it in the target occupation's vocabulary. These become your resume's headline lines.
- Have but mislabeled. You do the work but call it by your old industry's name. A teacher's 'classroom management' is 'stakeholder coordination and conflict resolution' in operations language. Rewrite it; do not re-learn it.
- Missing. Genuinely absent. These feed the plan in Step 5. Flag structural blockers here as well — professional licenses, clinical hours, security clearances. They cannot be shortcut, so if a candidate role is gated by one, either pick a closer role or budget for the long road deliberately.
Step 4: Validate demand and salary ranges with the BLS
Before you invest months, check that the destination is worth arriving at. Open each surviving candidate in the Occupational Outlook Handbook:
- Pay. Read the median and the percentile range, not just the median. Entry-level switchers usually land below the median, so if even the median does not clear your floor, stop here.
- Outlook. Each occupation carries a 10-year employment projection — the projected growth rate and how it compares with the average for all occupations. Growth meaningfully above average is a tailwind; projected decline means you will be competing for shrinking demand.
- Entry requirements. BLS summarizes typical education, licenses and on-the-job training — a useful cross-check against O*NET's Job Zone.
- State and metro data. BLS publishes regional wage data through the Occupational Employment and Wage Statistics program, so you can see what the occupation actually pays where you live. National medians can be badly misleading for any given metro.
One more caution: projections are national and long-horizon. They do not see this year's hiring freeze or next year's boom in your city. Cross-check with live postings — if you cannot find ten open jobs in your target title in your region within an hour of searching, treat that as a data point too.
Step 5: Plan the bridge
Take the gap you found in Step 3 and convert the missing items into a plan with three parts:
Courses and credentials. Target only the items in your missing bucket — nothing else. For most professional roles, the gaps are learnable skills on a four-to-twelve-week timeline (a spreadsheet language, a CRM platform, basic SQL, a design tool), not degree programs. Prefer learning that produces an artifact you can show.
Proof-of-work projects. Switchers have no job history in the target field, so they need artifacts instead: a portfolio piece, a volunteer engagement, a freelance gig, or an internal project at your current job that exercises the skill. Two small, real projects beat six certificates.
Network contact. Data tells you what the role requires; only people inside it can tell you what hiring actually looks like. Aim for a handful of informational conversations with people who have the target job — not just people at target companies — and ask them the same questions the BLS answered nationally: how did you get in, what do interviewers screen for, and what did you need on day one?
Then sequence the move. Apply first where your overlap is strongest, and accept an adjacent title as an entry point if it is the fastest route inside. Track your response rate as feedback: if tailored applications produce no interviews, the gap is larger than you estimated — and the tables from Steps 2 and 3 tell you exactly where to look again.
Where AI tools help — and where they do not
Inside this method, AI is genuinely useful: it can expand your candidate list in Step 1, decode a job description into a requirement inventory, and rewrite your bullet points in the target role's vocabulary (job-description analyzers such as CareerHelp's automate that decoding step). What AI cannot do is tell you whether a role fits your life — it has no access to your regional market, your salary floor or the truth about what you have practiced. That is what ONET and BLS are for. The rule that keeps AI safe in a career switch: every role an AI suggests must survive the ONET overlap check and the BLS demand-and-pay check before you invest in it.
Checklist
- Task-level inventory of your current role (verbs, not titles)
- Three to five candidate occupations with SOC codes, generated via My Next Move skills search and O*NET related occupations
- Side-by-side O*NET comparison: skills, tasks, Job Zone
- Overlap table sorted into have / mislabeled / missing, with structural blockers flagged
- BLS check: median and percentile pay, 10-year outlook, entry requirements, regional wages
- Bridge plan: specific courses, one or two proof-of-work projects, informational conversations
- Live-posting check for regional demand before you commit