Trade school video: substantiating placement claims

The number in the video is the number you have to defend. If a video says 87% of graduates are placed within six months, someone can ask you to produce the cohort, the definition, the follow-up method and the exclusions, and "marketing rounded it" does not survive the question.

Trade schools and bootcamps are unusual because three separate regimes read the same sentence. Consumer-protection substantiation says an objective claim needs a reasonable basis before it runs. If the school participates in federal student aid, federal rules treat misrepresentations about the employability of graduates as a distinct problem. And your accreditor almost certainly defines how a placement rate is computed — cohort, window, what counts as in-field, who is excluded — in a way that may not match what a marketer would produce from the same spreadsheet.

The fix is not to stop using numbers. Numbers are the reason someone considers a nine-month program over a four-year degree. The fix is to make every number traceable before it is spoken aloud.

Build the claims-to-source table first

Before a single script is written, one table. It makes everything else fast, and it is what you hand over when someone asks.

Claim as it appears on screen	Underlying metric	Source of record	Definition owner	Last verified	On-screen caveat
"87% placed"	In-field employment rate, 2025 completer cohort	Accreditor annual report submission	Institutional research	2026-04	Cohort year + definition, held 3s
"median starting wage"	Median reported first-position wage	Graduate outcomes survey, n reported	Institutional research	2026-04	n, cohort year, region
"employers hire from here"	Named partner list	Signed partner agreements	Career services	2026-06	Partner, not guarantee
"financial aid available"	Title IV participation	Program participation agreement	Financial aid	2026-01	"For those who qualify"
"job-ready in 22 weeks"	Published program length	Catalog	Registrar	2026-08	Full-time schedule stated

Five rules govern the table:

The claim column contains the exact words, not a paraphrase. "Nearly nine in ten" and "87%" are different claims with different substantiation needs.
The definition owner is a person, not a department. When a reviewer asks how placement is defined, one named human answers within a day.
Last verified is a date, and stale rows expire. This is what stops a 2024 rate running in a 2027 campaign.
The caveat column is a production spec — what appears on screen, where, for how long. A caveat that lives only in a PDF is not on the video.
If a row cannot be filled, the claim does not ship. Not softened — removed. "Great job outcomes" is the same claim with the evidence stripped out, which is worse, not safer.

The parallel discipline for product claims is in spec sheet claims that don't survive testing; the failure mode is identical.

Gainful employment and where the disclosure lives

Gainful-employment disclosure obligations attach to Title IV programs, and the details of what must be reported and displayed have moved more than once. Do not take the current specification from a marketing article, including this one. Get it from your financial aid office in writing, dated, and put that date in the claims table.

What a video team needs is narrower and more durable:

Program-level disclosure is program-level. One institution-wide number used across creative for eight programs is a mismatch waiting to be noticed.
A disclosure that must appear at a particular point in enrollment is a compliance workflow. A video does not discharge it.
The video should not contradict the disclosure. It does not have to reproduce it.

The practical version: your video links to the program's disclosure page, and the numbers spoken are the numbers on that page, from the same record. If marketing is computing its own figure from a different export, you have already lost.

The instructor-plus-lab format

The format that carries this material is not the montage. It is one instructor, in the lab, with the equipment running, talking about what a student can do at the end of week three that they could not do at the start. Six to nine minutes, cut down into a dozen shorter assets.

Why this format: the audience is deciding whether a program is real, and real is demonstrated by an instructor who obviously knows the trade, in a room full of the actual equipment, using its vocabulary. It cannot be faked, and it survives a claims review untouched because nobody in it is making a claim.

Cold open on the work. Arc struck, machine cutting, terminal running. Two seconds, no talking.
Who the instructor is, in trade terms. Years in the field, licences held, what they did before teaching. This is the credential the audience actually weighs.
The equipment, named. Specific machines, specific software versions, specific brands. Vague equipment reads as a school that does not have any.
One skill, taught briefly. Not a full lesson — a demonstration of the level of instruction. The viewer should learn something small and real.
Week-by-week progression. First month, middle, last. Concrete tasks, not learning objectives.
What the first job looks like. Roles, shifts, tools, the parts that are unglamorous. Honesty here converts better than a salary figure and needs no substantiation.
Schedule and cost mechanics. Hours per week, whether it fits alongside a job, payment structures. No financial-aid claim without the caveat from the table.
The next step, named. An advisor, an open lab night, a tour. Not "apply now."

Nothing in beats one through six touches the claims table. That is the point — most of the asset ships without a compliance review, and the beats that do carry numbers can be re-cut without touching anything else. Explainer video formats handle the diagrammatic sections.

Salary figures and graduate stories

Six ways a well-intentioned salary claim goes wrong, all of which show up in career-school creative.

Occupational data presented as graduate outcomes. National wage data for an occupation describes the occupation. Used in a video about your program without saying so, it becomes an implied claim about your graduates. Name the source on screen and say it describes the field.

Median labelled as average, or the reverse. In wage distributions these differ meaningfully. Say which you have.

Survey respondents presented as the cohort. "Graduates report earning X" where X is the median of the third who answered the survey is not the graduate median. State the n and the cohort.

Top-of-band figures with "up to." The phrasing regulators recognise instantly, because it is what you write when the typical number disappoints.

Overtime and premium pay folded into a base figure. Break it out — the audience wants the structure.

Experienced-worker wages next to a first-job promise. A five-year wage in a video about a twenty-two-week program implies a timeline nobody said out loud.

For every figure that survives all six, the caveat goes on screen as timed text at the moment the number is spoken, not in a card at the end. Add timed text overlays to video is the mechanic; timing the caveat to the claim is what reviewers check.

Graduate testimonials are the strongest asset here and the fastest way to manufacture an unsubstantiated claim, because a graduate saying "I doubled my income in six months" becomes your claim the moment you publish it. Endorsement rules attach — see FTC endorsement disclosure — and generalised expectations created by an atypical story are the specific problem.

If the outcome shown is not typical, say so on screen, in the same shot, in the same size type as the claim.
Never pay for a testimonial without disclosing the material connection.
Never composite two graduates into one story.
Never generate the graduate. A synthetic person presented as a completer is a fabricated endorsement however it was produced — the fake review rule and AI testimonials sets out where that line sits.

Generated material has a clean role here and it is not people: facility exteriors you cannot shoot, seasonal plates, diagram backgrounds, equipment close-ups you would otherwise buy as stock — see replacing stock footage budgets with AI. Anything synthetic gets labelled per AI ad disclosure compliance.

Make a number change cost one shot, not one shoot

Schools run stale figures because updating them means re-editing everything. Build so it does not.

Put every claim in its own shot, with the number as an overlay rather than spoken by the instructor. When the cohort updates, you swap an overlay on a saved timeline and re-export. Keeping the piece as a re-renderable draft — reusable editor drafts — turns the annual refresh into an afternoon across the whole library. A 480p preview pass is free, subject to a short per-user cooldown, so the reviewer sees the change before you spend anything, and the final export is charged once regardless of clip count.

Keep one register mapping every published asset to the claims-table rows it depends on. When a row expires, you know which videos come down.

FAQ

Can we say "job placement assistance" without a placement rate?

Yes, and it is often the better claim, because it describes a service you provide rather than an outcome you cannot control. Then describe the service concretely: resume review, employer introductions, how many partner employers, what career services actually does in week one after completion. Specific service description outperforms a vague outcome number with this audience anyway.

Do these rules apply to a non-accredited bootcamp with no federal aid?

The federal aid rules and accreditor methodology do not, but consumer-protection substantiation does, and so does the state agency licensing private career schools in most states. Non-accredited operators read that as freedom. In practice it means fewer defined methodologies to point at when asked to justify a number, which makes the claims table more important, not less.

How do we handle a program with genuinely bad numbers?

Do not run the number, and do not run a vaguer version either. Run the format that does not need one — instructor, lab, equipment, week-by-week, what the first job looks like. It is the discipline any regulated local business uses when the outcome claim is off limits, and it is close to what driving school video marketing has to do. Sign-off then goes to the definition owner named in the claims table, on the specific cut rather than the script, because scripts change in the edit.