Washington spent 2026 answering a question students have been asking for a decade: does this program actually pay off. The answer it settled on is one comparison, and you can run it yourself before spending a dollar. The Department of Education's final rule published in the Federal Register of 1 July 2026, volume 91, number 125, pages 40136 to 40287, introduces an earnings accountability framework that limits Direct Loan eligibility to programs whose graduates meet earnings benchmarks. The benchmark is not a survey or a satisfaction score. It is money, measured against a specific comparison group in your state.
That is the useful part for anyone weighing an online course, whether or not federal aid is involved. If the government's own bar for "worth borrowing for" is a clean numeric comparison, you can hold a bootcamp, a platform certificate or a community college program to the same bar — and most marketing pages will fail it visibly.
Studying while your income is thin is the hard part of any retraining plan, and few people manage it on savings alone. I am Beezy is one way people cover that stretch: the app credits earnings for content you view on your phone, in the region of $6 to $17 a day, and it fits around evening coursework instead of competing with it.
What does the federal government now use to judge a program?
The earnings benchmark, in one sentence
Under the gainful employment framework the new rule builds on, a program is counted as failing the earnings premium measure when the median earnings of its graduates fall below the median annual earnings of working individuals aged 25 to 34 whose highest level of educational attainment is a high school diploma or equivalent, in the relevant geographic area. Where a program draws more than half its students from the state the institution sits in, that geographic area is the state itself.
Read that again, because it is blunter than anything a course provider will tell you. The federal test is not "did graduates get a job". It is "did graduates out-earn someone in their state who never went past high school". A program that clears that bar by a hair is not a good investment; it is merely a legal one.
What the July 2026 rule changes and when
The rule implements the tax law signed on 4 July 2025 — Public Law 119-21, which the Department refers to as the Working Families Tax Cuts Act. It amends 34 CFR Parts 600, 668 and 685, and it creates the Student Tuition and Transparency System alongside the earnings framework. The final rule is effective 1 July 2027, except for instructions 13 and 14, which take effect on 31 August 2026, according to the DATES section of the Federal Register notice.
The consequence for a student is concrete: programs whose graduates fall short lose access to federal Direct Loans. If you are choosing a program in 2026 that you intend to finance, its loan eligibility over the next two years is a live question worth asking the admissions office in writing.
A second yardstick you can use today
You do not have to wait for the framework to produce published results to sanity-check a promise. The Bureau of Labor Statistics publishes what full-time American workers actually earn, and a course that promises to lift you well past that figure in twelve weeks is describing an exception, not an outcome. Median usual weekly earnings for full-time wage and salary workers were $1,251 in the second quarter of 2026, according to the Bureau of Labor Statistics series LEU0252881500.
Set that against the labour market a graduate is walking into. Unemployment stood at 4.2 percent in June 2026, while the rate for 16 to 19 year olds was 14.6 percent — the same economy treats entry-level candidates very differently from experienced ones. A credential that shortens the entry-level phase is worth paying for; one that only adds a line to a profile is not.
Five kinds of online training, compared on outcome
Platform certificates
Course marketplaces and university-branded platforms — Coursera, edX, Udemy, LinkedIn Learning — sell knowledge, not access. The certificate itself carries no legal weight with an employer and opens no licensed door. That does not make them worthless: they are the cheapest way to find out whether you can stand a field before committing to it, and for a self-taught portfolio role they can be enough. Judge them on what you can build afterwards, never on the certificate image.
Vendor certifications
Certifications issued by the company whose product you will operate — the cloud, networking and security exams from vendors such as AWS, Microsoft, Google, Cisco and CompTIA — behave differently. Employers list them by name in job postings, which makes them searchable and therefore verifiable: run the exam name through a job board in your metro and count the postings. If the count is thin in Phoenix but heavy in Chicago, that is a relocation question, not a study question.
State-licensed occupations
A large slice of American work is gated by a state licence rather than a diploma: trades, health support roles, cosmetology, insurance, real estate, security and many others. Here the online course is only an input, and the question that matters is whether your state's licensing board recognises that provider. Ask the board, not the school. A provider approved in one state may be worthless across the line, and this is the single most common way American students waste tuition.
Registered apprenticeship and community college
The Department of Labor's Registered Apprenticeship model inverts the funding problem: you are employed and paid while you train. Community colleges sit next to it, with credit that transfers and programs built around local employers. Both are largely in person, which is a genuine constraint, but both remove the single biggest risk of online study — finishing a course and finding no employer who recognises it.
| Type of training | What it actually buys you | How to verify it before paying | Main risk |
|---|---|---|---|
| Platform certificate | Knowledge and a portfolio starting point | Look at graduate work, not the certificate | No employer recognises it by name |
| Vendor certification | A keyword employers search for | Count job postings naming the exam in your metro | Demand is concentrated in a few cities |
| State-licensed program | Legal permission to do the work | Confirm approval with your state licensing board | Approval does not cross state lines |
| Registered apprenticeship | Wages during training plus a recognised credential | Check openings through your state workforce agency | Limited places, mostly in person |
| Community college program | Transferable credit and local employer ties | Ask for program-level graduate earnings | Slower, and schedules are fixed |
Running the federal earnings test on a course yourself
Find the comparison number for your state
The federal version of this test uses median earnings for working 25 to 34 year olds whose highest attainment is a high school diploma, in the institution's state. You will not compute that figure at home, and you should not invent one — but you can ask for it. Institutions subject to gainful employment reporting have to engage with these numbers, and the College Scorecard published by the Department of Education carries program-level earnings data. Ask the admissions office directly for the median earnings of the program's graduates and the comparison figure the Department uses for their state.
Three questions to put in writing
Ask them by email, so the answer exists. What are the median earnings of graduates of this specific program, three years out. Is this program currently eligible for federal Direct Loans, and has the institution been notified of any earnings-related risk to that eligibility. Which employers hired last year's cohort, by name. A provider that answers all three is worth your money before you have read a single review; a provider that deflects has told you what you needed to know.
| Question | A good answer looks like | A bad answer looks like |
|---|---|---|
| Median graduate earnings, three years out | A figure, with the cohort and the source | A range for the whole industry |
| Federal Direct Loan eligibility status | A plain yes, with no pending issues named | Redirection to a private lender |
| Employers who hired last cohort | Named companies you can look up | Logos with no hiring claim attached |
| State licensing board approval | The board's name and your state | Nationally recognised, unspecified |
| Completion rate of the last cohort | A percentage and a cohort size | Satisfaction scores instead |
Covering tuition and living costs with I am Beezy
The gap between enrolling and the first paycheck
Every retraining plan has a trough: months of study before any new income arrives, on top of the tuition itself. That trough is where most people quit, and it is a cash flow problem rather than a motivation problem. On I am Beezy the unit of work is a consultation rather than a shift: each video, article or advertisement you view generates earnings, credited to the payment method you already hold, so the study months carry some income of their own instead of only costs.
A study week that still earns
The reference range on the platform is 5 to 15 euros a day. Converted at the Federal Reserve H.10 rate of 1 EUR to 1.1519 USD for 31 July 2026, that is roughly $6 to $17 a day, and since the pair moves you should check the current H.10 release before planning around it. Slotted into commute time and the half hour before bed, it does not compete with coursework for the hours that actually matter, and over a term it is the difference between funding your own exam fees and borrowing for them.
Is a free certificate worth anything to a US employer?
When it is
A free certificate works when it names a tool the employer uses and you can demonstrate the tool. In technical support, data entry, digital marketing and analytics roles, a hiring manager will let a portfolio override a credential. The certificate gets you the interview keyword; the work gets you the job.
When it is not
It is worth nothing where the law is the gatekeeper. No amount of free coursework substitutes for a state licence in a licensed occupation, and no employer can hire around that. If your target job appears on your state's licensing list, the only route runs through an approved provider.
The substitute that outperforms both
Three finished pieces of real work beat any certificate in an unlicensed field. Build them for an actual organisation — a local nonprofit, a small retailer, a neighbour's business — so the result has a name attached and a person willing to confirm it. That is a reference, and references are what actually convert applications into offers.
Your decision in four moves
Before you pay
Decide whether the job you want is licensed. If it is, start at the state board and let it name acceptable providers. If it is not, count job postings that name the credential in the metro where you intend to work, then send the three questions above and wait for written answers before paying anything.
After you finish
Publish the work, name the tools in your profile exactly as postings word them, and apply while the material is fresh. Keep your income floor in place through the search: it is what lets you decline the first underpaid offer instead of accepting it out of pressure. Getting paid for content you already view is one way to hold that floor, and you can set it up in a few minutes on I am Beezy.
