Because ADP processes payroll for a huge share of American employers, an enormous fraction of the pay stubs that land in rental, lending, and hiring files carry ADP branding or an employer header produced through ADP. That familiarity cuts both ways. Reviewers develop an instinct for what an ADP stub “should” look like, but ADP output is configured by each client employer, so genuine stubs vary in logo placement, section order, and terminology. A fabricated stub that copies the general look can pass a glance test easily. What it copies far less reliably is the internal consistency of real payroll: earnings math, tax math, and year-to-date accumulation that all agree with each other.
Two cautions frame everything below. A polished, professional-looking stub is not proof of authenticity—careful forgeries reproduce ADP-style layouts convincingly, which is why material decisions should escalate to source verification rather than stop at file review. Equally, an awkward-looking stub is not proof of fraud: employers customize ADP formats, employees print from portals at odd sizes, and scans degrade everything. Treat each signal in this guide as a prompt for a better question or a stronger document request, never as a verdict about the applicant.
A polished pay stub is not proof of authenticity. Unusual metadata is not proof of fraud. Every signal here is a reason to ask a better question—never a verdict.
What a genuine ADP pay stub typically contains
ADP formats are employer-configured and change over time, so treat the following as commonly observed traits of genuine ADP-generated stubs rather than guarantees. The load-bearing question is never whether one visual detail matches a template you have seen before—it is whether the whole document is internally consistent:
- The employer is typically identified by name and address in a header block, often alongside ADP branding or an “ADP” reference in the footer, though many client configurations show only the employer's identity.
- A pay period is stated as an explicit date range with a separate pay date, and the pay date commonly falls on or shortly after the period end, consistent with the employer's cadence.
- Earnings appear in a structured table where hourly rows typically show rate, hours, and a current amount that equals rate multiplied by hours, with salaried rows showing a flat period amount.
- Deductions and taxes are itemized on labeled lines—federal income tax, Social Security, Medicare, applicable state and local lines, plus benefit deductions—rather than lumped into one figure.
- Most line items carry a paired current-period column and a year-to-date column, and the YTD figures grow plausibly with the pay date's position in the calendar year.
- Net pay is stated explicitly, commonly with a check or advice number and a deposit reference, and it equals gross earnings minus taxes and deductions on the same stub.
- A stub downloaded from an employee portal is typically a text-layer PDF whose figures can be selected and copied; a stub that is one flat image per page is a scan or a reconstruction.
Field-level checks anyone can run
1. Ask for the self-service download
The strongest version of an ADP stub is the file the employee downloads themselves from their employer's ADP portal or app. If your policy allows, ask the applicant to pull the stub fresh and send that exact PDF rather than a photo, screenshot, or rescan. Never ask for portal credentials.
2. Fingerprint the file on receipt
Compute a SHA-256 hash of the exact file you received before you review it, so the document you evaluated is provably the document you decided on. Ocolta's free scanner records this fingerprint locally in your browser without uploading the stub anywhere.
3. Recompute net pay from the stub's own lines
Gross earnings minus itemized taxes minus itemized deductions must equal the stated net pay, to the cent. An edit that inflates gross pay frequently forgets a downstream line. Work the subtraction explicitly and document any gap instead of eyeballing plausibility.
4. Multiply out every hourly earnings row
For each hourly line, rate times hours should equal the current amount, with overtime rows reflecting a sensible premium over the base rate. ADP-generated stubs carry this structure precisely; a row where the multiplication misses is a specific, documentable question.
5. Chain the year-to-date columns
If you hold two consecutive stubs, this stub's YTD for each line should equal the prior stub's YTD plus this period's amount. YTD figures must never decrease across the year. On a single stub, ask whether YTD totals fit the pay date's position in the year.
6. Sanity-check the withholding percentages
Social Security is commonly withheld at 6.2% of Social Security taxable wages up to the annual wage base, and Medicare at 1.45%. Pre-tax deductions change the taxable bases, so recompute against the stub's own taxable-wage figures where shown, and treat large unexplained deviations as questions.
7. Test the pay cadence
Weekly, biweekly, and semimonthly schedules produce pay dates that land in a predictable rhythm. Line the stub's period and pay date up against other stubs in the file, and against payroll deposits on any bank statement provided—the net amounts and dates should correspond.
8. Follow the check or advice numbering
ADP stubs commonly carry a check, voucher, or advice number. Across multiple stubs from the same employer, that number should progress in a consistent direction. Duplicated numbers, or numbering that jumps backward between consecutive pay dates, deserve an explanation.
9. Zoom in on the numbers that matter
At 400% zoom, edited figures often betray themselves: digits slightly off the baseline, a near-match font, or a broken right alignment in a numeric column. Compare a suspect digit against the same digit elsewhere on the stub before drawing any conclusion.
10. Cross-check employer identity details
The employer name, address, and any EIN shown should match the application and each other, and an EIN should follow the standard two-digits-hyphen-seven-digits format. Mismatched or vague employer identification on an otherwise crisp stub is a reason to strengthen the request.
11. Read the creator and producer metadata
A PDF records the software that produced it. A stub presented as a fresh ADP portal download whose metadata names a general-purpose design or editing tool, or whose modification date long postdates creation, merits a request for the original file—not an accusation.
12. Question a stub full of round numbers
Real payroll arithmetic produces odd cents: percentage-based taxes and benefit premiums rarely land on whole dollars. A stub where gross, every withholding line, and net are all round figures is not automatically fake, but it fits template output more than genuine ADP processing.
What Ocolta’s free scan checks automatically
Several of the checks above are mechanical, and mechanical work belongs to software. Ocolta’s free Integrity Scan runs entirely in your browser—the file never leaves your device—and reports the supported signals it can actually observe: whether the declared file type matches its binary signature, which software the PDF says created and produced it, creation and modification timestamps, incremental revision pointers, embedded scripts or attachments, and a SHA-256 fingerprint of the exact file you reviewed. For paystubs it also recomputes the core arithmetic—gross pay minus deductions against net pay—and explains benign causes to consider for each observation. The PDF tamper checker runs the structural subset on any PDF. Every result keeps “unable to determine” on the table; the scan never claims a pay stub is authentic or fraudulent.
When to escalate beyond file review
File review is triage, not proof. When the decision is material and questions remain, move to employment verification through an authorized income-verification service or a documented verification call to the employer using contact information you looked up independently—never a number printed on the stub itself, which a forger controls. Run every applicant through the same written policy with the same criteria, keep records of what you checked, and obtain legal review before any document finding contributes to an adverse action.
When the file itself deserves a deeper look before you escalate externally, Ocolta’s AI Deep Review ($5 per document, no subscription) asks a model to organize fraud-risk and AI-generation indicators—evidence locations, confidence, benign explanations, limitations, and next checks. It reports observations, not verdicts, and it can return “unable to determine.” See pricing for 5-review and 25-review packs.
Frequently asked questions
Can Ocolta tell me whether an ADP stub is fake?
No tool can. Ocolta's free scan runs locally in your browser and checks the file's structure, metadata, revision history, and payroll math—net-pay reconciliation, earnings-row multiplication, YTD chaining—reporting supported observations including an explicit unable-to-determine outcome. It never declares a stub authentic or fraudulent, and it does not contact ADP or the employer.
What is the fastest first check on an ADP stub?
Recompute net pay: gross earnings minus itemized taxes and deductions must equal the stated net, and each hourly row's rate times hours must equal its amount. Then confirm YTD columns exceed current amounts sensibly. This takes minutes, is fully objective, and sloppy fabrications fail it more than any visual check.
Is odd metadata proof that an ADP stub was forged?
No. Legitimate stubs get printed and rescanned, exported through browsers, merged into application packets, and stamped by document tools—each step rewrites metadata. Metadata anomalies justify requesting the employee's own fresh portal download; they never justify an accusation on their own.
Do all genuine ADP pay stubs look the same?
No. ADP's output is configured per client employer, so genuine stubs differ in branding, section order, and line labels—some show prominent ADP marks, others only the employer's header. Judge internal consistency of the math and identifiers rather than resemblance to any single ADP template you have seen.
Do not tell an applicant that a document is “fake” based on file review alone. State the observation, request the same follow-up your written policy requires in comparable cases, apply consistent criteria to every applicant, and preserve an unable-to-determine outcome. Obtain legal review before using document findings in an adverse decision.