Scribd Just Paid $3 Million Because Its Job Ads Lacked Salary Ranges. Virginia Just Gave Every Applicant the Right to Sue. The Compliance Software Doesn't Exist.

The Problem

Start with the invoice. In September 2026, Scribd agreed to pay $3 million to settle a class action alleging its Washington job postings failed to disclose wage scales, salary ranges, and benefits information as the state's Equal Pay and Opportunities Act requires. The settlement class: 9,330 people who applied for Scribd jobs between January 2023 and August 2026, each in line for roughly $219. The company's offense was not underpaying anyone. It was posting job ads without numbers in them. Three million dollars for missing numbers.

Scribd is not an outlier. It is a data point on a curve that bends straight up. Washington employers faced 54 employment class actions in 2023; by 2025 the number had jumped to 773, and 658 more were filed in just the first half of 2026, according to Davis Wright Tremaine, driven substantially by pay transparency claims. One Seattle firm, Emery Reddy, filed 31 pay transparency class actions on its own. The settlements keep landing: Saint-Gobain Adfors agreed to $1.63 million, Cosco Fire Protection to $332,000. The plaintiffs' theory is brutally simple. Every job posting without a range is a potential class, every applicant a potential class member, and the damages sought run $5,000 or actual damages per person plus attorneys' fees. A single noncompliant posting, scraped and filed, becomes a seven-figure case. The posting is the product liability.

Now the map, because Washington was the warning and the country is the market. As of August 2026, sixteen states plus Washington, DC require employers to disclose pay in job postings: California, Colorado, Connecticut, Hawaii, Illinois, Maine, Maryland, Massachusetts, Minnesota, Nevada, New Jersey, New York, Rhode Island, Vermont, Virginia, and Washington. Five of those laws took effect in the last eighteen months: Illinois and Minnesota in January 2025, New Jersey in June 2025, Vermont in July 2025, Massachusetts in October 2025, Maine and Virginia in July 2026. Connecticut's amendments requiring benefits in ads took effect October 1, 2026, yesterday. Delaware's law lands September 2027. More are in legislatures now.

The trap is not the count. It is the variation. The laws differ on everything that matters: who is covered (four employees in New York, fifty in Hawaii), what must be disclosed (Colorado, Maryland, Minnesota, New Jersey, and Washington want benefits and compensation details too, not just base pay), whether remote postings count (they do, in several states, which means one remote listing can trigger five states' obligations at once), and how violations are punished (Massachusetts fines escalate to $7,500 to $25,000 per offense for fourth and subsequent violations; postings made within 48 hours count as a single offense, which is the kind of detail that decides real cases). Employment lawyers now advise multi-state employers to find the strictest standard and apply it nationally. As one Minnesota attorney told Bloomberg, the learning curve is steepest for smaller and mid-size companies without a broad national footprint: the companies with no employment counsel on retainer, no compliance department, and a careers page full of postings nobody has audited since 2024. That is the buyer. There are hundreds of thousands of them.

And the postings themselves are still mostly noncompliant. Indeed's data puts salary disclosure at 50 percent of US postings, a record high, up from 10 percent in 2019, which means half of all postings still omit pay. Monster's analysis of postings from February 2025 to January 2026 found that even in the most transparent sectors, roughly four in five ads omit pay information: tech discloses on only 10 to 11 percent of postings, healthcare and legal on 18 to 21 percent. Three in four job seekers now expect a salary range in the ad. The gap between what candidates expect, what the law requires, and what employers actually post is the market. It is measured in millions of noncompliant postings.

The Math: What a $3 Million Settlement Is Worth as Software

Price the risk first, because the risk is the budget. A Washington pay transparency class action now settles between $332,000 and $3 million, with defense costs on top running well into six figures even for cases that never certify. A mid-size employer posting 150 jobs a year across six disclosure states, with each posting a potential class of every applicant, carries a standing exposure in the low seven figures against a compliance cost, today, of either employment counsel reviewing postings at $400 to $800 an hour or nothing at all. Most choose nothing. The settlements are the marketing.

Now the buyer population. The Census Bureau counts roughly 1.96 million US firms with 50 or more employees. Narrow to the realistic buyer: multi-state employers that post jobs continuously in disclosure jurisdictions, large enough to be worth suing, small enough to lack in-house employment counsel. Call it 60,000 firms, about 3 percent of the 50-plus-employee population, an assumption stated plainly and revisited below. Price the product at $5,000 a year, deliberately under the $17,900 entry price of the closest existing tool, Ongig's Text Analyzer, because a focused compliance product should cost less than a general JD-writing platform. Sixty thousand firms at $5,000 a year is $300 million a year in US software revenue. That is the working TAM: not the whole HR tech market, just the employers for whom a missing salary range is now a litigation event.

The legal-exposure cross-check supports it. If 3,000 employers nationally face posting-violation claims over the next three years at an average all-in cost of $400,000, that is $1.2 billion in settlements and defense spend. Software that prevents the claim at $5,000 a year prices at barely 1 percent of the loss it insures against. Nobody buys software at 1 percent of expected loss and regrets it. Insurance is the oldest software category there is.

The Gap in the Market

Every regulated vertical that digitizes under enforcement repeats the pattern. The funded incumbents built for an adjacent job, which was writing better job descriptions or managing internal pay equity. The consultants sell hours against a problem that needs monitoring. The generic tools solve the requisition, not the posting. And the employers staring at a litigation curve that went from 54 to 773 filings in two years have no system that watches every live posting and keeps the record that proves compliance. Their lawyers told them the rule. Nobody sold them the evidence.

Their lawyers told them the rule. Nobody sold them the evidence.

The Solution

RangeCheck: the posting-compliance engine for the pay transparency era. Not a JD writer, not a pay equity platform, not an ATS. A control layer that watches every job posting the company has live, scores it against the posting-disclosure rules of every jurisdiction it touches, fixes what it can, flags what it cannot, and keeps the timestamped evidentiary record that turns a $3 million class action into a dismissed complaint. Five modules, priced for HR teams instead of law firms.

1. Posting crawler and auditor (core SaaS): the daily-use wedge. It inventories every live posting from the ATS, the careers page, and the major job boards, then scores each one against the jurisdictional rule engine: range present and in good faith, benefits disclosure where required, remote-reach analysis for multi-state exposure, promotion-transparency obligations where they exist. A 300-person company with 120 live postings gets a compliance score per posting, per state, every morning. This module is the moat: once two years of audited posting history live in your system, the employer is not migrating.

2. Jurisdictional rule engine (core SaaS): the seventeen regimes, encoded and maintained: coverage thresholds, disclosure content, remote-posting reach, cure periods, penalty schedules, private-right-of-action flags. Updated as laws change, which they do quarterly now: Connecticut's benefits amendments took effect this week, Delaware lands in 2027. The engine is the product's brain and its hardest asset to replicate, because every rule has edge cases that only show up in enforcement, and the engine learns from every demand letter the customer base receives.

3. Auto-remediation (core SaaS): when the auditor finds a posting missing a range in a disclosure state, it does not just flag. It pulls the approved range for the role from the compensation bands, drafts the compliant posting language, routes it through the configured approval chain, and pushes the fix to the ATS and boards. The Virginia cure clock is fifteen business days from written notice; the remediation module is built to clear it in one. Speed is the feature. The cure period is the deadline.

4. Evidentiary record ($4K/year add-on): the litigation defense, productized. Every audit, every fix, every approval, timestamped and immutable: the complete history of what every posting said, in every jurisdiction, on every day it was live. When the demand letter arrives, the employer hands counsel an exhibit instead of a scramble. When the class action files, the record shows good-faith compliance across the class period. Scribd paid $3 million without this. The add-on prices at a rounding error against that number, which is exactly why it sells.

5. Demand-letter response ($5K per incident): the services wedge. When the fifteen-day cure notice arrives, and for customers in Virginia and Washington it will, the module generates the remediation plan, executes the posting fixes across all boards, assembles the evidence file for counsel, and documents the cure inside the statutory window. High-margin incident revenue that converts to SaaS: the employer that just survived a cure clock does not go back to the spreadsheet.

Revenue Model

Unit economics for a mid-market employer: 400 employees, 150 postings a year across six disclosure states, two remote roles. Platform cost: $12K SaaS plus $4K evidentiary record, $16K a year all in. Compare the alternative: one Washington class action at the low end of the observed settlement range, $332,000, plus defense costs, plus the plaintiffs' attorneys' fees the statute awards. Or the counsel alternative: outside employment counsel reviewing 150 postings a year at even two hours each and $500 an hour is $150,000, for a point-in-time review with no monitoring and no evidentiary record. The platform costs a tenth of the counsel review and a twentieth of the cheapest observed settlement. Acquisition runs through employment-law firms as referral partners, HR associations, and the ATS marketplaces, at roughly $4K per logo. With the litigation curve as the tailwind, retention is structural: nobody cancels the system that holds their litigation defense. At five-year retention the LTV is about $80K against a $4K CAC. The ratio survives even deep skepticism about the incident revenue.

Market Size

TAM: 60,000 multi-state US employers exposed to posting-disclosure liability, at $5,000 a year for the compliance platform, is $300M a year. The buyer count is an assumption, roughly 3 percent of the 1.96 million US firms with 50 or more employees, and the price is anchored under the closest existing tool's $17,900 entry point. The legal-exposure cross-check: $1.2 billion in national settlement and defense spend over three years against software priced at 1 percent of the loss. Working TAM on the order of $300M a year, recurring, growing as the jurisdiction count grows.

SAM: not every exposed employer buys in year one. The sharpest near-term pain sits where the enforcement already is: employers posting in Washington and Virginia, the two private-right-of-action states, plus the high-penalty regimes of Massachusetts, California, New York, and Colorado. That enforcement-front population is on the order of 20,000 employers, or $100M a year in reachable software spend, and it is the portion of the market where the general counsel has already read the Scribd settlement.

SOM (year 3): 700 customers at a blended $7,200 a year in SaaS and add-ons: ~$5M in ARR at roughly 85% gross margin, a twentieth of the SAM, reachable through ten employment-law firm referral partnerships and the ATS marketplace channel. The wedge is the 200-to-2,000-employee company posting in three or more disclosure states: big enough to be worth suing, too small to have the lawyers.

Why Now

The settlements just started landing. The Scribd $3 million settlement received preliminary approval in August 2026 and hit the press in the last two weeks. Saint-Gobain's $1.63 million and Cosco's $332,000 are the same story at different scales. For three years the risk was theoretical; the plaintiffs' bar filed, the courts certified, and now the checks are being written. Nothing sells compliance software like someone else's settlement. Every general counsel in America can now put a number on the posting without a range. The number is $3 million.

Virginia doubled the private-enforcement map. Until July 2026, Washington stood alone in letting applicants sue directly over postings, which is why the class-action wave concentrated there. Virginia's new law adds a second private right of action with a fifteen-day cure clock, and the plaintiffs' firms that built the Washington playbook now have a second venue. The enforcement model is replicating faster than the statutes: expect the filing pattern to follow the law within eighteen months. Software bought after the first Virginia class action costs the same as software bought before it. The protection does not.

The jurisdiction count grows quarterly. Connecticut's benefits-in-ads amendments took effect October 1. Delaware's law lands September 2027. More states have bills in committee. Every new jurisdiction multiplies the rule-engine's value and the spreadsheet's failure modes at the same time, because each new law differs on thresholds, content, and remedies. A compliance product gets stronger with every statute. A shared doc gets longer.

Remote work multiplied every posting's exposure. Several states' laws reach remote postings tied to the state, which means a single remote listing can trigger five states' obligations simultaneously. The remote-work era turned every posting into a multi-jurisdictional filing. Nobody's 2024-vintage posting template was built for that. The auditor that scores per-posting, per-state exposure is the only honest answer to the remote question.

The EU directive went live in June. The EU Pay Transparency Directive took effect in June 2026 for companies with 100-plus employees, mandating pay-gap disclosure and employee access to pay data. Multinationals now face posting and reporting obligations on both sides of the Atlantic, and the US vendors that win the domestic compliance layer get pulled into the global one. Build for the seventeen states first. The continent is the expansion pack.

Startup Costs

Break-even lands around month 22: 400 customers at a blended $14K a year in SaaS and add-ons is $5.6M in ARR at roughly 85% gross margin, against a ~$1.3M annual burn for a team of eleven. That ramp assumes 80 customers by end of year one from the pilots converting and the law-firm referral channel activating, then the ATS marketplace and HR association channels compounding, and both scale without headcount, because the firms bring the trust, the marketplace brings the distribution, and the settlements bring the urgency. Year two is the business: once the evidentiary record holds two hiring cycles, the employer's counsel, its board, and its insurer all read from your audit trail, and at that point the switching cost is not the software license but the litigation defense, which no general counsel will vote to retype. Nobody migrates the exhibit file in year two of a class-action wave.

Limitations

The 60,000-firm buyer count is an assumption, not a census: 3 percent of the 1.96 million US firms with 50-plus employees, posited as multi-state active hirers in disclosure jurisdictions. If the true addressable base is half that, the TAM halves with it. The $5,000 price is anchored under Ongig's entry point, but Ongig sells a broader platform; a focused tool may need to price lower to win the mid-market, or higher if the evidentiary record proves its worth in the first defended case. Read the $300M as directional.

The Washington filing counts, 54 in 2023 to 773 in 2025, cover all employment class actions, not pay transparency alone; the Davis Wright analysis attributes the surge substantially to pay transparency, wage-and-hour, classification, and non-compete theories together. The pay transparency share of the wave is large but not the whole wave. The settlement figures, Scribd's $3M, Saint-Gobain's $1.63M, Cosco's $332K, are preliminary-approval amounts that can shift at final approval.

Virginia's private right of action is three months old. The prediction that plaintiffs' firms replicate the Washington playbook there is a forecast, not a fact; the fifteen-day cure provision may dampen filings relative to Washington, where no cure period gates the private action. If Virginia under-produces litigation, the urgency thesis weakens in the second-largest venue.

The "strictest standard nationally" advice that employment lawyers give large employers is a real substitute for the rule engine at the top of the market: a company that posts full ranges plus benefits everywhere has less need for per-state scoring. The product's answer is that the strictest standard is expensive in comp-strategy terms, breaks down for remote roles, and still leaves no evidentiary record. But for the Fortune 500, the substitute is real, which is why the wedge is the mid-market, not the enterprise.

Strongest Counterargument

Ongig already does this, the ATS vendors will bundle it, and the lawyers' actual advice, post the strictest standard everywhere, makes the whole rule engine unnecessary. You are building a feature of Greenhouse with a $5,000 price tag and a litigation story. Ongig has the JD-writing workflow, the ATS integrations, and the pay transparency flag today; when the demand grows, they deepen the flag into an audit trail and your startup is a checkbox on their pricing page. Mid-market HR buys suites, not point tools, and you are a point tool with a scary pitch deck.

This is the right objection, and it deserves a straight answer in four parts. First, the category error: Ongig is a writing platform. Its product answers "is this a good job description." The compliance engine answers a different question: "prove what every posting said, in every jurisdiction, on every day it was live." Those are different products the way accounting software and tax software are different products: adjacent, integrated, and never the same company. A flag is not a file, and the file is what the court reads.

Second, the ATS bundle argument misunderstands the ATS business. Greenhouse and Lever sell requisition workflow; their salary fields are inputs, not judgments. Building a seventeen-jurisdiction rule engine with quarterly legal updates and an immutable litigation-grade audit trail is a compliance business, not a workflow feature, and ATS vendors have consistently partnered rather than built for compliance depth. The ATS is the source system. The control layer on top of it is a separate company, the way payroll compliance sits on top of payroll.

Third, the "strictest standard" substitute fails exactly where the product wins. Posting the strictest standard everywhere means disclosing benefits detail in states that do not require it, constraining comp strategy in markets where the company wanted flexibility, and still producing no record of what was posted when. It is the expensive answer for companies that can afford expensive answers. The mid-market cannot, which is why the wedge is the 200-to-2,000-employee company, not the Fortune 500.

Fourth, the point-tool argument is answered by the switching cost. HR does buy suites, but litigation defense is not a suite feature: it is an evidentiary record with a chain of custody, and the company that holds two hiring cycles of timestamped posting history holds the customer. Nobody rips out the exhibit file to save a line item. The moat is not the feature. The moat is the record.

What You Can Do

If you run HR at a multi-state employer: audit your live postings this week, not this quarter. Pull every posting on your careers page and the boards, check each one for a good-faith range and the benefits disclosure your states require, and remember the remote roles: a remote posting can trip five states at once. Virginia's fifteen-day cure clock starts when the notice arrives, not when you get around to it. The spreadsheet from 2024 is not an audit.

If you are general counsel: price the exposure before the plaintiffs' bar does it for you. Count your postings in Washington and Virginia, multiply by the observed settlement range, and compare it against a compliance platform that costs less than one outside-counsel posting review. Then ask your HR team what the postings actually said last Tuesday. The gap between those two answers is the case.

If you are a job seeker: the range in the posting is now a legal right in sixteen states plus DC, and 75 percent of candidates expect to see it. When it is missing, you are looking at either an employer that does not know the law or one that does not follow it. Neither is a great sign. In Virginia and Washington, the missing range is also a potential claim, which is worth knowing before you apply.

If you are building this: start with the crawler and the rule engine, not the dashboard, because the per-posting audit is the daily-use wedge and the evidentiary record is the switching-cost engine. Hire the employment-law co-founder before the second engineer: in this market credibility is a feature and the demand letter is the sales call. Encode Virginia's cure clock first, because the fifteen-day deadline is the product's heartbeat. Decide on day one that you are the company the court reads.

The Bottom Line

America's job postings spent a decade without numbers in them. The numbers are now the law in sixteen states plus DC, the plaintiffs' bar turned the missing numbers into a class-action industry, and the settlements, $3 million for Scribd, $1.63 million for Saint-Gobain, $332,000 for Cosco, are the market's price discovery. The funded incumbents built better job descriptions and internal pay equity. The ATS vendors built requisition workflow. Nobody built the system that watches every live posting, scores it against seventeen rule sets, fixes it inside the cure clock, and keeps the record that defeats the suit. That system is worth on the order of $300 million a year. It sells to buyers the settlements already identified, through channels the law firms already built. The lawyers told everyone the rule. Somebody has to sell the evidence.

The evidence has never been worth more than it is right now.