On August 5, 2026, the Senate Commerce Committee voted 15 to 13 to advance the SCREEN Act, a federal bill that would force age verification onto any "covered platform" carrying content deemed harmful to minors. It did not actually advance, because too few senators showed up to meet quorum. That near miss is easy to file under political trivia. It should not be. The bill's definition of a covered platform reaches well past dedicated adult sites, similar state laws are already live in Louisiana, Texas and California, and retailers selling anything from wine to vape pens are already bolting ID checks onto checkout. Every one of those checks is a new step in your funnel, and every new step has a completion rate most teams have not tested.
Age verification is spreading from adult content into mainstream e-commerce checkout, driven by state laws already in effect in Louisiana, Texas and California and a near-miss federal bill, the SCREEN Act, that stalled in Senate Commerce on August 5, 2026. Document-based ID checks can cut conversion 10 to 20 percent at that step, facial age estimation cuts it only 2 to 5 percent, and self-declaration checkboxes satisfy neither the law nor your legal team. If any product line you sell is age-restricted anywhere you ship, test your verification method now, before a regulator tests it for you.
What just happened with the SCREEN Act, and why should a CRO team care?
The Stopping Child Predators and Sexual Exploitation Now Act, known as the SCREEN Act, is a federal bill that would require covered platforms to verify a user is not a minor before serving content deemed harmful to minors, and to bar VPN and proxy traffic from bypassing the check unless the service can otherwise confirm the user is outside the United States. On August 5, 2026, the Senate Commerce Committee held a markup where the bill cleared committee 15 to 13, but the vote did not formally advance because attendance fell short of quorum. In the same session, the Kids Online Safety Act advanced with unanimous support. The SCREEN Act is stalled, not dead, and Congress has a habit of reviving stalled age-verification bills within a session or two.
The Electronic Frontier Foundation flagged the real issue before the markup even happened: the bill's covered platform test is broad enough to sweep in services that have nothing to do with adult content, because it turns on whether a platform carries content a regulator could later argue is harmful to minors, not on whether it markets itself as adult. A general retail site selling alcohol, vapes, kitchen knives or mature-rated media could plausibly fall inside that definition. That ambiguity is why this is a checkout story, not just a policy story.
Which state laws are already forcing age checks into checkout, right now?
States have not waited on Congress.
| Law | Status | What it requires |
|---|---|---|
| Louisiana platform-level age verification | Effective July 1, 2026 | Age verification at both platform and app store level, with data handling rules |
| Texas App Store Accountability Act | Effective January 1, 2026 | App stores must verify age and secure parental consent before minors can download or purchase |
| California Age-Appropriate Design Code | Partial effect since March 2026, Ninth Circuit narrowed the injunction | Businesses must estimate user age and cannot reuse that data for any other purpose |
| Utah S.B. 73 | Signed March 19, 2026 | First state law targeting VPN circumvention of age checks |
Layer state alcohol, tobacco and vape rules on top of this and the picture gets more crowded. Regulators increasingly treat a checkbox or typed date of birth as non-compliant, since neither stops a minor willing to lie. If your store ships any age-restricted SKU into a state with an active law, the honor-system age gate you shipped in 2019 is now a liability.
Do not wait for a law that names your exact category. Pull every SKU that could plausibly be read as age-restricted, alcohol, vape, cannabis-adjacent, mature media, even kitchen knives in a couple of states, and check it against where you ship. If the answer is yes for even a handful of orders a month, that flow needs a real verification method, not a checkbox you added years ago.
How much does age verification actually cost you in conversion?
This is the number that gets skipped in most compliance conversations, and the one your finance team will ask about the moment legal signs off on a vendor. Identity verification vendor Didit publishes its own comparison: document-based checks, where a shopper photographs a government ID, cause a 10 to 20 percent drop in conversion at that step, while facial age estimation, a few seconds looking at a webcam, causes only a 2 to 5 percent drop. Treat that as vendor-reported rather than independently audited, but it matches what every checkout optimizer already knows: the more a step resembles paperwork, the more people abandon it.
Retail Technology Innovation Hub's roundup of age-estimation providers for e-commerce, published August 13, 2026, shows where the market is headed to close that gap. Vendor Incode reports a mean absolute error of just 0.95 years in the sensitive 13 to 17 bracket, combining estimation with a document check for borderline cases, and prices per approved check rather than charging for failed attempts, a model that tells you vendors expect drop-off here.
| Verification method | Typical friction | Legal defensibility in 2026 |
|---|---|---|
| Self-declaration checkbox | Lowest, near zero | Weak, increasingly rejected as non-compliant |
| Typed date-of-birth field | Low | Weak, easy for anyone to falsify |
| Facial age estimation | Low, seconds, no document required | Moderate to strong, accepted as a valid non-document method |
| Document upload plus liveness check | High, meaningfully higher abandonment | Strongest, the fallback for borderline or high-risk cases |
The pattern across every vendor is escalation, not uniformity. Nobody who has actually optimized this funnel puts every shopper through a document check. They start with the lowest-friction method that still satisfies the law, and escalate only a subset of borderline cases.
Facial age estimation versus document upload: which one keeps buyers in the funnel?
Facial age estimation analyzes a live camera image for roughly three seconds to produce an estimated age bracket, then discards the image immediately, no ID number, no stored photo, no name attached. It is the closer of the two methods to a frictionless checkout step, which is why supermarket chains piloting self-checkout age gates for alcohol have leaned on it rather than asking every shopper to scan an ID. The tradeoff is accuracy at the margins: someone close to a legal threshold gets flagged for a stronger check rather than approved outright, the correct failure mode for compliance but still an extra step for a real customer.
Document upload is the fallback, not the default, in every well-run 2026 deployment we have seen described. It is slower, asks for more personal information, and is the step most likely abandoned on mobile, where photographing an ID cleanly is harder than it sounds. It is also the method regulators trust most, because it is hardest to falsify. The right architecture is estimation first, document check only on escalation: cheap, fast check in front of everyone, expensive, slow check reserved for the cases that need it.
If you are shopping for a verification vendor, ask for approval-rate and drop-off numbers broken out by method, not one blended figure. A vendor that cannot separate estimation from document-check completion rates has not built the escalation logic that protects your funnel.
Is this only a problem if you sell age-restricted products?
Today, mostly yes. But the SCREEN Act's covered platform language, the pattern of state laws expanding scope year over year, and California's AADC decision all point the same direction: the definition of what requires an age check keeps widening, and rarely narrows back. A retailer with no age-restricted SKUs today should still know which product lines sit closest to that line, because the runway for adding verification well is weeks, not days. Bolting a vendor's default widget onto checkout the week a law takes effect gets you the 20-plus percent conversion hit instead of the 5 percent one, and adding the step usually means restructuring part of checkout, closer to website development work than a settings toggle. Run paid traffic into that funnel and an untested age gate wastes ad spend on top of the lost sale.
There is a quieter reason to get ahead of this: trust. A verification step that feels invasive, or stores data it does not need, reads as a red flag to privacy-conscious shoppers even when technically compliant. California's AADC explicitly bars reusing age-estimation data for any other purpose, precisely because regulators anticipated retailers repurposing it for marketing. Don't. The step should do exactly one job.
What should you actually test before a law forces your hand?
Treat this like any other funnel change worth running through conversion optimization, not a compliance box to check once and forget.
- Compare completion rate between your current age gate, if any, and a facial-estimation flow, measured at that specific step, not blended into overall checkout conversion.
- Segment mobile and desktop separately. Document upload performs worse on mobile, where photographing an ID cleanly is harder than it sounds.
- Test the escalation copy, not just the mechanic. Shoppers bumped to a document check need a clear, non-accusatory explanation or they abandon out of irritation.
- Track false-escalation rate alongside conversion. A miscalibrated model pushes far more people into the slow path than a vendor's marketing suggests.
- Confirm what data the vendor retains and say so on the page itself. Stated data minimization reduces abandonment at sensitive steps.
Frequently asked questions
What is the SCREEN Act and did it pass?
A federal bill requiring age verification on services carrying content deemed harmful to minors, with a covered platform test critics say reaches beyond dedicated adult sites. The Senate Commerce Committee voted 15 to 13 in favor on August 5, 2026, but it did not formally advance because too few senators met quorum, and it has not passed the full Senate or House.
Do online stores selling age-restricted products need age verification in 2026?
Increasingly yes. Louisiana's law took effect July 1, 2026, Texas's App Store Accountability Act took effect January 1, 2026, and California's Age-Appropriate Design Code took partial effect in March 2026. More than half of US states now have an active age-verification law, with alcohol, vape and cannabis retailers most exposed.
How much does age verification hurt checkout conversion?
It depends on the method. Vendor Didit reports document-based verification causes a 10 to 20 percent drop in conversion at that step, while age estimation causes only a 2 to 5 percent drop. A DOB entry popup performs worst on both counts, since it satisfies no serious legal standard while still adding a step shoppers abandon.
Is a checkbox or date-of-birth field enough to comply with age-verification laws?
No. Checkboxes and typed date-of-birth fields are increasingly treated as non-compliant under the state laws taking effect in 2026, which call for verifiable, auditable checks rather than an honor system. Retailers relying on a checkbox should expect to replace it.
What is facial age estimation and is it more accurate than a document check?
It analyzes a live camera image to estimate a shopper's age bracket without collecting a name, ID number or stored biometric template, and the image is typically deleted right after. Vendor Incode reports a mean absolute error of 0.95 years in the sensitive 13 to 17 bracket. It is not infallible, which is why deployments escalate borderline cases to a document check.
What should a CRO team do about age verification right now?
Audit whether any product line is age-restricted under a law your state or a customer's state already enforces. If so, test verification like any checkout step: measure completion rate by method, escalate from age estimation to document checks only for borderline cases, and never let a vendor's default flow ship untested.
The takeaway
The SCREEN Act stalling on a quorum count is not the end of this story, it is a preview. State laws already require real age checks for a growing list of product categories, and the definition of what counts keeps expanding rather than shrinking. Treat verification like any other checkout step: measure it, choose the lowest-friction method that still satisfies the law, and escalate only when you have to. Ship the compliant version that also converts, because the alternative is a regulator or a slow checkout doing the choosing for you.