Coffee Meets Bagel Hit 70% Verification in 30 Days. Here's How.

Match Group spent years trying to convince the industry that verification was essential for trust. Bumble made it a launch feature. Yet adoption remained stubbornly low, with single-digit completion rates across most platforms where it was optional. Coffee Meets Bagel just reported that 70% of active users completed its new AI-powered video selfie verification within the first month of rollout. That is not a feature launch. That is a behavioral shift, and the industry needs to understand what caused it, because the answer changes what you should be building right now.

The High Intent Take

The trust calculation for online daters has changed. Not because the technology got meaningfully better, but because the problem got worse. The 2023 and 2024 wave of AI-generated profile photo scandals and bot proliferation stories shifted member psychology from "verification is an intrusion" to "verification is the baseline." CMB's 70% adoption rate in 30 days is evidence that the permission now exists. Operators who were waiting for that signal have it.

The harder question is whether verification actually fixes anything, or whether it simply shifts the trust problem one layer down. CMB is measuring phone number exchanges, which is the right metric, not matches, not messages, but willingness to move off-platform. If that conversion holds over time, this becomes a model worth copying. If verified users still ghost and still misrepresent themselves in person, adoption will crater as quickly as it rose. Adoption is the easy part. Sustained behavior change is the test.

The Privacy Trade Users Are Now Willing to Make

CMB is not the first to offer verification, but it may be the first to see this level of voluntary uptake without mandating the feature. Bumble's photo verification, launched in 2020, remains optional and has never disclosed completion rates publicly. Tinder introduced similar functionality in 2021; anecdotal reports from operators suggest single-digit adoption where it is not required. Hinge made verification available in 2022 with minimal fanfare and equally minimal uptake. The pattern was consistent enough that most operators concluded verification and voluntary adoption were incompatible.

Three factors distinguish CMB's implementation. First, the video selfie is checked against all profile photos, not just one. That raises the technical bar for fake accounts in a way that a single-photo check does not. Second, CMB enlisted Persona, a third-party identity verification provider, to handle biometric processing, a choice that may reassure privacy-conscious users but introduces real questions about data retention and regulatory compliance, particularly under GDPR. Third, and most importantly, the timing worked. CMB launched this well after the initial wave of AI-generated profile scandals had run through media cycles and settled into member consciousness as a known problem. If member tolerance for verification has genuinely shifted, it is because the problem got worse, not because the product improved. That distinction matters for how you build.

"Rachel Tee, CMB's Head of Trust and Safety, framed the feature as a response to 'a core source of user fatigue.' That is diplomatic language for what members have been saying more bluntly: they do not trust profile photos anymore. When your members stop believing what they see, verification stops being a friction point and starts being a selling point."

According to CMB, the verification process is completed within 24 hours, and early data suggests verified users are progressing to phone number exchanges at higher rates than unverified counterparts. That claim deserves scrutiny. CMB is offering correlation, not causation. Higher phone number exchange rates among verified users could reflect the verification effect, or it could mean that serious daters are more likely to verify in the first place. Selection bias, not verification, may be driving the outcome. Without a controlled experiment, you are measuring a signal, not proving a mechanism. CMB should run the experiment and publish the result if it wants the industry to follow.

Singapore as the Canary

CMB has already deployed more aggressive verification in specific markets. In Singapore, where romance scam complaints surged, the company introduced government ID checks. According to CMB, that intervention led to a significant drop in scam-related complaints. The Singapore playbook suggests that selfie verification may be a halfway house, a test of member tolerance before rolling out document-based identity checks globally. That trajectory has real implications for operators watching from the sidelines.

If 70% of users will hand over a video selfie, how many would submit a driving license or passport? The answer likely depends on perceived necessity. Singapore's regulatory environment and high-profile scam cases created a context where invasive checks felt justified. Whether that translates to London, New York, or Berlin is an open question, and the answer will vary by platform culture and declared user intent, not just geography. A casual-use platform will face different member resistance than one explicitly positioned around serious relationships.

"The data handling stakes are real. Persona processes the biometric data, but what happens to it after verification? How long is it retained? What happens if CMB is acquired or Persona is breached? UK and EU regulators have been explicit that biometric data is special category data under GDPR, requiring explicit consent and heightened security. CMB will need to demonstrate compliance not just at launch, but continuously. Any breach or regulatory action would set the industry's verification efforts back years."

What Competitors Should Be Watching

The metric to track is not verification completion rates. It is what happens next. CMB is measuring phone number exchanges because that is the indicator that verification is leading to real-world behavior change. If verification just shifts the trust problem, members verify, match, then ghost because the in-person experience does not match the profile, adoption will crater. The industry needs a longer data window than 30 days to know whether this is a structural fix or a temporary lift driven by novelty.

Competitors should also watch for demographic variation. CMB's user base skews toward serious relationship-seekers, which may make them more willing to verify than users on casual-first platforms. A 70% completion rate on Coffee Meets Bagel does not mean Tinder or Feeld would see the same uptake. Platform culture and declared intent both matter. Operators who lift and drop CMB's implementation without adjusting for their own member psychology will not replicate the result and should not expect to.

If verification becomes table stakes, smaller operators face a build-or-buy decision. Integrating third-party providers like Persona adds cost and complexity, but building in-house requires trust and safety expertise that many do not have. Platforms that cannot or will not verify risk becoming the dumping ground for accounts that cannot pass checks elsewhere. That is a user quality death spiral that compounds fast and is very difficult to reverse once your reputation as the unverified option is established. The cost of verification, absorbed now, is lower than the cost of that reputation. For most operators, the math on that comparison is not close.