How Do Carriers Calculate Your Caller-ID Reputation Score?

Insurance agencies running outbound dialer operations hear it from their producers every week: "Nobody is answering. Are my calls showing up as spam?" The answer is not a simple yes or no. Behind every outbound call, carriers run a real-time reputation evaluation that determines whether your number rings clean, gets a warning label, or is blocked before it ever reaches the prospect. Understanding how that scoring works is the difference between a dialer session that connects and one that burns leads.

What actually goes into a carrier reputation score?

A call reputation score is not a single number published on a public dashboard. Each major carrier runs its own proprietary algorithm through partnerships with analytics providers like Hiya, TNS Call Guardian, and First Orion, which monitor calling patterns across billions of calls and assign reputation assessments independently per number PhoneBurner, 2026. A number can be clean on AT&T and flagged as "Scam Likely" on T-Mobile simultaneously because the carrier networks do not share a unified scoring model Convoso, 2025.

While carriers do not publish their exact methodologies, the behavioral signals that drive scoring are well understood from industry research and carrier guidance. CallLogic identifies seven primary factors: complaint rate, call volume and velocity spikes, low answer rates, short call durations, calling outside business hours, mixing use cases on one number, and CNAM mismatches CallLogic, 2026.

The complaint rate is the single most damaging signal. Every time a recipient taps "Report Spam" on their handset, that complaint feeds directly into the carrier's reputation system. Even a small number of complaints against a high-volume dialing number can trigger a label. The FTC received over 2.6 million Do Not Call complaints in fiscal year 2025, with robocalls representing the majority of those complaints FTC DNC Data Book FY2025. That volume of consumer feedback is continuously ingested by carrier analytics engines, which means the threshold for flagging a number is lower than most agency operators assume.

How does STIR/SHAKEN attestation feed into the score?

STIR/SHAKEN is the FCC-mandated caller authentication protocol that assigns one of three attestation levels to every call. A-level (Full Attestation) means the originating carrier verifies both the caller's identity and their right to use the calling number. B-level (Partial Attestation) means the carrier knows the customer but cannot verify their right to the specific number, common when numbers are ported between carriers. C-level (Gateway Attestation) is the lowest tier, where the carrier can only verify the call entered through a known gateway Sipstack, 2026.

The TNS 2026 Robocall Investigation Report found that in 2025, 85% of all voice traffic between Tier-1 carriers was signed and verified using STIR/SHAKEN, and 93% of that signed traffic carried A-level attestation. Smaller carriers trailed significantly, with only 17.5% of inter-carrier traffic signed BiglySales, 2026. This means that agencies using smaller VoIP providers may see a higher share of B or C-attested calls, which are more likely to be blocked or labeled regardless of the calling behavior.

Critically, A-level attestation does not guarantee a clean label. STIR/SHAKEN verifies identity, not behavior. CallLogic notes that carriers evaluate behavioral profiles independently of attestation, and a fully attested number with poor calling patterns will still be labeled or blocked CallLogic, 2026. Attestation is the identity layer; behavioral analytics is the trust layer. Both matter.

Which calling patterns push a reputation score into the danger zone?

PhoneBurner identifies eight call-pattern factors that carriers weigh when calculating risk: inconsistent call volume from a single number, cold-starting new numbers with high volume, short-duration calls under five seconds, double or triple dialing the same number, high share of ignored calls or straight-to-voicemail, consistently short answered conversations, manual blocks and spam reports from recipients, and dialing inactive or unassigned numbers PhoneBurner, 2026.

Bandwidth confirms the same pattern from the carrier side: carriers and analytics providers increasingly block or label calls that fit "typical robocalling campaign characteristics, such as high call volumes and short or unanswered calls together with other indicators of potential fraudulent activity" Bandwidth, 2026.

The threshold numbers are instructive. BiglySales reports that answer rates below 15 to 20 percent signal to carrier analytics engines that recipients are actively avoiding the number, which is treated as strong evidence of unwanted calling behavior BiglySales, 2026. For an insurance agency dialing 500 to 1,000 calls per day per producer, a contact rate that dips below 20% on any single number is a leading indicator that a label may be forming before the flag is visible on a handset.

What structural mistakes accelerate reputation damage?

Several operational patterns cause reputation damage that agencies often miss until answer rates have already collapsed. CallLogic flags four structural risks: routing all outbound activity through a single number concentrates every complaint and volume spike against one caller ID, rapidly rotating new numbers to escape labels accelerates flagging across the entire number pool, mixing inbound and outbound calls on the same line creates behavioral inconsistency, and suddenly changing local area codes mimics the neighbor-spoofing pattern that carriers actively flag CallLogic, 2026.

New numbers carry their own risk. Bandwidth notes that across all carriers, newly assigned phone numbers are more frequently mislabeled on their first several calls, not because of prior bad use but because of the absence of call history, making them indistinguishable from fraudulent number rotation in the carrier's view Bandwidth, 2026.

The consumer context keeps getting tougher. Hiya's 2026 State of the Call report found that 86% of unknown calls now go unanswered, and nearly half of consumers believe spam calls are getting worse, not better Hiya, 2026. Meanwhile, the FTC received over 2.6 million Do Not Call complaints in fiscal year 2025, and Hiya's 2026 State of the Call report found that 86% of unknown calls go unanswered Hiya, 2026. In that environment, the carrier algorithms are calibrated to be aggressive, and agencies operating without deliberate reputation management are burning numbers and leads against a system designed to filter them out.

How can insurance agencies audit their own reputation posture?

You cannot look up a universal score, but you can run a practical audit across the three carrier ecosystems. PhoneBurner recommends a diagnostic sequence: confirm the label on multiple devices and carriers by testing on iOS and Android handsets on different networks, isolate which specific numbers are affected, verify CNAM registration and caller ID consistency, audit call velocity and complaint sources, implement fixes, and monitor changes over several weeks PhoneBurner, 2026.

On the CNAM front, Bandwidth confirms that while spam labels are separate from Caller ID Name information, registering accurate CNAM helps recipients identify your calls and lowers the chance of mislabeling because consumers are less likely to report a call as spam when they recognize the name on screen Bandwidth, 2026.

For agencies managing dozens or hundreds of DIDs, the operational checklist is straightforward: pull a 30-to-60-day report of answer rate, call volume, and call duration per number, flag any DIDs with answer rates below 20% or unusually short average call durations, rest flagged numbers for several weeks rather than immediately rotating to a new number, register every active DID with the major analytics registries, and, for numbers ported from another carrier, allow 24 to 72 hours for SPID records to propagate before running full volume Sipstack, 2026.

What questions do agents ask most about carrier reputation scoring?

Can a number be flagged on one carrier and clean on another?

Yes. Each major carrier contracts with different analytics providers. AT&T uses Hiya, T-Mobile uses First Orion, and Verizon uses TNS Call Guardian. A number flagged on T-Mobile may ring clean on Verizon because the algorithms, complaint databases, and reputation thresholds differ between providers Convoso, 2025. This is why testing across multiple carrier networks is essential to understanding your real reputation posture.

How fast can a reputation score change?

Reputation scores are recalculated continuously based on real-time calling data. A single high-volume campaign on a clean number that generates a handful of spam reports can trigger a label within hours. Conversely, resting a flagged number and allowing complaint data to age off typically takes several weeks to meaningfully improve the score PhoneBurner, 2026. Reputation decays fast and recovers slowly, which is why proactive management beats reactive remediation every time.

Does registering with the Free Caller Registry actually help?

Registration alone does not prevent labeling, but it adds a verified identity signal that reduces the likelihood of automated flagging, particularly on new numbers that lack calling history. PhoneBurner includes registry registration as step one of its five-step number protection plan, noting that the combination of registration plus behavioral discipline is more effective than either tactic alone PhoneBurner, 2026.

How should agencies build reputation management into daily operations?

The reputation scoring system is not going away. Carriers have invested billions in spam detection infrastructure, and the consumer expectation of call screening is now baked into the handset experience. The agencies that thrive under this system are the ones that treat caller ID reputation as a daily operational metric rather than a crisis response lever.

Three disciplines separate agencies with stable answer rates from those constantly chasing flags. First, maintain per-DID performance dashboards with answer rate, average call duration, and complaint indicators. Second, operate a number rotation model where DIDs are rested before they burn rather than after, giving each number enough idle time to reset behavioral signals. Third, ensure every number in the pool carries A-level STIR/SHAKEN attestation by verifying that your telephony provider maintains current Robocall Mitigation Database filings and proper SPID records.

The agencies winning on answer rates in 2026 are not the ones with the best scripts or the most aggressive dialing speeds. They are the ones who understand that the carrier network is the first gatekeeper, and they manage their reputation posture accordingly.