- In 2026, carrier analytics (Hiya, TNS, First Orion) and STIR/SHAKEN attestation scoring penalize the pattern of your DID rotation, not just per-number dial volume. A clean high-volume number now often outlasts a randomly-rotated pool.
- Most pay-per-call insurance buyers report answer rate at the campaign level. That masks the fact that 70–80% of productive volume comes from the top quintile of DIDs, and those numbers decay on a measurable curve.
- The signature metric is DID half-life: days until a number’s effective answer rate falls below 50% of its day-one baseline. Disciplined pools hold 9–12 days in the pools we audit. Undisciplined pools collapse in 4–6.
- The fix is a four-tier pool (warm-up, production, cooldown, quarantine) with daily dial caps, 72-hour cooldowns after any flag event, and a weekly review you can run in 30 minutes.
- At $1–$3/month wholesale DID pricing, replacing a degraded number breaks even against a 20% answer-rate drop in well under 48 hours. Most buyers under-replace and over-extend.
We’re media buyers and lead-gen operators sharing what we see in the field. This isn’t legal advice. TCPA and carrier-analytics rules are genuinely complicated and vary by state, carrier, and vertical. Talk to an actual attorney before changing your consent flows or vendor contracts.
If Your Answer Rate Is Decaying Faster Than Your Dial Volume Explains, The Pool Is The Problem
If you’re a pay-per-call insurance buyer watching blended answer rate slide from 14% to 9% over two weeks, the instinct is to blame volume. It’s almost never volume. It’s the pattern.
A pay per call insurance buyer DID rotation strategy built in 2022, buy a block of 200 numbers, rotate randomly, replace when flagged, now generates exactly the fingerprint that carrier analytics flag in 2026. STIR/SHAKEN attestation scoring (the call-authentication framework the FCC mandated for U.S. voice carriers) and the analytics networks behind Hiya, TNS, and First Orion are scoring behavior shapes, not just per-number call counts.
This piece is operational mechanics. You’ll leave with a measurable metric (DID half-life), a four-tier pool architecture, and a Monday-morning review process. No theory.
Random Rotation Now Looks Like Spoofing To Carrier Analytics
The carrier-side analytics layer doesn’t care that you bought your numbers legitimately. It cares about pattern signatures.
What Carrier Analytics Actually Measure
Hiya, TNS, and First Orion ingest call-event signals from the terminating carriers (T-Mobile, Verizon, AT&T) and score numbers on a small set of behaviors:
- Short-duration call rate: how often the called party hangs up inside 6 seconds.
- Geographic mismatch: caller ID NPA-NXX doesn’t match the originating route’s geography.
- Velocity: dials per number per hour, and bursts after dormancy.
- Cluster behavior: groups of numbers from one originator behaving identically.
- Attestation downgrades: the originating carrier’s identity confidence dropped.
Random rotation across an undifferentiated pool produces a textbook cluster signature. Every number inherits the worst behavior of its peers because the analytics network treats them as a unit. One newly flagged number drags down the reputation of the 199 sitting next to it.
Why A Clean History Beats A Fresh Number In 2026
The counterintuitive part: a single number that’s been dialing steady volume for 90 days with normal call-duration distribution often outperforms a brand-new number ripped from a fresh block. Reputation is sticky in both directions. Carrier analytics weight history, and a stable history is an asset you actively destroy when you cycle numbers too fast.
This is why “buy more numbers” stopped working around mid-2024 and stopped working badly in 2026.
Attestation A, B, And C Aren’t Labels. They’re A Conversion Tax.
STIR/SHAKEN assigns one of three attestation levels to every outbound call. The originating carrier picks the level based on what it knows about you.
| Attestation | What it means | Typical buyer-side impact |
|---|---|---|
| A (Full) | Originator verified the caller and confirmed they own the number | Lands clean on most carriers, no spam label |
| B (Partial) | Originator verified the caller but can’t confirm number ownership | Often labeled “Spam Likely” on T-Mobile and Verizon |
| C (Gateway) | Originator can only verify the call entered its network | Near-guaranteed spam label on tier-1 carriers |
How Attestation Is Assigned
Your VoIP provider assigns attestation at the originating carrier. If your provider can’t prove you legitimately own the DIDs you’re dialing from, you get B. If the call originates from a gateway with no caller identity, you get C. ATIS publishes the technical spec, but the operator-level fact is simpler: your provider choice caps your attestation ceiling.
Attestation isn’t just a label question. It’s a conversion question. Calls landing with Attestation B convert through buyer-side IVR (the interactive voice menu that routes the call) at meaningfully lower rates than Attestation A. The gap shows up largest on T-Mobile and Verizon endpoints, where the “Spam Likely” overlay shows even on answered calls.
If your provider can only ship B, no amount of rotation discipline fixes the conversion ceiling. Solve attestation first, then optimize rotation.
DID Half-Life Is The Only Metric That Tells You When A Number Is Actually Cooked
Campaign-level answer rate lies. Here’s the math.
The Formula
Effective answer rate = (calls answered + IVR completions on Attestation A/B routes) ÷ total dials per number per day.
DID half-life (days) = days until effective answer rate drops below 50% of day-one baseline.
If a number opens at 18% answer rate on day one and falls to 9% on day 10, its half-life is 10 days. Simple. The reason nobody tracks this is that 70–80% of productive volume in a typical pool comes from the top quintile of numbers. The blended campaign number hides the decay of your workhorses behind the noise of your replacements.
The Ringba And Retreaver Setup
Neither Ringba nor Retreaver ships number-level answer rate in default reports. You build it with a custom tag:
- Create a tag on every outbound number (
did_id,did_first_seen_date,did_tier). - Pipe call events to a custom report grouped by
did_idand date. - Export the CSV weekly.
- In a spreadsheet, compute per-DID effective answer rate by day, then plot the decay curve.
- Half-life is the day the curve crosses 50% of the day-one value.
That’s it. Thirty minutes of setup, fifteen minutes of weekly review.
Half-Life Benchmarks Worth Knowing
For final expense and Medicare outbound under disciplined rotation (tier-tagged, dial-capped, cooldown-enforced), production-tier numbers typically hold a half-life in the 9–12 day range. Undisciplined pools (same volume, no tier structure) typically collapse to 4–6 days. The difference is roughly 2x the per-number working life. That’s the entire economic case for tiering.
Tiered Pool Architecture Is The Fix: Warm-Up, Production, Cooldown, Quarantine
This is the part most buyers skip. The dialer doesn’t know which numbers are healthy. You have to tell it.
The Four Tiers
Warm-up tier. New numbers, no history. Cap at low daily volume (20–50 dials/day) for the first 5–7 days to build a normal-looking call pattern. No bursts. No 400-dial days on a fresh number. That’s how new numbers get instantly clustered with bad actors.
Production tier. Attestation A, clean history, established baseline answer rate. Cap at roughly 250 dials/day per number for final expense and Medicare outbound. The 250 threshold isn’t magic. It’s where the answer-rate curve starts to bend across most pools we audit. Above 400 dials/day, degradation accelerates fast. In pools we audit, single numbers crossing that line typically show 18–25% effective answer rate degradation inside 14 days.
Cooldown tier. Any number that triggered a Hiya, TNS, or First Orion flag event in the last 72 hours. Pull from rotation immediately. Verify flag status via Free Caller Registry and the major reputation portals before re-promoting. 72 hours is a floor, not a target. Some flags clear in a day, some take a week.
Quarantine tier. Numbers that failed cooldown verification (flag persists after 72 hours) or are showing degraded answer rate without obvious flag trigger. Candidates for replacement. Don’t put these back in production hoping it sorts itself out.
Daily Dial Caps By Tier
| Tier | Daily cap per number | Selection priority |
|---|---|---|
| Warm-up | 20–50 dials | Last (only when production capacity exhausted) |
| Production | 200–250 dials | First |
| Cooldown | 0 dials | None, resting |
| Quarantine | 0 dials | None, replace or hold |
Configure your dialer to draw from production tier first, warm-up second, and never from cooldown or quarantine. In Ringba, this is a routing rule on the number pool. In Retreaver, it’s tag-based selection logic. Our operator’s guide to Ringba, Retreaver, and Invoca goes deeper on the routing logic side.
When To Replace Versus Extend Cooldown
The replacement math is straightforward. At wholesale DID pricing of roughly $1–$3/month plus minor porting and setup time, the breakeven against a 20% answer-rate degradation on a number doing 200 dials/day is well under 48 hours of continued use.
Replace when: flag persists after one full 72-hour cooldown cycle, or the number’s half-life has dropped below 4 days for two consecutive cycles. Extend cooldown when: flag cleared on at least one carrier and the number has a history of recovering.
Your Weekly DID Rotation Review Is A 30-Minute Process
Monday morning. Coffee. Spreadsheet.
- Export number-level answer rate from Ringba or Retreaver for the prior seven days.
- Compute per-DID effective answer rate and half-life. Anything with half-life under 4 days flags for review.
- Check reputation on the suspect numbers via Free Caller Registry, Hiya Connect, and TNS Call Guardian. Note which carriers are showing labels.
- Move numbers between tiers based on the data. New flags go to cooldown. Cleared flags go to production. Persistent flags go to quarantine.
- Decide replacement vs. extend on quarantine-tier numbers using the breakeven math above.
- Promote warm-up numbers that finished their 5–7 day ramp without incident into production.
Thirty minutes if your tagging is set up. Two hours if it isn’t, and you only do that setup once.
Reputation Monitoring Endpoints Worth Checking
The analytics network that determines what your prospect sees is not monolithic. T-Mobile uses one set of signals, Verizon another, AT&T a third, and the labels are rendered by Hiya, TNS, and First Orion depending on the device and carrier. A number flagged on T-Mobile but clean on Verizon and AT&T is a candidate for partial-use, not full retirement, but only if your dialer can route based on terminating carrier, which most can with a LERG lookup (NANPA’s number administration data feeds this).
How Tier Discipline Reads In An Audit
One more thing worth flagging. Aggressive random rotation across large undifferentiated pools is increasingly the pattern regulators associate with spoofing-adjacent behavior. A disciplined tiered pool with documented dial caps, cooldown logs, and per-number history reads very differently than “we bought 500 numbers and the dialer picks randomly.” If somebody comes asking questions, the documentation that comes out of a weekly review process is the documentation you want to have. Our TCPA lead buyer checklist covers the broader stack.
did_tier this week. Even before you build the half-life report, just knowing which numbers are warm-up versus production versus cooldown lets your dialer make smarter routing decisions. Most accounts see a 5–10% answer-rate lift inside two weeks just from the routing change.Local Presence Dialing Still Works. The Lift Shrinks Under Attestation B.
Quick note for buyers running local-presence stacks (matching the caller ID NPA to the prospect’s area code): yes, it still works in 2026. The lift hasn’t disappeared. But if your provider ships Attestation B because they can’t verify number ownership across rotating local-presence inventory, the “Spam Likely” overlay eats most of the local-presence advantage on T-Mobile and Verizon endpoints.
If you’re running local presence, push your provider for full Attestation A on the rotation block specifically. Some providers can. Many can’t. It’s worth the conversation before you blame the strategy.
If your contact rate dropped through 2025 and you’re not sure whether the cause is your dialer, your data, or your rotation pool, our outbound contact rate diagnostic walks the decision tree.
FAQ
Why does my answer rate degrade faster after I expanded my rotation pool from 50 to 200 numbers?
Larger undifferentiated pools generate stronger cluster signatures in carrier analytics. When 200 numbers from one originator dial identically, the analytics network treats them as a unit, and one bad actor drags the reputation of the rest. A 50-number pool with disciplined tiering will almost always outperform a 200-number random pool. Size isn’t the lever. Structure is.
How many dials per day can I run on a single DID before answer rate degrades?
For final expense and Medicare outbound, 200–250 dials/day on a production-tier number with Attestation A holds a stable curve. Crossing 400 dials/day typically triggers measurable degradation within two weeks. The threshold varies by vertical and provider, so measure half-life on your own pool before assuming numbers from someone else’s stack apply.
If a number gets a “Spam Likely” label on T-Mobile but not Verizon, should I retire it?
Not necessarily. Carrier-specific flags mean carrier-specific routing. If your dialer can route by terminating carrier (using a LERG lookup), you can continue using the number for Verizon and AT&T traffic while it cools down on T-Mobile. If your dialer can’t route that way, move the number to cooldown and re-evaluate after 72 hours.
Does cooling down a flagged number for 72 hours actually clear the flag, or do I need to file a remediation request?
It depends on what triggered the flag. Some flags clear naturally as call patterns normalize. Others, particularly First Orion and Hiya flags driven by complaint volume, may require an explicit remediation filing through the provider’s portal. Always verify clearance via Free Caller Registry and the major reputation portals before promoting the number back to production.
What’s the right ratio of warm-up to production numbers in a pool doing 5,000 dials/day?
If you’re running 5,000 dials/day at 250/number, you need about 20 production-tier numbers active at any time. Keep roughly 25–30% additional capacity in warm-up to cover cooldown rotation. A healthy pool looks like 20 production plus 6–8 warm-up plus a rolling cooldown bench of whatever flagged recently. Total active inventory: 30–40 numbers, not 200.
Can I verify attestation level on outbound calls before the prospect picks up?
Your VoIP provider can tell you what attestation they’re shipping on your traffic. Some surface it in dashboards. Many require you to ask. There’s no real-time pre-answer check available to the buyer, which is why provider selection is so critical. Once the call leaves your dialer, the attestation is set.
How is this different from just buying more numbers and replacing the flagged ones?
That’s the model that’s collapsing. Buying more numbers without tier structure means new numbers inherit the cluster reputation of the block they came in on, and replacement-as-strategy generates exactly the velocity signature that carrier analytics flag. Tiered rotation with measured half-life lets you extract roughly 2x the working life from each number and stop treating DIDs as disposable.
Talk To Elevarus About Your Pool Before The Next Carrier Tightening Cycle
Most pay-per-call insurance buyers don’t need more numbers. They need a tiered pool architecture, a half-life metric they actually track, and cooldown discipline that holds up when a carrier tightens its analytics again, which happens roughly every quarter now.
If you’re running final expense, Medicare, U65, or MedSupp pay-per-call and your blended answer rate is sliding, talk to our pay-per-call team. We audit existing Ringba and Retreaver setups, build rotation frameworks from scratch, and route exclusive call traffic into buyer stacks that need volume without the reputation drag. Book a free strategy call with Elevarus and ask about exclusive lead routing for your vertical and volume need. We’ll look at your actual pool data, not give you generic advice.
Worth reading next: Pay Per Call Insurance Buyer Pricing Tiers 2026: The Duration Map That Beats the Bid Sheet and If Your Contact Rate Dropped From 11% to 5% in 2025, It’s Not Your Dialer.
