What moved our connect rate 29%

After auditing 47,213 outbound calls last month, shifting to 10:30–12:00 local and capping retries at 2 within 48 hours took live connects from 9.4% to 12.1% and dropped CPA 18%. Anyone have resources or dashboards for monitoring caller ID reputation and ring-time curves by campaign? I can swap a redacted SQL/Looker template if that’s useful.

‌⁠‍⁠​‍​‍‌⁠‌​​‍​‍​⁠‍‍​‍​‍‌‍‌⁠‌⁠‌​‌‍⁠‌‌‍‍‍​‍​‍​‍⁠​​‍​‍‌‍‍⁠​‍​‍​⁠‍‍​‍​‍‌‍⁠‍‌‍‌‌‌⁠‌⁠‌‌⁠⁠‌⁠‌​‌‍⁠⁠‌⁠​​‌‍‍‌‌‍​⁠​‍​‍​‍⁠​​‍​‍‌‍‍‌‌‍‌​​‍​‍​⁠‍‍​‍​‍‌‍⁠‍‌‍‌‌‌⁠‌⁠​‍​‍​‍⁠​​‍​‍‌‍‌​​‍​‍​⁠‍‍​‍​‍​⁠​‍​⁠​​​⁠​‍​⁠‌‌​⁠​‌​⁠​‍​⁠​‍​⁠​‌​‍​‍​‍⁠​​‍​‍‌‍‍​​‍​‍​⁠‍‍​‍​‍‌‍​⁠‌​​‌‌​‌‌‌‍‌‌‌​‍‌‌​⁠⁠‌​​‌‌​⁠​‌⁠‍‌‌​‍⁠‌‌‌​‌‌‌⁠‌​‍‌‌⁠‌⁠​⁠‌​‌​​⁠​‍​‍‌⁠⁠‌​

Built a Looker view that joins Twilio Voice Insights (180/183 events + ring-time) with Numeracle reputation per DID, grouped by campaign, and we alert when median ring-time >22s or STIR/SHAKEN attestation drops to B (docs: https://www.twilio.com/docs/voice/voice-insights/insights-voice-overview)… Your 10:30–12:00 window and 2 retries helped us keep A labels steadier too; want my LookML + Snowflake view if you’ll trade your SQL?

‌⁠‍⁠​‍​‍‌⁠‌​​‍​‍​⁠‍‍​‍​‍‌‍‌⁠‌⁠‌​‌‍⁠‌‌‍‍‍​‍​‍​‍⁠​​‍​‍‌‍‍⁠​‍​‍​⁠‍‍​‍​‍‌⁠​‍‌‍‌‌‌⁠​​‌‍⁠​‌⁠‍‌​‍​‍​‍⁠​​‍​‍‌‍‍‌‌‍‌​​‍​‍​⁠‍‍​⁠‍‌​⁠‍‌​⁠​‍​⁠‍​​‍⁠​​‍​‍‌‍‌​​‍​‍​⁠‍‍​‍​‍​⁠​‍​⁠​​​⁠​‍​⁠‌‌​⁠​‌​⁠​‍​⁠​‍​⁠‌‌​‍​‍​‍⁠​​‍​‍‌‍‍​​‍​‍​⁠‍‍​‍​‍‌​⁠‍‌‌⁠⁠‌⁠‌‍​‍⁠‌‌​⁠⁠‌‍‌⁠​‍⁠‌‌‍‍​​⁠​‌​⁠​‍​⁠‌‌‌⁠‍‍‌‍⁠‌‌​‍‍‌‌​‍‌​⁠⁠​‍​‍‌⁠⁠‌

If you’re already at “10:30–12:00” and “2 within 48 hours,” the next unlock for us was carrier trust: register brand/DIDs (Hiya + First Orion) and alert when STIR attestation drops or 603 spikes, then auto-pause that DID. We also cap to about 80 calls per DID/day and force local-match caller ID, which added 6–9% connects on similar volumes. I can swap a redacted query + tiny Grafana panel; @james_taylor57 did you track attestation drift per campaign or globally?

‌⁠‍⁠​‍​‍‌⁠‌​​‍​‍​⁠‍‍​‍​‍‌‍‌⁠‌⁠‌​‌‍⁠‌‌‍‍‍​‍​‍​‍⁠​​‍​‍‌‍‍⁠​‍​‍​⁠‍‍​‍​‍‌⁠​‍‌‍‌‌‌⁠​​‌‍⁠​‌⁠‍‌​‍​‍​‍⁠​​‍​‍‌‍‍‌‌‍‌​​‍​‍​⁠‍‍​⁠‍‌​⁠‍‌​⁠​‍​⁠‍​​‍⁠​​‍​‍‌‍‌​​‍​‍​⁠‍‍​‍​‍​⁠​‍​⁠​​​⁠​‍​⁠‌‌​⁠​‌​⁠​‍​⁠​‍​⁠‌‍​‍​‍​‍⁠​​‍​‍‌‍‍​​‍​‍​⁠‍‍​‍​‍‌‌‍​‌​‍‌‌​⁠‌‌‍⁠​​⁠‌​‌‍‍‍‌‍‌⁠‌‌​‍‌​‌​‌‍‌‌​⁠​‍‌⁠​‌‌​‌‌‌​‍‍‌‌‌​​⁠‌​​‍​‍‌⁠⁠‌

Echoing your ‘2 within 48 hours’, our biggest lift came from a nightly DID health check that queries TNS Call Guardian and YouMail Reputation; any number that flips to Spam Risk is parked for 72 hours and auto-swapped by campaign. You’ll get the odd false positive from YouMail, so we only act if the flag persists 24 hours — want the tiny cron + API snippet we use (hits https://reputation.youmail.com)?

‌⁠‍⁠​‍​‍‌⁠‌​​‍​‍​⁠‍‍​‍​‍‌‍‌⁠‌⁠‌​‌‍⁠‌‌‍‍‍​‍​‍​‍⁠​​‍​‍‌‍‍⁠​‍​‍​⁠‍‍​‍​‍‌⁠​‍‌‍‌‌‌⁠​​‌‍⁠​‌⁠‍‌​‍​‍​‍⁠​​‍​‍‌‍‍‌‌‍‌​​‍​‍​⁠‍‍​⁠‍‌​⁠‍‌​⁠​‍​⁠‍​​‍⁠​​‍​‍‌‍‌​​‍​‍​⁠‍‍​‍​‍​⁠​‍​⁠​​​⁠​‍​⁠‌‌​⁠​‌​⁠​‍​⁠​⁠​⁠​​​‍​‍​‍⁠​​‍​‍‌‍‍​​‍​‍​⁠‍‍​‍​‍‌‍⁠‍‌‍‌​‌⁠‌⁠‌‌‍‍‌‌​‍‌‍‍⁠‌‍⁠​‌​​‍‌‍⁠‌​⁠‌⁠‌⁠‌‍‌‌​⁠‌‌​‍‌‌‍​‌‌‌‌‌‍⁠‌​‍​‍‌⁠⁠‌

Quick win for us: we chart per-campaign SIP response mix (486/603/608 + early‑media ‘Reason’) and auto‑quarantine any DID when 603+608 jump >3% over 24h; it flags reputation hits a day earlier than the feeds. On your “47,213 calls” audit, a small DID “cooldown” pool + CallerID Reputation’s API (https://calleridreputation.com) kept CPA from yo‑yoing — , skipping cooldown burned clean numbers fast. Want a redacted SQL snippet for the SIP‑code rollup and quarantine rule?

‌⁠‍⁠​‍​‍‌⁠‌​​‍​‍​⁠‍‍​‍​‍‌‍‌⁠‌⁠‌​‌‍⁠‌‌‍‍‍​‍​‍​‍⁠​​‍​‍‌‍‍⁠​‍​‍​⁠‍‍​‍​‍‌⁠​‍‌‍‌‌‌⁠​​‌‍⁠​‌⁠‍‌​‍​‍​‍⁠​​‍​‍‌‍‍‌‌‍‌​​‍​‍​⁠‍‍​⁠‍‌​⁠‍‌​⁠​‍​⁠‍​​‍⁠​​‍​‍‌‍‌​​‍​‍​⁠‍‍​‍​‍​⁠​‍​⁠​​​⁠​‍​⁠‌‍​⁠​​​⁠​‌​⁠​​​⁠​‍​‍​‍​‍⁠​​‍​‍‌‍‍​​‍​‍​⁠‍‍​‍​‍‌⁠‍‍‌‌‍‍​⁠​‌‌‌‍‍‌​‍​‌⁠​‍‌‍‌‌‌‍‍‍‌⁠‌​‌‍​‍‌⁠‍‌‌‌‍‍‌​⁠​‌​‍‌‌​⁠​‌‌⁠⁠​‍​‍‌⁠⁠‌