1. Mathematical Cadence Modeling: Frequency vs Engagement Velocity

Every recipient list has a finite engagement capacity. When a sender doubles their weekly email frequency from 2 to 4 sends per week, aggregate open volume rarely doubles. Instead, marginal opens decline, unsubscriptions accelerate, and spam complaint velocity increases exponentially:

• Linear Decay of Engagement: While the first email of the week captures high-intent readers, subsequent dispatches to identical cohorts yield diminishing returns.
• The Spam Complaint Inflection Point: When frequency exceeds recipient expectations, users stop searching for unsubscribe links and instead click the native “Report Spam” button in Gmail and Outlook.
• ISP Algorithmic De-prioritization: As user open ratios drop below 12%, machine-learning spam filters (Gmail Gemini, Microsoft SmartScreen) begin automatically sorting subsequent broadcasts into the Promotions or Junk folders.

2. ISP Volume Standard Deviation: Avoiding Anomalous Spike Penalties

A classic operational mistake is sending zero emails on Monday and Tuesday, blasting 350,000 emails on Wednesday morning, and resting until the following week. This erratic volume profile triggers defensive rate-limiting mechanisms:

3. Recipient Fatigue Index (RFI): Dynamic Algorithmic Throttling

Rather than sending the same number of weekly emails to every subscriber, modern deliverability systems compute a Recipient Fatigue Index (RFI) based on recent interactions:

• High Affinity (RFI < 20): Subscribers who opened or clicked within the last 7 days. Can safely receive 3 to 5 communications per week without complaint risk.
• Medium Affinity (RFI 21 – 60): Subscribers inactive for 8 to 30 days. Frequency automatically downshifted to 1 or 2 high-value digests per week.
• Fatigued / Cold (RFI > 60): Subscribers inactive for 31 to 90 days. Frequency restricted to 1 monthly digest, followed by sunset suppression if inactivity persists.

4. Strategic Silence Windows: Post-Campaign Reputation Healing

Following high-intensity promotional windows—such as Black Friday, Cyber Week, or quarterly inventory clearance sales—subscriber complaint rates and unsubscribe volume naturally rise. Continuing to send at peak volume during the immediate aftermath of a campaign causes serious deliverability damage.

Implementing a mandatory 48-hour to 72-hour Silence Window allows ISP Bayesian counters to normalize, complaint velocity to subside, and outbound IP reputation to heal before regular editorial schedules resume.

-- KumoMTA Automated Traffic Smoothing & Rate-Limiting
kumo.on('init', function()
  -- Smooth outbound dispatches across hourly time windows
  kumo.configure_domain('gmail.com', {
    max_connection_rate = '150/s',
    max_deliveries_per_connection = 80,
    max_concurrent_connections = 40,
    idle_timeout = '20s',
  })
  kumo.configure_domain('yahoo.com', {
    max_connection_rate = '100/s',
    max_deliveries_per_connection = 100,
    max_concurrent_connections = 30,
    idle_timeout = '30s',
  })
end)

5. Dedicated MTA Queue Rate Curve Configuration

Standard cloud ESPs lack granular controls to smooth volume spikes over time. Dedicated bare-metal MTAs (AutoSMTP with PowerMTA/KumoMTA) enable engineers to program custom delivery rate curves that automatically spread large marketing jobs evenly across designated time windows, preventing ISP spam triggers.

6. Cadence Engineering Comparison Matrix