- October 2022
- Engineering Memo · External Release
Sending frequency — how often the same recipients receive campaigns — is one of the most direct levers affecting ISP reputation. More frequent sending produces more engagement signals per unit time (positive) and more complaint opportunities per unit time (negative). The net reputation effect of a given sending frequency depends on the engagement quality of the list: highly engaged lists can sustain daily sending without complaint accumulation; moderately engaged lists reach complaint rate thresholds at weekly or biweekly frequency that would not be reached at monthly frequency.
This note documents how ISP reputation models respond to different sending frequencies, what the measurable thresholds are for major ISPs, and how to calibrate sending cadence against list quality to maintain inbox placement across frequency changes.
How ISP Reputation Models Process Frequency Signals
ISP reputation models observe sending patterns over rolling time windows. Gmail's domain reputation model evaluates signals over 30-day windows with heavier weighting on recent sends. A sender who is at the borderline of Medium/High reputation and doubles their sending frequency will, within 2–4 weeks, have twice the complaint volume attributable to their domain — because the same contacts who generate occasional complaints at weekly cadence generate twice as many complaints at daily cadence.
The specific mechanism: complaint rate is calculated as complaints per delivered messages over the observation window. If a sender delivers 100,000 messages per week and generates 60 complaints per week (0.06% complaint rate), doubling to 200,000 per week while the per-message complaint rate stays constant produces 120 complaints per week — still 0.06%. So far, no reputation impact from the frequency increase alone.
The issue arises when frequency increase drives per-message complaint rate increase — which it does for virtually all programmes, because more frequent sending fatigues some recipients who were previously tolerant. The recipient who was willing to receive weekly campaigns marks the daily campaign as spam. The per-message complaint rate rises from 0.06% to 0.08%, and with double the volume, the weekly complaint count is now 160 rather than 60. The ISP's reputation model sees a 2.7× increase in weekly complaint volume from the same domain. Even though the per-message rate rose only modestly, the absolute complaint volume increase produces visible reputation impact.
Figure 1 — Sending Frequency vs. Complaint Rate: The Fatigue Effect
Higher sending frequency with low-engagement lists crosses the 0.08% alert threshold (and 0.10% ISP limit) quickly. Highly engaged lists sustain much higher frequencies.
The Frequency-Complaint Rate Threshold by Engagement Tier
The relationship between sending frequency and complaint rate is not uniform — it is moderated by the engagement quality of the list segment being sent to. Highly engaged segments (30-day openers) can sustain significantly higher frequencies before complaint rates cross ISP concern thresholds. Lower-engagement segments reach those thresholds at much lower frequencies.
Empirical observations across managed sending environments suggest the following sustainable frequency ranges by engagement tier, where "sustainable" means complaint rates remain below 0.06% (well below the 0.08% alert threshold):
High engagement (opened in last 30 days): daily to 3×/week sending is typically sustainable. These recipients are actively engaged with the sending programme and have established a pattern of email interaction that reduces fatigue-driven complaints. The ceiling exists but is high enough that most programmes don't approach it for this segment.
Moderate engagement (opened in last 31–90 days): weekly to biweekly is typically sustainable. More frequent sending to this cohort produces progressive fatigue as recipients who were previously engaged but have become less active find daily or near-daily emails intrusive. Complaint rates for this cohort climb noticeably above weekly frequency.
Lower engagement (no open in 61–90 days): monthly or less is sustainable. This cohort is at the boundary of list retirement — their complaint rates are elevated even at low frequencies, and more frequent sending amplifies this risk significantly. The correct approach for this cohort is a single re-engagement campaign before retirement, not sustained high-frequency sending in an attempt to drive re-engagement through volume.
ISP-Specific Frequency Sensitivity
Different ISPs have different sensitivity to frequency-driven complaint accumulation, reflecting their different complaint processing architectures and reputation model update frequencies. Gmail updates domain reputation and spam rate data on a 24–48 hour lag, which means complaint accumulation from high-frequency sending becomes visible in Postmaster Tools within 2–3 days of the frequency increase. The reputation model's response — if the complaint rate increase is sustained — begins affecting inbox placement within 1–2 weeks.
Yahoo's complaint processing (through the CFL FBL) is near-real-time — complaint reports arrive within minutes of each complaint event. Yahoo's reputation model is more reactive to sudden complaint spikes, which can produce throttling (TS02 codes) within hours of a high-frequency send to a low-engagement segment generating unusual complaint volume. The response is faster than Gmail's but also recovers faster when the problematic sending stops — Yahoo's complaint rate calculation tends to be more recency-weighted than Gmail's broader 30-day window.
Microsoft's SNDS complaint data updates daily and its reputation model responds more slowly than Yahoo's. High-frequency sending to low-engagement segments typically produces SNDS IP status degradation over 1–2 weeks rather than immediately, giving slightly more time for detection and correction before delivery impact becomes severe.
Table 1 — ISP response characteristics to frequency-driven complaint rate increases
| ISP | Complaint data source | Detection lag | Reputation impact lag | Recovery speed after fix |
|---|---|---|---|---|
| Gmail | Postmaster Tools spam rate | 24–48 hours | 1–2 weeks | Slow: 3–8 weeks of clean sending |
| Yahoo | CFL FBL near-real-time | Minutes to hours | Hours to days (TS02 throttle) | Moderate: 1–3 weeks |
| Microsoft | SNDS daily update | 24 hours | 1–2 weeks | Moderate: 2–4 weeks |
Infrastructure Configuration for Frequency Management
The infrastructure components that support sustainable high-frequency sending are the same components discussed throughout this note series: real-time FBL complaint processing that suppressescomplainants before the next send, engagement-based list segmentation that routes high-frequency campaigns only to high-engagement segments, per-IP complaint rate monitoring that alerts before ISP thresholds are crossed, and pool isolation that ensures a high-frequency promotional campaign's complaint rate doesn't contaminate the transactional pool's reputation.
The specific configuration that high-frequency sending requires that weekly sending does not: more responsive complaint suppression (if campaigns are running daily, a complainant who is not suppressed immediately will receive additional sends the very next day, compounding the FBL complaint volume); more frequent engagement cohort re-classification (at daily cadence, a contact who engaged with Monday's campaign should be re-classified before Wednesday's campaign is segmented, not at the weekly batch re-classification run); and tighter monitoring thresholds (alerting at 0.05% complaint rate rather than 0.08%, to provide sufficient reaction time before the ISP's 0.10% threshold is reached given the faster complaint accumulation rate at high frequency).
High-frequency sending is not inherently problematic — many highly successful email programmes send daily to engaged lists with excellent ISP relationship outcomes. It is problematic when the frequency exceeds what the list's engagement quality can sustain, when complaint processing is not rapid enough to prevent complainants from receiving multiple sends, or when the monitoring resolution is insufficient to detect the faster complaint accumulation that high frequency produces. Infrastructure that correctly handles high-frequency sending addresses all three of these: real-time suppression, rapid engagement re-classification, and tightly calibrated alert thresholds.
The Frequency Increase Process: How to Test Before Full Deployment
Increasing sending frequency — moving from weekly to 3×/week, or from biweekly to weekly — is a decision that carries reputation risk if the list's engagement quality does not support the higher cadence. Testing the frequency increase on a subset of the list before full deployment provides data on the complaint rate impact before the change affects the full sending programme's reputation.
The A/B frequency test methodology: split the active list into two equal segments matched by engagement profile. For 4 weeks, send the control segment at the current frequency and the test segment at the proposed higher frequency. After 4 weeks, compare: complaint rate per segment (from FBL data), unsubscribe rate per segment, and ISP-segmented engagement rate per segment (to catch inbox placement differences that engagement rate changes may signal). If the test segment's complaint rate is more than 0.02 percentage points above the control segment's rate, or if unsubscribe rate is more than 50% higher, the list does not support the higher frequency without additional engagement filtering to restrict the higher cadence to only the highest-engagement contacts.
The test results also reveal which ISPs are most affected by the frequency increase. If the test segment shows elevated complaint rates at Yahoo but not at Gmail, the frequency ceiling is being approached specifically for Yahoo-destined mail — which may indicate that the Yahoo-specific audience demographic has different frequency tolerance than the Gmail audience, or that the Yahoo FBL's faster complaint processing is revealing complaints that Gmail's slower aggregation is smoothing out.
Frequency Segmentation: Not One Cadence for the Entire List
The most sophisticated approach to frequency management is not a single sending cadence applied uniformly to all contacts, but a variable cadence calibrated to each contact's engagement level. Highly engaged contacts receive daily sends; moderately engaged contacts receive weekly sends; lower-engagement contacts receive monthly sends. This variable-cadence model maximises revenue from high-engagement contacts (who can sustain and appreciate higher frequency) while protecting ISP reputation from the complaint accumulation that would result from applying daily sending cadence to the moderate and lower-engagement cohorts.
The implementation in MailWizz: campaign segmentation rules that exclude contacts based on their last engagement date. The daily campaign targets only the 30-day engaged cohort. The weekly campaign targets the 31–60 day cohort (and optionally the 30-day cohort if they haven't received the daily campaign already). The monthly campaign targets the 61–90 day cohort. Contacts beyond 90 days of no engagement do not receive any campaign sends until they self-identify through a re-engagement sequence.
The variable cadence approach requires more segmentation work per campaign cycle, but the deliverability benefit is measurable: complaint rates remain below ISP alert thresholds even as the high-frequency daily sends reach the engaged cohort, because the moderate and lower cohorts — where the frequency-driven complaints would otherwise originate — receive appropriately lower-frequency sends that match their tolerance level. The ISP sees a complaint rate for the domain that reflects the blended rate across all cadence tiers, which is lower than what a uniform daily send to all cohorts would produce.
The Revenue Calculation: Is Higher Frequency Worth It?
Higher sending frequency typically produces more total revenue — more opportunities to convert per time period — but at a cost in unsubscribe rate and complaint rate that has a longer-term revenue impact. The decision about optimal sending frequency is an economic calculation that must account for both dimensions.
The short-term revenue contribution of a frequency increase: if moving from weekly to 3×/week produces 3× the campaign sends per week, and each additional send generates 60% of the revenue of the first send in the same week (diminishing returns as the same audience receives multiple sends), the frequency increase produces approximately 2.2× the weekly email revenue. This is the commonly cited justification for higher frequency.
The long-term revenue cost of the frequency increase must be subtracted. Higher frequency increases unsubscribe rate — contacts who were comfortable with weekly sends but not with 3×/week sends. Each unsubscribe removes a contact permanently from the programme. The lifetime value of each unsubscribed contact (the expected revenue they would have generated at weekly frequency for their remaining engagement lifetime) is the opportunity cost of the frequency-driven unsubscribe. This calculation typically shows that the short-term revenue gain from higher frequency is partially or fully offset by the accelerated list attrition it drives, particularly for the moderate-engagement cohort where frequency sensitivity is highest.
The optimal frequency for a given list segment is the frequency that maximises the net present value of future email revenue from that segment — accounting for both the campaign revenue per send and the expected reduction in segment size from frequency-driven attrition. This optimum varies by segment: for high-engagement contacts, the optimum may be daily; for moderate-engagement, weekly; for lower-engagement, monthly. The variable-cadence model described above implements this optimisation by applying different frequencies to different engagement tiers.
Communicating Expected Frequency at Opt-In
Frequency management begins at list acquisition — setting correct expectations at the point of opt-in dramatically reduces frequency-driven complaints from new subscribers. A contact who opted in expecting weekly emails and then receives daily emails is much more likely to mark the daily sends as spam than a contact who opted in explicitly acknowledging daily sends. The opt-in communication ("Get daily deals delivered to your inbox" vs "Join our weekly newsletter") sets the frequency expectation that determines how recipients interpret high-frequency sending.
Contacts whose opt-in explicitly acknowledged high frequency have lower complaint rates at that frequency because the sends match their expectations. This expectation-setting effect is a list quality input that infrastructure systems can act on: tracking opt-in language or opt-in source per contact and using this data to inform both the frequency applied to each contact and the engagement model that determines their threshold for frequency-driven complaints.
The infrastructure implication: the contact database should store opt-in context including the expected frequency communicated at signup. This field is then available as a segmentation criterion in MailWizz, allowing the campaign manager to apply different frequency cadences to contacts who signed up for different frequency promises, ensuring that the sends received by each contact are consistent with what they were told to expect. This reduces complaint rates not through engagement filtering alone but through fundamental expectation management at acquisition — addressing the root cause of frequency-driven complaints rather than just managing the consequence.
Daily Monitoring Requirements for High-Frequency Programmes
High-frequency sending requires daily monitoring of metrics that weekly-sending programmes can review less often. The daily monitoring cadence for a programme sending 3×/week or more:
Morning check (before first send): Gmail Postmaster Tools spam rate for the past 24–48 hours — any spike that corresponds to yesterday's sends? Yahoo FBL complaint count from the previous day's sends. DNSBL status for all sending IPs. If any of these show anomalies, investigate before sending the day's first campaign. At high frequency, an elevated complaint rate from yesterday's campaign will be compounded by today's campaign if not investigated first.
Post-campaign check (within 4 hours of each send completion): per-ISP delivery rate and deferral rate from the accounting log. FBL complaint count attributed to the just-completed send. Any hard bounces above the programme's normal rate. This check catches campaign-specific problems — a list segment that performed differently than expected, a content change that drove unusual complaint rates — while there is still time to adjust before the next send.
Weekly review: 7-day trend for Gmail Postmaster Tools domain reputation (not just the current tier, but the direction of the trend over the week). Per-segment unsubscribe rates versus baseline. Cohort engagement distribution (is the high-engagement cohort growing or shrinking as a proportion of the active list?). These weekly metrics are the leading indicators that a frequency increase is approaching the sustainable ceiling — declining domain reputation trend, accelerating unsubscribe rates, and shrinking high-engagement cohort are the signals that warrant a frequency reduction before complaint rates reach ISP alert thresholds.
The monitoring discipline for high-frequency programmes is not optional — it is what makes high-frequency sustainable. A weekly sender who misses a week of Postmaster Tools review loses a week of early warning. A daily sender who misses a day of monitoring may miss the complaint spike from the previous day's campaign that would have prompted segment adjustment before today's campaign compounds the problem. The monitoring cadence must match the sending cadence to provide the early warning that high-frequency reputation risk requires.
Programmes that implement the variable-cadence model, real-time FBL processing, tightened monitoring thresholds, daily pre-send and post-campaign checks, and opt-in frequency expectation setting consistently sustain high-frequency sending without the reputation deterioration that makes high frequency unsustainable for programmes without these disciplines. High-frequency email is not a deliverability risk — high-frequency email without the infrastructure and practices that make it sustainable is a deliverability risk. The distinction is entirely in the operational disciplines that determine whether complaint rate stays below ISP thresholds as the frequency increases beyond what simple, undifferentiated bulk sending can accommodate.
Building the Business Case for Frequency Discipline
Email marketing teams that want to maximise sending frequency sometimes face resistance from deliverability and infrastructure teams who understand the reputation risk. The correct framing for this conversation: the deliverability concern is not about frequency per se, but about frequency relative to list quality. The business case for frequency discipline is not "send less" — it is "send high-frequency to the contacts who can sustain it, and lower frequency to those who cannot, to maximise revenue from the programme while protecting the infrastructure's reputation capacity that makes all sending — at any frequency — possible."
The loss scenario that makes the case concretely: a programme that increases sending frequency to the entire list without engagement segmentation may see a 20% short-term revenue increase from more sends per week. If this frequency increase drives Gmail domain reputation from High to Medium over 6 weeks — which the elevated complaint rate from the moderate-engagement cohort produces — the resulting inbox placement decline (from ~92% to ~75% inbox at Gmail) reduces revenue from all campaigns, not just the campaigns to the moderate-engagement segment. The 20% revenue gain from higher frequency is erased by the 17-point inbox placement decline, which across the full programme (including the high-engagement segment that was sending correctly) represents a net revenue loss that the frequency increase caused but the entire programme bears.
This downstream reputation cost — where one segment's frequency-driven complaints reduce inbox placement for a different, well-managed segment through the shared domain reputation — is the systemic risk that the variable-cadence, engagement-segmented model avoids. By isolating the high-frequency sends to only the contacts whose engagement quality supports it, the shared domain reputation is protected by the behaviour of the entire programme's sends, not degraded by the most aggressive part of it. This is the business case for frequency discipline that resonates with business stakeholders who understand revenue risk: protecting the domain reputation that makes the high-frequency revenue possible is worth the operational complexity of variable-cadence segmentation.
The infrastructure, monitoring, and sending practice disciplines described in this note are the operational foundation that makes high-frequency sending sustainable rather than self-destructive. They are not constraints on revenue potential — they are the enablers of it. Programmes that invest in real-time FBL processing, engagement-based segmentation, daily monitoring, and variable-cadence delivery consistently generate more long-term revenue from their email channel than programmes that optimise for maximum short-term frequency at the expense of domain reputation quality. The reputation capacity that ISPs extend to consistently clean, well-managed senders is the asset that high-frequency revenue extraction depends on — and the discipline required to maintain it is the investment that makes the extraction sustainable over months and years rather than quarters.
Infrastructure Assessment
Our managed infrastructure includes real-time FBL processing, rapid engagement re-classification, tightened monitoring thresholds for high-frequency programmes, and per-ISP complaint rate alerting calibrated to the programme's sending cadence. Request assessment →