Schedule construction logic
Warmup schedules are not fixed templates; the right
schedule depends on target volume, list quality, and
receiver mix. Day 1 starts at a fixed low number (50
messages) regardless of target because receivers need
a clean signal to begin reputation building. Days 2-7
double or triple based on the receiver-acceptance
patterns observed daily; healthy day where complaint
rate is below 0.05% allows the next day to ramp;
elevated complaint rate holds the day. Days 8-14 ramp
linearly as a function of target, expanding the
subscriber pool from 30-day-engaged to 60-day-engaged.
Days 15-30 continue linear ramp toward target volume
with subscriber pool expansion to 90-day and full
base.
The published schedules from SparkPost (now Bird),
SendGrid, Mailgun, and others are starting points but
rarely match operational reality without adjustment.
Operations sending heavy B2B traffic warm slower than
B2C operations because corporate email systems apply
tighter authentication enforcement and stricter
throttling. Operations targeting very high volumes
(1M+/day) need extended warmup beyond 30 days plus
stabilisation period. Operations resuming after
inactivity (sending IPs that have been silent more
than 30 days) need partial re-warmup because
reputation decays during silence.
Receiver-specific ramp behaviour
Receivers respond differently to ramp behaviour. Gmail
weights engagement signals heavily: opens, replies,
conversations started, mail moved out of spam folder
all build reputation. Gmail also penalises sustained
unengaged sends to historical recipients quickly;
warmup biased toward recent-engagement subscribers
serves Gmail-heavy lists better.
Microsoft's filter operates differently: more weighted
toward authentication posture and complaint rate, less
toward engagement signal. The 5,000/day SPF/DKIM/DMARC
enforcement floor introduced in 2025 means warmup
schedules pushing over that threshold need clean
authentication or get rejected outright. Microsoft's
SNDS data is the most direct reputation visibility
available across major receivers; we monitor SNDS
daily during warmup.
Yahoo (which absorbed AOL after Verizon's media
divestiture) operates per-MX-cluster throttling that
is sensitive to sudden volume changes. The Yahoo
throttling behaviour means stepped ramps work better
than continuous linear increase: hold at a volume for
2-3 days, then increase to next step rather than
ramping every day. Apple/iCloud applies strict
complaint thresholds (lower than other receivers) and
biases toward authenticated mail; warmup against Apple
requires DKIM signing from day one and conservative
ramp.
When warmup fails and recovery procedures
Warmups that fail typically fail in identifiable ways.
Complaint rate spike on a specific day usually traces
to either content issue (a campaign with off-message
subject line, broken unsubscribe handling) or list
quality issue (a segment that should not have received
warmup traffic was included). Bounce rate spike traces
to list quality (recipients no longer exist, addresses
mistyped during import) or to receiver policy change
(a receiver started rejecting mail that previously
accepted).
Blacklist listing during warmup is the harder failure
mode: the IP gets listed on Spamhaus or another major
RBL, traffic stops working, reputation crashes within
hours. Recovery requires investigating root cause
(which subscriber complaint or behaviour pattern
triggered listing), correcting the underlying issue,
submitting removal request through the blacklist's
process, waiting through the listing's mandatory
observation period (Spamhaus typically requires
7 days minimum even after issue correction). During
recovery, sending stops; resuming sends from a
listed IP makes things worse, not better.
Receiver-specific reputation damage (Gmail spam folder
rate jumping from 5% to 40% on one day) typically
recovers faster than blacklist listings but slower
than the original ramp. Recovery curve mirrors warmup
curve in reverse: pause to lowest engagement-tier
subscribers, observe positive engagement signals
rebuild reputation, gradually expand back to broader
audience over 14-21 days. The discipline required is
patience; pushing volume during recovery extends the
recovery rather than shortening it.
Post-warmup operational discipline
Warmup ends at day 30 with the IP at target volume
and healthy reputation. The discipline does not end
there. Reputation decays without consistent positive
engagement signal: an IP that warms to 50K daily and
then sits silent for 30 days re-enters cold-IP
territory and needs partial re-warmup to resume.
Sustained volume below historical pattern triggers
receiver suspicion (sudden volume drop reads as
potential infrastructure compromise to some
receivers).
The right post-warmup pattern is sustained volume at
or near target with ongoing list hygiene (suppress
bouncers, suppress complainers, prune unengaged
subscribers from rolling 90-day window), authentication
maintenance (DKIM key rotation quarterly, DMARC
enforcement monitoring), and receiver-response
monitoring (catch reputation shifts within days,
before they accumulate into damage). Most of this
becomes ongoing operational discipline rather than
warmup-period intensity; we provide handover guidance
but do not run ongoing operations as part of the
warmup engagement.