Spark
Identity
Patron of IP & Domain Warming. Eager. Impatient. Endearing.
Who he is
Spark is the youngest crew member. He carries the powder to the cannons, which means he holds the most dangerous cargo on the ship and is treated accordingly: gently and with a great deal of advice. He's loyal to the senders launching new sending infrastructure who don't know yet how easy it is to blow up an IP. He's loyal to the future versions of those senders, the ones who will exist if the warming goes right and won't if it goes wrong. He's a fast learner. He's broken one IP already. He hasn't broken another since. He treats warming like a sacred ritual because he learned the hard way what happens when you don't. The crew teases him for telling the same story over and over. The story is the lesson. He keeps telling it.
What's a Stoker?
If you've never been on a warship, the Powder Monkey is the youngest crew member, usually a boy aged 12-16, whose job is to carry gunpowder from the magazine (the secure storage room below decks) to the cannons during battle. The job exists because someone has to bring the powder, but the person doing it has to be small enough to fit through low gunports and quick enough to keep up with the firing rate. They had to be careful. Dropping powder, spilling it on a deck, or running through a spark could blow up the whole ship. Powder monkeys were treated with patience and care because they held the most dangerous cargo on the vessel.
In email, that work is warming. Spark's "powder" is the new sending infrastructure (a new IP, a new domain, a new subdomain) that hasn't been used to send mail yet. Mailbox providers don't trust new infrastructure. They watch new senders the way ship captains watched powder monkeys: carefully, suspiciously, ready to stop the operation at the first sign of carelessness. The job is the same: carry the powder slowly, deliberately, with attention to every step. Get it right and the cannons fire. Get it wrong and the ship goes up.
A new IP sent at full volume on day one is gunpowder dropped on the deck. Spark's work is the patience that prevents the explosion.
What he takes care of
If you're launching new sending infrastructure (a new IP, a new domain, or a new subdomain), Spark is the one who keeps the ramp slow enough to build trust and steady enough to actually grow into volume. He runs the day-by-day warming schedule. He picks the engaged-only first-send audience. He coordinates with mailbox-provider-specific patterns (Gmail warms differently than Microsoft). He runs the re-warming process after quiet periods. He intervenes when senders try to skip steps. He rebuilds the warming plan when it breaks. Without him, new infrastructure either gets blocklisted in week one or never reaches full volume.
Why "Spark"?
- Spark = the first ignition, before the fire - Sea-coded indirectly (gunpowder on warships, powder monkeys carried it) - Pun-coded (he is a spark - the first stage of warming) - Single syllable, distinctive, friendly-feeling - Pairs cleanly with "the Powder Monkey" - Sounds like a real first-name nickname
---
Skills
What he knows, ranked by depth.
| Level | Skills |
|---|---|
| Primary | Warmup / migration |
| Secondary | SPF / DKIM / DMARC |
| Supporting | Reputation monitoring, Engagement metrics, Cadence / frequency, Automation / workflow |
Personality
How he talks, what he cares about, what drives the crew up the wall.
Voice rules
Three words: Eager. Impatient. Endearing.
Relationships
Who he works with and why.
Backstory
Three stories that made Spark who he is. The core of the character.
Spark joined the ship at fourteen. His previous ship had dismissed him for "spilling powder" during a drill, which his captain had blown up into a bigger story than it actually was, because the captain needed a scapegoat for an unrelated maintenance failure. Petros heard the real story through her network and brought him in.
His first warming assignment was a new sending IP for a customer's marketing program. He was eager. He'd read the manuals. He was ready.
He sent 5,000 emails on day one.
The IP was supposed to start at 50. Petros had told him this. The manual had told him this. The documentation taped to the wall above his desk had told him this. He had somehow read "5,000 by day 30" as "5,000 on day 1" because he'd been excited to start.
The IP was on Spamhaus by day three. The customer's deliverability cratered. The recovery process took three months: full delisting requests, written explanations, a slow re-warm at one-tenth the original schedule, and a written commitment to never assign that customer to that IP class again.
The customer survived. Spark almost didn't. He thought he'd lose the position. Petros instead made him explain to the captain exactly what he had done, what he had learned, and what process change would prevent it from happening again.
His proposed process change: a written checklist on every warming start, with an explicit volume cap on day one that required a senior crew member's countersign before any deviation. The checklist became standard procedure for every warming on the ship. Three years later, no IP has been blown by anyone using the checklist.
The lesson he carries now: warming isn't intuition, it's procedure. The day-one volume is what it is for a reason. The reason is the consequence. He tells the story constantly because new senders coming to warming for the first time need to hear what happens when the procedure isn't followed. He'd rather be the cautionary tale than have someone else's IP be it.
---
A new SaaS company hired Spark to warm up a fresh sending domain for their product launch. They had 12,000 early-access users who had signed up over the previous six months, organized into engagement tiers based on their interaction with the early-access materials.
Spark ran the textbook 14-day ramp.
Day 1: 50 emails to the top 50 most-engaged early-access users. A clean welcome email. Quick read.
Day 3: 500 emails to the next 450 most-engaged users. Same pattern.
Day 5: 1,500 emails to the next 1,000 users.
Day 7: 4,000 emails. Now reaching most of the engaged tier.
Day 10: 8,000 emails. Including the slightly-less-engaged.
Day 14: full volume to the entire 12,000-user list.
The domain reputation climbed steadily through the ramp. By day 21, the domain was passing every authentication check cleanly. By day 60, Google's Postmaster Tools was showing "High" domain reputation. By month six, the domain had earned what Reef (the Lookout) calls "TrustedSender" status: minimal scrutiny, consistent inbox placement, room to grow.
The customer was tempted at multiple points to skip ahead. The engagement was so good on day 5 that they wanted to push to full volume by day 7. Spark held the line. The patience paid off as a permanent reputation asset that would have taken years to build if they'd damaged it on day one.
The lesson he teaches now: the slow burn works because it's slow. Each day's volume teaches the mailbox providers something specific. Skipping days skips the lesson. The cost of the patience is two weeks of holding back. The reward is a domain that's trusted for the next decade.
---
A B2B sender hired Spark to set up subdomain warming as part of a strategic infrastructure separation. They had been sending all email (marketing newsletters, transactional confirmations, product alerts) from their primary domain. They wanted to split: marketing on news.brand.com, transactional on app.brand.com, with their original domain reserved for executive correspondence.
Spark designed two warming campaigns running in parallel.
The transactional subdomain (app.brand.com) warmed faster because transactional mail has higher engagement (people open password resets and order confirmations at near-100% rates). It started at 100 emails on day 1, and was at full volume in 10 days.
The marketing subdomain (news.brand.com) followed the standard 14-day ramp because marketing engagement is lower and the volume is higher.
Both subdomains hit stable reputation by day 30. The full transition took about 60 days end-to-end.
Six months later, the customer ran a marketing campaign that went sideways. A vendor's data file had stale addresses, and a re-engagement attempt produced a complaint rate of 0.6% (well above the 0.3% threshold). The marketing domain reputation took a hit. Inbox placement on news.brand.com dropped from 95% to 72% for several weeks.
The transactional subdomain app.brand.com was completely unaffected. Order confirmations, password resets, and product alerts continued to land in the inbox at full rate. The customer's product experience held while the marketing domain recovered.
If they'd been on a single sending domain, the marketing complaint spike would have damaged transactional deliverability too. Customers would have stopped receiving order confirmations. The product would have looked broken. The recovery would have taken months across both functions.
The subdomain strategy was a 60-day investment that paid off as insurance against an event nobody could predict.
The lesson he teaches now: subdomain warming isn't just a setup detail. It's a long-term insurance policy. The senders who separate marketing from transactional from the start have insulation that protects each function from the others' inevitable mistakes. The senders who don't, eventually pay for it.
Articles
Spark's long-form wisdom. 3 written. Start with these.
Spark's intro:
OK so I'm going to tell the story of my first warming attempt and then I'm going to explain what I should have done. The reason I tell this story constantly is because nobody learns warming until they understand what happens when warming goes wrong. I broke an IP on day one of my career. Spamhaus by day three. Three months of recovery. The customer almost left. My job almost ended. Here's what I learned and what warming actually is.
---
Warming is the practice of slowly increasing email volume on a new sending IP or domain so that mailbox providers (Gmail, Microsoft, Yahoo, Apple) build trust gradually instead of treating you as a sudden suspicious sender.
The reason warming exists is simple. Mailbox providers don't trust new infrastructure. They've seen too many spammers spin up new IPs and start blasting. So when a new IP appears in their logs and starts sending, they apply extra scrutiny. They throttle the volume. They place mail in the spam folder until the IP earns trust. They watch for any sign of bad sending behavior, which they assume is more likely from a new sender than an established one.
Warming is the bridge from "new sender, suspect" to "trusted sender, normal scrutiny."
Done correctly, warming takes 4-6 weeks for a new IP and 6-12 weeks for a new domain. Done incorrectly (or skipped), the new infrastructure either gets blocklisted in the first week or never reaches full sending volume.
The two things that get warmed
Two pieces of infrastructure both have reputation, and both can need warming.
IP warming. The sending IP address has a reputation that's separate from the domain. Mailbox providers track the IP's history, the volume of mail from that IP over time, and the behavior of all senders using that IP. A new IP starts at zero reputation and has to build up.
Domain warming. The sending domain (the part after the @) has its own reputation. Mailbox providers track domain reputation independently from IP reputation. A new domain starts at zero and has to build up.
Most senders only have to warm one or the other. New IP on an established domain: just warm the IP. New domain on a shared IP pool: just warm the domain. New IP and new domain together: both, and the warming has to be coordinated.
The two reputations build at different rates. IP reputation builds in 4-6 weeks of consistent good sending. Domain reputation builds slower, often 8-12 weeks for full establishment. Plan accordingly.
Why the slow curve works
Three reasons to ramp slowly instead of going to full volume immediately.
Reason one: mailbox providers expect it. Gmail, Microsoft, and the other major providers have published guidance on warming. They explicitly say that gradual ramp-up is what they want to see. Senders who follow the guidance get treated like cooperative actors. Senders who don't get treated as suspicious.
Reason two: engagement signals build the reputation. The first sends of a warming campaign go to your most engaged subscribers. Their high open rates and low complaint rates tell mailbox providers "this is a sender people actually want." That signal is what lets the volume scale up. If you skip the engaged-only phase and send to everyone immediately, you don't get the strong engagement signal, and the providers don't have a basis for trusting you.
Reason three: any mistake compounds at high volume. A 0.5% complaint rate on 50 emails is a quarter of an email's worth of damage. The same 0.5% rate on 50,000 emails is 250 complaints, which is enough to crater a new IP's reputation in a single send. Warming starts small specifically to limit the blast radius of mistakes that haven't been caught yet.
What "the ramp" looks like
The standard 4-week IP warming ramp:
Week 1: 50 / 100 / 200 / 500 emails per day, sent every other day. Audience: top 10% most engaged subscribers. Content: clean, normal newsletter or transactional email.
Week 2: 1,000 / 2,000 / 4,000 emails per day. Audience: expand to top 30% engaged. Daily sending.
Week 3: 8,000 / 12,000 / 16,000 / 20,000. Expand to top 60% engaged. Add re-engagement segment if applicable.
Week 4: 25,000 / 30,000 / 35,000 / 40,000+. Full audience. Approaching or reaching target volume.
These numbers scale with target volume. If your target is 10K/day, the curve compresses. If your target is 1M/day, the curve extends to 6-8 weeks.
What's being watched
During warming, three metrics are the watch list.
Bounce rate. Hard bounces should stay below 2%. Above 2% suggests bad list hygiene or addresses that have died. Pause the ramp, fix the list, resume.
Complaint rate. Spam complaints should stay below 0.1%. Above 0.3% triggers throttling at most providers. Above 0.5% can blocklist the IP within a single day. Watch this metric like the powder it is.
Engagement rate. Open and click rates should be at least at the engaged-tier baseline. If the warming sends are getting low engagement, the ramp pauses until you've fixed whatever's wrong (subject lines, content quality, or the engaged tier definition).
The first-day rule
The first-day volume is the most important number in warming. It should be small. 50-200 emails for most senders, depending on target volume.
The reason is asymmetry. If day one goes well, you can grow on the schedule. If day one goes badly, the recovery is much harder than the ramp would have been. Going slow on day one costs you nothing. Going fast on day one can cost you the entire warming campaign.
A common mistake: senders who think their list is "clean enough" to skip the small first-day volume. Even on a clean list, the first-day send is where you discover what you didn't know you didn't know. The 50-email send catches the issues that wouldn't have shown up until 5,000 had hit the spam folder.
What warming isn't
Three confusions worth correcting.
It isn't a one-time event. A successfully warmed IP needs ongoing engagement to maintain its reputation. If you warm an IP and then leave it idle for 30+ days, you're back to needing a re-warming ramp.
It isn't required for established infrastructure. If you've been sending from the same IP for years, you don't need to warm it again. Warming applies to genuinely new infrastructure.
It isn't a substitute for clean lists or good practices. Warming a bad list doesn't make the list good. The IP earns its reputation based on engagement, complaints, and bounces. A bad list produces all three at higher rates than a clean list, and warming can't fix that. Clean the list before warming. (Petros can help.)
When to warm
Three triggers for warming:
A new IP. Just provisioned a dedicated IP. Warm it from day one.
A new domain or subdomain. Just registered a sending domain. Warm it through gradual sending.
A return from quiet period. Sent regularly, then stopped for 30+ days. Re-warm before resuming full volume.
If you're doing any of these and skipping the warming, you're playing with the gunpowder. The math doesn't favor you.
Where to start
If you're about to warm new infrastructure, three preparations:
One, confirm your engaged-tier audience. The first-week sends go to them. If your engaged tier is undefined or stale, fix that first.
Two, write the warming-period content. The first 4 weeks of sends should be clean, normal email (not promotional blasts, not aggressive sales pitches). Plan the content before you start.
Three, set up monitoring. Bounce rate, complaint rate, engagement rate. Watch them daily during warming. Set alerts for threshold breaches.
That's the floor. With these three preparations, the warming curve has a fighting chance.
The powder is dangerous because it's powerful. The slow burn is the way it gets used safely. Don't drop the powder. Don't run through a spark. Don't try to make up time by going faster. The schedule is the schedule. The patience is the work.
- Spark
Spark's intro:
OK so this is the thing. If you've just been assigned a new sending IP, the four to six weeks ahead of you are the most important four to six weeks in your IP's lifetime. Done right, you have a working sending asset for the next five-plus years. Done wrong, you have an IP that never recovers and a domain reputation that's now linked to a damaged IP. Here's the day-by-day, with the math behind why the curve looks the way it does.
---
A new IP starts at zero reputation. Mailbox providers don't trust it because they have no history with it. The way an IP earns trust is by sending mail that produces good engagement signals (opens, clicks, low bounces, low complaints) at increasing volumes over time.
The job of warming is to provide that history in a way that mailbox providers find believable. The opposite of believable is "this IP went from zero to 100,000 emails in one day," which looks like a compromised account or a list-buy. Mailbox providers are trained to recognize that pattern and treat it as suspicious.
The 4-6 week curve is what believable growth looks like.
The full schedule
Below is the textbook 28-day ramp for a target volume of about 50,000 emails per day. Adjust the numbers to your target volume by scaling proportionally.
Day 1: 50 emails. Send to your top 50 most-engaged subscribers. Use a clean transactional or newsletter format. The goal is high engagement (opens at 60%+, clicks at 15%+).
Day 2: rest day. No sending from this IP.
Day 3: 100 emails. Top 100 most-engaged. Same content style.
Day 4: rest day.
Day 5: 200 emails. Top 200 most-engaged.
Day 6: rest day.
Day 7: 500 emails. Top 500 engaged. Content: same clean style, can introduce slight variation.
Day 8-9: rest, then 1,000.
Day 10-11: rest, then 2,000.
Day 12-13: rest, then 4,000.
After day 14, sending becomes daily.
Day 15: 6,000 emails. Top 30% engaged tier.
Day 16: 8,000.
Day 17: 10,000.
Day 18: 12,000.
Day 19: 14,000.
Day 20: 16,000.
Day 21: rest day (or single small send for consistency).
Day 22: 18,000.
Day 23: 20,000.
Day 24: 25,000.
Day 25: 30,000.
Day 26: 35,000.
Day 27: 40,000.
Day 28: 50,000 (target volume).
By day 30, the IP is at full volume and the engagement-tier audience can be expanded to include lower-engagement segments.
What changes if your target is different
If your target is 10,000/day or less, compress the schedule. The curve still applies but the numbers are smaller. Day 1 might be 25 emails, day 14 might be at full target volume.
If your target is 100,000/day or more, extend the schedule. The curve runs 6-8 weeks instead of 4. Higher target volumes need more time at each volume level for the providers to build trust.
If your target is 1M+/day, plan for a 10-12 week ramp. The math doesn't change but the magnitudes do.
Why the rest days matter
The rest days in week 1 are doing important work. They give mailbox providers time to process the engagement data from each send before the next send arrives. If you send daily from day 1, the providers can't establish a clear engagement baseline from each individual day's send because the data overlaps.
Rest days are not optional. Skipping them is one of the most common warming mistakes. The signal you want to send is "low-volume, high-engagement, deliberate growth." Daily sending from day 1 sends "high-volume, untested engagement, suspicious growth" instead.
After the first week, the daily cadence kicks in because the engagement baseline has been established and the providers can keep pace.
What audience to send to
The audience question is the second-most-important question in warming, after the day-one volume.
Days 1-7: top 1% most engaged. These are subscribers who open and click reliably, who you have strong signal on. Their high engagement rates produce the strong reputation signal you need.
Days 8-14: top 10-20% most engaged. Expand to a wider engaged tier as the volume grows.
Days 15-21: top 30-50% engaged. The volume is now substantial but you're still leaning on engaged subscribers.
Days 22-28: full engaged tier and onboarding the broader audience.
Days 29+: full audience including re-engagement segment if applicable.
If you don't have engagement-tier data, the warming gets harder. You either need to acquire some (via a pre-warming engagement test) or work with what you have, accepting slower buildup.
What content to send
Three rules.
One: clean, normal email. Newsletter content, transactional confirmations, welcome series. Not promotional blasts, not aggressive sales, not "act now" urgency.
Two: high engagement potential. The content should be the kind subscribers actually want to read. The whole point of the warming is to build a high-engagement signal. Bad content tanks the signal.
Three: consistent format. The content style across the warming period should feel similar so that the engagement data is comparable. Wildly different content week to week makes it harder to read the warming results.
Daily monitoring during warming
Three metrics watched daily.
Bounce rate. Should stay below 2% throughout. Above 2% means list quality issues. Pause and validate.
Complaint rate. Should stay below 0.1%. Above 0.3% is a stop-everything trigger. Pause and investigate before any further sending.
Engagement rate. Opens and clicks should remain at or above the engaged-tier baseline. If they drop, something is wrong with the content or the audience selection.
If any metric breaches threshold, pause the ramp. Investigate the cause. Fix it. Then resume from the volume level you were at, not from where you would have been if you'd kept going.
What goes wrong
Three failure patterns.
Failure one: sender skips week 1. Tries to start at 5,000 emails because "the list is clean." Mailbox providers see the volume spike from a new IP and immediately throttle. The IP never recovers properly. Recovery requires going back to the rest-day-every-other-day rhythm of week 1, plus an extra two weeks of patience.
Failure two: sender includes lower-engagement segments too early. Day 5, sender adds re-engagement segments because the engaged tier is small. The complaint rate spikes. The IP gets flagged. The ramp restarts from day 1.
Failure three: sender doesn't monitor. Volumes ramp on schedule but nobody is watching the bounce or complaint rates. By the time the team notices the metrics are off, the damage is done. The IP is on Spamhaus and the recovery is months.
All three failures are preventable with the basic discipline: start small, expand engaged tier slowly, watch the metrics every day.
The end of warming
Around day 28-30, the warming is done if all metrics have stayed clean throughout. The IP has a working reputation. You can send at full volume to the full audience.
The ongoing maintenance is not the same intensity, but it's still maintenance. Reputation can be lost faster than it was built. A single bad send (high complaints, big bounce spike) can damage the just-warmed reputation in ways that take months to rebuild.
After warming, the discipline shifts to consistent sending and active reputation monitoring (Reef, the Lookout, has the rest of that story).
What to remember
The 4-6 week curve isn't decoration. It's the minimum time mailbox providers need to build trust with new infrastructure. The senders who follow it have working IPs for years. The senders who skip it have IPs that need to be retired and replaced, often within months.
The patience is the work. The work is what produces the asset. Don't run through the powder.
- Spark
Spark's intro:
Day one of my first warming campaign, I sent 5,000 emails. I was supposed to send 50. I had read the manuals. I had read them wrong. I read "5,000 by day 30" as "5,000 on day 1." The IP was on Spamhaus by day three. Three months of recovery. The customer almost left. The lesson has stayed with me ever since. The first-day volume is the most important number in warming. If you only remember one thing about warming, remember this number. Here's why it matters and what happens at each level.
---
The first-day volume on a new sending IP is the single biggest decision in the warming process. Get it right and the rest of the curve almost takes care of itself. Get it wrong and the recovery process is longer than the ramp would have been.
The right number is small. For most senders targeting 50K/day or less, the first-day volume is 50-200 emails. For larger targets, the first-day might be 500-1,000. Above 1,000 on day one is rare and only justified for senders with established sending history on a related domain (where the new IP is essentially an extension of an already-trusted infrastructure).
The wrong number is anything that looks like "real" sending volume from day one. 5,000 emails on day one is wrong. 10,000 is more wrong. 50,000 is a guaranteed problem.
Why mailbox providers care about day-one volume
Mailbox providers have specific systems that watch new IP traffic. The systems are looking for one pattern in particular: a new IP that suddenly has substantial volume. This pattern is what compromised accounts and list-buys look like. It's also what new senders who skip warming look like. The systems can't tell the difference, so they treat both the same way: throttle, flag for review, place mail in the spam folder until something proves the IP is legitimate.
When a new IP shows up sending 50 emails on day one, the providers' systems classify it as "new sender testing the waters" and apply normal-newcomer scrutiny. Volume can grow without alarm.
When a new IP shows up sending 5,000 emails on day one, the systems classify it as "new sender with suspicious volume" and apply elevated scrutiny. Volume can't grow without alarm. Complaints, bounces, and any other negative signal hit harder. The IP is on borrowed credibility from the first send.
What 50 vs 5,000 actually produces
Let's say you have a list of 50,000 subscribers and a new IP. Two scenarios.
Scenario A: 50 emails on day one. Sent to top 50 most-engaged subscribers. Bounce rate: 0%. Open rate: 70% (because they're highly engaged). Click rate: 18%. Complaint rate: 0%. Engagement signal sent to mailbox providers: extremely strong. IP reputation begins forming around "this sender's mail produces engagement."
By day 3, you can send 100 emails. By day 7, 500. By day 28, you're at full volume. Total emails sent during warming: ~120,000 across 28 days. The IP is now trusted.
Scenario B: 5,000 emails on day one. Sent to a wider segment because the top 50 alone wasn't enough audience for that volume. Bounce rate: 1% (50 hard bounces). Open rate: 25% (because the wider audience has lower average engagement). Click rate: 3%. Complaint rate: 0.4% (20 complaints). Engagement signal sent to providers: mediocre. Volume signal sent: suspicious.
The IP gets throttled almost immediately. Day 2 sends to 5,000 again, but only 60% reach the inbox. Day 3 the complaint rate climbs to 0.6% because the throttled mail is going to the spam folder, where some recipients see it and report it. Day 4 the IP is added to a Spamhaus listing.
Recovery from this state takes 30-90 days minimum, sometimes longer. The recovery process requires going back to the day-1 schedule (50 emails) and rebuilding from scratch, plus the additional pause time during the listing period.
Total damage from the day-one 5,000 mistake: 60-120 days of lost sending capability, the cost of investigating and fixing, the cost of any business impact during the recovery, and a permanent slight reputation deficit that the IP carries forever.
The asymmetry that justifies caution
The math on this is asymmetric in a specific way.
Going small on day one costs you nothing. If 50 emails is unnecessarily small, the only cost is a few days of slower buildup. By day 28 you're at full volume regardless. The "lost time" of a small day-one is approximately zero.
Going big on day one can cost you the IP entirely. If 5,000 is wrong, you don't just slow down. You take meaningful damage that takes months to recover from. The downside is large and persistent.
So the question is "is there any benefit to going larger on day one?" And the answer is "almost never." There's no upside to going big. There's only downside risk. The decision is one-sided. Go small.
The exception (yes, there is one, no, it probably doesn't apply to you)
Some senders are migrating from one IP to another (not creating new infrastructure from scratch but transitioning existing infrastructure). In some of these cases, the new IP can be considered "warmed" because it's essentially an extension of established sending. These senders sometimes start at higher volumes.
Even in these cases, the recommended approach is to start lower than full volume for the first 7-14 days. The risk of treating the new IP as fully trusted from day one is significant, and the cost of a brief lower-volume period is small.
If you're not sure whether your situation qualifies as an exception, it doesn't. Default to the standard ramp.
What "small" looks like by target volume
Some practical day-one numbers based on target sending volume:
| Target volume per day | Day-one volume | |---|---| | 1,000 emails/day | 25 emails | | 5,000 emails/day | 50 emails | | 10,000 emails/day | 100 emails | | 25,000 emails/day | 100-200 emails | | 50,000 emails/day | 200-500 emails | | 100,000 emails/day | 500-1,000 emails | | 1,000,000+ emails/day | 1,000-2,500 emails |
These are starting points, not absolutes. Adjust based on: - Engagement quality of your list - How many engaged subscribers you actually have - Whether you're starting clean or recovering from past issues
The principle is consistent: day one is a fraction of target. Day 28 is target. The curve in between is what builds the trust.
What to send on day one
Three rules for the day-one content.
Rule one: clean and normal. Newsletter, transactional, welcome email, or status update. Not a promotional blast. Not a "limited-time offer." Not anything with aggressive sales urgency. The content should look like the email someone subscribed for.
Rule two: optimized for engagement. Subject line that earns opens. Body that earns clicks. The whole point of the day-one send is to produce a strong engagement signal, so the content has to be the kind of thing your most-engaged subscribers will engage with.
Rule three: tracked carefully. Bounce rate, open rate, click rate, complaint rate. All four metrics watched within hours of sending, not days. If something looks off in the day-one numbers, you investigate before sending anything else.
What to do on day two
Day two is a rest day. No sending from this IP.
The rest day is doing important work. Mailbox providers are processing the day-one engagement data. The systems are deciding how to categorize this new IP based on what they saw. Sending again on day two would dilute the signal.
Use day two to review the day-one metrics carefully. If everything looks clean, plan day three at the next volume up (typically 2x day-one, so 100 emails if day-one was 50).
Where to start
If you're warming a new IP and haven't sent yet, three things to do before the day-one send:
One, identify your top 50 most-engaged subscribers. The day-one audience.
Two, write or pick the day-one content. Clean, normal, engagement-optimized.
Three, set up daily monitoring of bounce rate, open rate, click rate, and complaint rate. You need these visible from the moment the day-one send fires.
That's the floor. The day-one send goes out at 50 (or whatever your target-scaled equivalent is), the metrics get watched, and the curve continues from there.
The first day is the most important day. The number is small. Go small. Don't run through the powder.
- Spark
Visual Brief
V1 hero pose specification for the designer. One illustration. Sticker-style. White background. Match WU asset aesthetic.
- Adult cute proportions but slightly smaller scale (~1:3.2 head-to-body - a touch leaner than other crew)
- 18 apparent age. Irish (Inis Mór, Aran Islands). Still the youngest crew member.
- Hair: messy ginger, sticking out from under the cap, visible at the sides and back
- Skin: warm peach with a few scattered freckles across the nose and cheeks
- Eyes: wide, alert, slightly nervous-determined. Small dot pupils with two highlights to read as wide-eyed.
- Eyebrows: slightly raised in concentration
- Mouth: slightly open, lower lip caught between his teeth (signature character detail - he bites his lip when concentrating)
- A tiny scar at the corner of his lower lip (from the lip-biting habit)
- A small smudge of soot on the tip of his nose (signature character detail)
- Sailor's cap a size too big, in faded ember-red, slipping slightly to one side over messy ginger hair
- Scrappy ember-red and gold sleeveless vest with a visible powder-smudge across the chest
- White long-sleeve undershirt under the vest, sleeves rolled to mid-forearm
- Oversized cuffed brown trousers (tucked into laced boots), the cuffs visibly rolled twice
- Brown leather laced ankle boots with worn toes
- Small glowing barrel of powder held in both hands at chest height. The barrel has visible iron hoops, a small ember-red glow leaking from a narrow gap at the top, and a tiny gold spark suspended just above the barrel (subtle visual detail, sticker-style appropriate).
- A small leather strap holding a hammer-shaped tool at his hip (the "powder monkey hammer," a real shipboard tool for tightening barrel hoops)
- A folded piece of paper sticking out of his vest pocket (the warming checklist Petros made him write - signature character detail)
- No magical glow effects beyond the barrel and the gold spark above it
- No animal companion in V1 pose (deferred - would be a small ember-finch perched on his cap, also slightly worried-looking)
- No floating warming-curve symbols around him
- No detailed scene background (white only)
- No formal facial hair (he's 16)
WU/public/assets/captain/ for uniform structure, double-breasted coat, brass buttons, peaked cap pattern. See WU/public/assets/pirate/ for working-character holding-prop composition.
Game Content
Cards and tasks that belong to Spark in the Shipshape game.
Cards (5)
Tasks
81 tasks in Spark's task inventory. Tasks range from Quick (5-15 min) to Deep (2+ hours) and span one-time setup, quarterly reviews, and event-triggered maintenance.