Domains and Warmup for Podcast Invite Campaigns

Most teams treat domain warmup as a deliverability chore. We run outbound for 50 plus B2B companies and have sent over 8 million invitations this year, and on an invite campaign warmup is a scheduling item, because the day you buy the domains is what sets the day your first guest sits down to record. Below, the mailbox math behind a 10,000 invite month, the warmup curve that has to finish before invite 1, the sender identity rules, the 4 week ramp to full volume, why invites hold reputation better than pitches, and the 4 numbers that call a domain going bad.

What Does a Podcast Invite Campaign Need for Domains and Warmup?

None of those numbers are unique to podcast invites. What is unique is the deadline attached to them. A pitch campaign can slip a week and nobody notices. An invite campaign is booking a calendar for a human being who has cleared time to host, so the infrastructure has a fixed date on the other end of it, and every day of warmup you did not plan for is a day the show sits empty.

That reframes the whole setup. You are not asking what infrastructure produces the best inbox placement. You are asking what infrastructure produces acceptable placement by a specific Tuesday. Those two questions have different answers, and the second one is the one that gets asked too late.

Sending Domain

A domain registered for outbound email only, usually a close variant of the company's real domain, redirecting to the real site. It carries the mailboxes that send invitations. If its reputation goes bad, it gets retired without touching the company's primary domain.

Warmup

A staged ramp of low-volume mail between a new mailbox and a network of real inboxes, with opens and replies, so mailbox providers see a normal sending history before real volume starts. It runs 21 to 28 days on a new mailbox and then continues in the background for the life of the campaign.

Why Is Warmup a Launch Date Problem, Not a Deliverability Problem?

Work the calendar backwards and the point makes itself.

Say a client wants their first recorded conversation in the first half of November. The recording has to be booked, so an invitation has to land in October. The invitation cannot go out until warmup is finished, which is 3 to 4 weeks. Warmup cannot start until DNS has propagated and the mailboxes exist, which is another few days. And domains should be registered a couple of weeks before any of that, because a domain that was bought yesterday and is sending today looks exactly like what a spammer does.

Add it up and the domain purchase sits 5 to 6 weeks ahead of the first invitation. Nobody feels that gap while they are picking a show name and arguing about cover art. They feel it the week they are ready to send and the mailboxes are 9 days old.

The list work is the part that can run alongside warmup, and it should. Verification, enrichment, and segmentation do not touch the sending reputation, so there is no reason to do them in sequence. Warmup is the only step on that table that cannot be compressed by working harder, which is exactly why it belongs at the front of the plan instead of the middle. We commit to a client's first recording landing within 14 days of the first invitation going out, and the only way that commitment is keepable is if the infrastructure was started 6 weeks before it.

How Many Domains and Inboxes Do 10,000 Invites a Month Need?

The domain count is arithmetic, not preference. Take the monthly invite target, divide by what one mailbox can carry, and you have your mailbox count. Divide that by mailboxes per domain and you have your domain count.

Here is the math we run on every build:

1. 20 cold sends per mailbox per day. That is the ceiling we hold, and we hold it whether the mailbox has been running for 3 weeks or 3 years. The temptation to push to 50 shows up every time somebody wants volume faster. It is the single fastest way to burn a domain that was fine yesterday. The reasoning is in how many cold emails you can actually send per day.
2. 22 sending days a month. Weekdays only, minus holidays. So one mailbox carries roughly 440 invitations a month.
3. 10,000 invites divided by 440 is 23 mailboxes. Round up for the mailbox you will inevitably have to pull out of rotation.
4. 3 mailboxes per domain. More than that concentrates too much volume behind a single reputation. Fewer than that multiplies your DNS surface for no benefit. 23 mailboxes lands at 8 domains.
5. Build in a spare domain. A ninth domain sitting warm and unused is cheap insurance. When one goes sideways in month 4, you swap instead of waiting 4 weeks for a replacement to warm.

Run that arithmetic at a few volumes and the sizing table writes itself:

The spare column is the item most builds skip and most builds later regret. The full decision tree for domain count by volume is in multi domain sending, 5 domains versus 50, the architecture rules sit in how to set up email domains for outbound, and the whole stack from send target backwards is in the cold email infrastructure setup. The short version: the number of domains is a function of volume, and anybody quoting you a domain count before asking your monthly send target is guessing.

One caveat on the top row of that table. Sizing the pool to the invite target only works if the invite target is real, and the invite target is a function of how many qualified names the market actually holds. A 20,000 invite month against a total addressable market of 6,000 companies is not a volume plan, it is a plan to mail the same people 3 times. Size the market first, in how to define an ICP for cold email, then size the infrastructure to it.

What Has to Land Before Warmup Day 1?

Warmup on a misconfigured domain teaches the mailbox providers something you do not want them to learn. Get the records right first, then start the ramp.

• SPF, DKIM, and DMARC, all 3, all aligned. Not 2 of 3. Google's sender guidelines and Yahoo's sender best practices both treat authentication as table stakes, and Outlook's high volume sender requirements put the same 3 records behind enforcement. The record-by-record walkthrough is in what SPF, DKIM and DMARC actually do and the DNS records explained.
• A redirect on every sending domain. The lookalike domain should resolve to the real website, not a parked page. A recipient who checks is a good sign, and a dead page turns a good sign into a bad one. More on the pattern in what a secondary domain is for.
• A real signature and a real profile photo on every mailbox. Cheap, and it changes how the message reads to a human before it changes anything for a filter.
• One-click unsubscribe wired up. Required for bulk senders, specified in RFC 8058, and more usefully the pressure valve that keeps an annoyed recipient from reaching for the spam button instead.
• An SPF record inside the lookup limit. RFC 7208 caps SPF evaluation at 10 DNS lookups, and stacking includes for a sending tool, a mailbox provider, and a redirect host blows through it quietly. A record that exceeds the limit fails, and a failing record during warmup is worse than no warmup at all.
• A DMARC policy you actually chose. Start at p=none with reporting on, read the reports for a week, then move up. The policy ladder is laid out at dmarc.org.
• A verified list waiting. Bounces on the first real batch undo weeks of warmup in an afternoon. This is what verification is protecting, and it matters more on day 22 than at any other point in the campaign. The thresholds that trigger a block are in cold email bounce rates, causes and fixes.

What Sender Identity Goes on a Podcast Invite?

An invitation is a message from a person. That reads as obvious until you look at how most invite infrastructure actually gets built, with 23 mailboxes across 8 domains and a From name that was typed into a signup form once and never revisited.

Sender identity belongs to the infrastructure, not to the copy. It is set at the mailbox level, it is identical on every send from that mailbox, and changing it later throws away part of the history you spent a month building. So it gets decided with the domains, before warmup starts, not during the week everyone is arguing about subject lines.

Sender Identity

The fixed set of attributes a mailbox sends under: the From name, the From address, the reply-to, the signature block, and the profile photo. Mailbox providers read these as part of the sending history, so a change mid-campaign reads as a new sender on an old domain.

Five rules we hold on every invite build.

1. The sender is a real person who is really attached to the show. Full first and last name in the From field. An invitation signed by a team alias is not an invitation, it is a mailing. The name should be findable on LinkedIn and should belong to somebody the guest could plausibly meet, because senior people look you up before they answer. A real name attached to a real show is the only version that survives that.
2. One identity per mailbox, held for the life of the mailbox. Providers read the From name and address as part of the reputation attached to the domain. Swapping the name on a warm mailbox to chase a new segment discards the engagement history sitting behind it, and that history took 4 weeks to build.
3. Signature variants across the pool, not one footer cloned 23 times. On our own invite campaign the signature rotates across 5 variants. One footer cloned across a pool is a pattern, and patterns are the thing filters are built to find. The same reasoning applies to the body, covered in how to personalize cold emails at scale.
4. Reply-to matches the From address. A mismatch between the two is a textbook phishing signal, and plenty of filters treat it as one regardless of what you intended. M3AAWG's published sending documents are the industry reference on this layer.
5. Cold invites go out on the sending domains. Everything after the yes moves to the show's own address. The moment somebody agrees to record, confirmations, reminders, and the calendar invite all come from a real mailbox on the show's own domain. The sending domains are disposable by design. A booked guest is not, and the relationship should not sit on infrastructure you might retire in month 7.

There is a sixth rule that is the easiest to break and the hardest to undo. The company's primary domain never goes into the sending tool. Not as a sender, not as a test, not for the one message somebody wants to send from the real address because it feels more credible. The whole point of the lookalike structure is that reputation damage lands on a domain you can retire, which is the argument in what a secondary domain is for.

One more identity detail that changes reply rates more than it should. The From name and the show name need to agree with each other. If the invitation comes from a producer and the show has a name the guest can look up, the message holds together. If it comes from a stranger with no show attached, the ask has nothing behind it. The full invite construction is in what to say when inviting a podcast guest and how to invite guests to your B2B podcast.

What Does the Warmup Curve Look Like Week by Week?

The curve itself is not controversial. Start small, climb slowly, and let the tool handle the opens and replies. What varies is patience.

1. Week 1, 5 to 10 a day. The mailbox has no history at all. This week exists to prove it is a real account behaving like one.
2. Week 2, 10 to 20 a day. Reply activity climbs. Placement should be sitting in the primary inbox across providers by the end of this week.
3. Week 3, 20 to 30 a day. The mailbox now looks like a person with an active thread history rather than a fresh install.
4. Week 4, 30 to 40 a day. Hold here. This is the week most people skip, and the difference between a mailbox that survives month 6 and one that does not usually got decided in it.
5. Day 22 onward, mix in real sends. Start real invitations at 5 to 10 a day per mailbox and step up to 20 over the following week. Do not go from warmup straight to full volume on a Monday morning.

The mistake worth naming: a warmup tool showing green at day 10 is reporting on its own network, not on how Gmail feels about you. Practitioner guides like MailReach's warmup walkthrough and Instantly's deliverability guide land in the same 3 to 4 week band for a reason. Our full day by day version is how to warm up a new email domain, with the concept explained plainly in email warmup explained.

How Do You Ramp From Warmup to Full Invite Volume?

Warmup finishing is not the same as the campaign being ready for its target volume, and conflating the 2 is how a pool that warmed cleanly gets burned in week 1 of real sending.

The reason is simple once you say it out loud. Warmup traffic goes to a cooperative network of inboxes that open, reply, and never complain. Real invites go to strangers who can mark the message as spam. Day 29 is the first time a domain has ever been exposed to actual complaint risk, so the campaign gets its own ramp on top of the one each mailbox already finished.

On a 23 mailbox pool built for 10,000 invites a month, the campaign ramp looks like this:

Three things about running that ramp that are not obvious from the table.

Send week 1 to your best segment. The cleanest verification, the tightest fit, the names most likely to reply well. The temptation is the opposite, to burn the weak segment first while the stakes feel low, and it is exactly backwards. Week 1 is when the domain is establishing what kind of mail it sends, so it should be the best mail you have. Qualification criteria are in how to qualify podcast guests before you invite.

Bring mailboxes live in cohorts, not all at once. If 8 domains finished warmup on 3 different days, they ramp on 3 different clocks. Forcing them onto one schedule means somebody is sending at week 3 volume on a week 1 mailbox.

Hold the follow-up steps off until week 2. A first invitation and a 3 step follow-up sequence firing in the same week doubles the volume hitting a mailbox that just came off warmup, and the follow-ups are the sends most likely to draw a complaint. Sequence timing is in cold email follow up sequence best practices.

Why Do Invite Replies Protect the Domain?

This is the part that makes invite infrastructure genuinely different from pitch infrastructure, and it is the reason we moved the whole business onto invitations.

Mailbox providers weight engagement. Replies, forwards, stars, and a message getting dragged out of promotions all read as wanted mail. Deletions without opening and spam complaints read the other way. Warmup exists to fake the first category until you have earned it for real.

An invitation earns it faster than a pitch does. Asking somebody to be a guest on a show gets answered, including by the people saying no, because declining an invitation feels like manners and ignoring a sales email does not. Across our book that lands reply rates around 4.6 percent against the 3.43 percent market median Instantly reports in its 2026 benchmark report, and every one of those replies is a positive engagement signal hitting the same reputation the warmup tool was simulating a month earlier.

Put the 2 campaign types side by side on the signals providers actually read and the gap stops being a matter of opinion:

Read down the right column and the reason the pool lasts longer becomes mechanical rather than philosophical. Every row is a reputation input, and the invite version wins most of them because the ask is different, not because the infrastructure is better. The same domains, the same warmup, the same volume ceiling, fed different traffic. Benchmarks by segment sit in cold email reply rate benchmarks, and the split between a reply and a reply worth having is in what a positive reply rate is.

There is a second-order effect worth knowing. A pitch campaign's reputation decays as the list gets colder, because the message never gets better and the recipients never get warmer. An invite campaign's reputation tends to hold, because the traffic includes real conversations with real people who answered on purpose. The infrastructure is doing the same job in both cases. Only one of them is being fed. The mechanics behind that gap are in invite versus pitch and why executives say yes to podcast invites.

Does Warmup Ever Stop Once the Campaign Is Live?

No, and treating it as a phase that ends is how good infrastructure quietly degrades in month 5.

Background warmup should keep running at roughly 20 to 30 percent of daily volume for the life of the campaign. An invite campaign is not a 6 week sprint. It runs for as long as the show does, which means the mailboxes go through holidays, list rebuilds, a pause while a client changes their target market, and the odd week where sending stops for a reason nobody wrote down. A mailbox with background warmup running comes back from those gaps. A mailbox that went dark for 3 weeks comes back looking new again.

The same logic applies to rotation. Mailboxes should share volume evenly rather than one domain carrying the campaign while the rest idle. Idle domains do not stay warm on their own, and the day you need to swap one in is the day you find out. Background warmup and even rotation are the 2 habits that separate infrastructure that lasts a year from infrastructure that needs rebuilding every quarter, which is the expensive version of this problem covered in how to recover a burned domain.

Which 4 Numbers Tell You a Domain Is Going Bad?

Reply rate is a lagging indicator. By the time it drops, the placement problem is weeks old. Watch these instead, weekly, per domain rather than per campaign:

1. Inbox placement from a seed test. The only direct read on where mail is landing. Under 60 percent primary placement, that domain comes out of rotation. Method in inbox placement tests.
2. Spam complaint rate. Google asks bulk senders to stay under 0.3 percent and to aim below 0.1 percent. Treat 0.1 percent as the ceiling. Google Postmaster Tools is where you read it for Gmail traffic.
3. Bounce rate per domain. A domain-level spike while the others hold steady is an infrastructure signal, not a list signal. A spike across all of them is the list.
4. Blacklist status. Weekly checks catch a listing while it is still 1 domain instead of 4. Details in blacklist monitoring and the wider dashboard in deliverability monitoring.

Run a deliverability test on the primary sending domain every week and log the result. The number itself matters less than the trend, and the trend is invisible unless somebody is writing it down. When placement slips, the fix is volume and rotation, not copy. Rewriting an invitation while the domain underneath it is degrading burns a rewrite and teaches you nothing, which is the argument in why your cold email is landing in spam and how to stay out of the spam folder.

The Practitioner Take on Invite Infrastructure

Every campaign we have watched fail on infrastructure failed for one of 2 reasons, and neither of them was a hard technical problem. Either the domains were bought too late and warmup got cut to 10 days to make a date somebody had already promised a client, or the daily volume got pushed past 20 because the target felt slow at week 3. Both are decisions made under time pressure by somebody who understood the risk and took it anyway.

So the useful discipline is not technical. It is putting the domain purchase on the calendar before the show name, the cover art, or the first list pull. Those things can be finished in a day. Warmup cannot, and no amount of budget shortens it.

The direction of travel makes the gap wider, not narrower. Google, Yahoo, and Microsoft have all tightened bulk sender enforcement in the last 2 years, starting with Gmail's bulk sender requirements, and none of them are loosening it. Authentication, complaint rates, and engagement history are becoming the whole game, which is the measurement stack laid out in what email deliverability actually is. That is bad news for anybody sending generic mail at volume and quietly good news for invite campaigns, because an invitation generates exactly the engagement the new rules reward. Build the infrastructure 6 weeks early, hold the volume ceiling, keep warmup running in the background forever, and check placement weekly. The copy can be fixed in an afternoon. The domain cannot.