There is a version of outbound that founders imagine, and it starts on Monday. Pick a tool, import a list, write a message, press send.
The real version starts with a credit card, a registrar, and about a month of nothing visible happening. Nobody wants to pay for that month. It is the most common reason I see a campaign die before the copy ever gets a chance.
So here is what actually goes into the sending side, in the order we build it.
1. Never send cold email from the domain your invoices come from
If your company emails live on yourcompany.com, cold campaigns do not go out from yourcompany.com. Ever. Cold email carries a reputation risk that has no upper bound, and the day your main domain gets flagged, your invoices, your support replies and your contracts start landing in spam folders alongside your outreach.
Buy separate domains. Close variants of your brand are fine, and they should point somewhere real rather than at nothing.
For a serious campaign that means 10 to 20 domains and somewhere between 20 and 40 mailboxes across them, more for a large company with a big market. On one software development client in Germany we ran 10 domains with 20 email accounts on top, and that was for a single team.
2. Authentication records before anything else
SPF, DKIM and DMARC get set up on every domain before a single message goes out. These are the records that tell receiving servers your mail is really yours. Without them you are a stranger claiming to be a company, and spam traps are built for exactly that.
This is a one-hour job for whoever manages your DNS, and skipping it undoes everything else on this page.
3. Warm up the mailboxes and actually wait
New mailboxes send small volumes to real inboxes and get replies, slowly, until they look like accounts a person uses. Ten days is the minimum I would accept. Two weeks is better, and on most builds we give those accounts 10 to 14 days before they touch a prospect.
Founders hate this step more than any other, because it is the part where they are paying and nothing is happening. It is also the part that decides whether the first campaign lands in an inbox or in a folder nobody opens.
4. Do the arithmetic on volume per mailbox
This is where most self-run campaigns quietly break.
Take a real month from that German client: 5,000 people contacted by cold email, 15,000 emails in total across the sequence, sent from 20 addresses. Spread across a month that is roughly 25 emails per mailbox per day.
Twenty-five. Not two hundred.
Now compare that to the usual do-it-yourself setup: two mailboxes, a tool with a monthly fee, and a thousand sends a day. That is not ambition, that is a bonfire. The volume you want does not come from pushing each mailbox harder. It comes from having enough mailboxes, which is why the domain count at the top of this page is what it is.
5. LinkedIn has a hard ceiling of about 2,800 people a month
That is what we plan around per account, per month. It is the practical ceiling for one profile behaving like a human being rather than a script.
Which means one founder with one profile cannot produce the volume most founders want, no matter how good the message is. On a financial advisory client in Australia the setup was 5 LinkedIn profiles and 20 email accounts, split so that each rep had one LinkedIn account, one email account for follow-ups, and three for cold email.
Split the territories too, so two of your people are never sitting in the same prospect's inbox in the same week. That comes out of the list, not out of the tool.
6. Start warm, then add the cold volume
The sequencing of channels matters as much as the setup.
We usually open on LinkedIn with email follow-ups behind it, so that a cold inbox is not where the relationship starts. Then, after the first two or three weeks, cold email volume goes on top of a system that is already running and already producing feedback.
The same logic applies to scaling. After 10 to 14 days of real data you add accounts, and you add them to whichever channel is actually working in that niche. Scaling is for things that have already proved they work. Doubling down on a channel that produced nothing in two weeks just produces nothing faster.
7. Clean data, or none of the above matters
Every address gets verified before it enters a campaign, verified addresses only, bounce rate under half a percent. Bounces are the fastest way to undo three weeks of careful warmup. I have written about what I look for in a list separately, because that part deserves its own argument.
How long this takes, honestly
Thirty to thirty-five days before the first campaign goes live is normal. On one UK client we spent 33 or 34 days building the foundations before anything went out, and that was a client who was in a hurry.
During that month there are no replies to report, no calls booked, nothing to screenshot. Then it runs, and it keeps running, and everything after that is iteration rather than construction.
What breaks, in order of how often I see it
Sending from the main domain. No authentication records. No warmup, or three days of warmup because somebody was impatient. One mailbox being asked to do the work of twenty. Unverified lists bouncing hard enough to undo the warmup. The exact same string of text sent ten thousand times with no variants, no spintax, nothing to distinguish one send from the next.
And the quiet one: a system built to contact a hundred people a month when the goal needs ten thousand. For 30 to 35 sales calls a month from outbound you are contacting at least 10,000 people, and that number needs infrastructure capable of reaching around a thousand people a day. You cannot bolt that on in week three. It is a decision you make at the registrar, before you write a word.
None of this is clever and none of it is fun. My team runs this part for clients because most companies would rather not think about DNS records at all, and that is a perfectly reasonable preference. But whoever does it, it gets done before the first send or it gets paid for later, usually in a month of silence that everybody blames on the copy.