The mechanism is worth understanding. Email providers score your domain on how recipients respond. Deletions without opening, spam complaints, and messages to dead addresses all push the score down, and once it drops far enough your mail lands in spam folders regardless of what it says. A new domain has no history to absorb that, so early mistakes carry disproportionate weight.
Warm the domain gradually if it is genuinely new. A handful of messages a day for the first fortnight, rising slowly, keeps the pattern consistent with ordinary business use. A domain that sends nothing for a month and then five hundred messages in a day looks exactly like what a filter is designed to catch.
Send individually rather than in bulk, which matters more than the number. A message with one recipient, no merge fields visible, and no bulk sending infrastructure behind it is treated differently from an identical message sent through a campaign tool. At this volume the manual approach is also faster than configuring the alternative.
The research requirement enforces the limit naturally. If each message takes ten minutes to prepare properly, twenty is a substantial part of a day and thirty is unrealistic. Any figure much beyond that is evidence the research has stopped happening.
Watch the bounce rate closely, because it is the fastest way to damage a domain. Guessed addresses and stale list data produce bounces, and a high bounce rate signals to providers that you are sending to a list you did not build. Verify addresses where you can and never guess formats in bulk.
Spam complaints matter more than bounces and are less visible. A single complaint per thousand messages is roughly the threshold at which providers take notice, which at twenty a day means one complaint every couple of months is already too many. Relevance is the only reliable protection.
Keep the plain text and drop the extras on a first message. No images, no tracking pixels, no attachments, and no shortened links. Each of those raises the probability of filtering, and none of them help a first approach.
Use a separate domain for outreach only if the volume genuinely warrants it, which for a first year business it usually does not. A subdomain or an alternative domain isolates risk and it also carries no reputation, so it starts from a weaker position than your main one.
Then stop entirely if the response rate is zero after thirty messages. Continuing at volume with a message that is not working is how domains get damaged, and the problem at that point is the offer or the audience rather than the sending.
Check where your messages actually land by sending to accounts on the major providers you do not otherwise use. A message in the spam folder produces no reply and no signal that anything is wrong.
Set up the authentication records for your domain before sending anything, since messages from a domain without them are considerably more likely to be filtered regardless of content.
Keep outreach separate from your ordinary correspondence in whatever way your provider supports, so that a problem with one does not affect the other.
Pause entirely for a week if you see a sudden rise in bounces or complaints, and find the cause before resuming rather than sending through it.
Track the reply rate rather than the send count, because the number of messages sent is an activity measure and the reply rate is the only one that indicates whether the approach works.