The campaign worked. Month one brought replies, month two brought calls, and somewhere around month three or four the same sequence started producing nothing. Nobody touched the copy. Nobody changed the offer.
Almost every time, the answer is one of three things, and they are not equally likely.
Most likely: you ran out of list. Second: your sending domains lost reputation while you were busy celebrating. Third, and rarest: you saturated a market that was smaller than you thought.
The reason this feels mysterious is that all three produce the same symptom. Fewer replies, from the same message, sent the same way. You cannot tell them apart by looking at the reply rate. You tell them apart by looking at what changed underneath it.
Cause one: the list ran out, and it ran out in order
This is the one almost nobody checks, because it does not feel like a cause. The list is still sitting there in the tool with thousands of rows in it.
Do the arithmetic. Twenty mailboxes, thirty sends a day each, twenty two working days a month is about thirteen thousand sends. Run a four step sequence and every contact absorbs four of those sends. So you are actually touching somewhere around three thousand new people a month, not thirteen thousand.
If you started with an eight thousand name list, you were out of fresh contacts before the end of month three. Everything after that was follow ups to people who had already decided.
That is the mechanical half. The important half is the order.
Nobody builds a list at random. You targeted the companies you know best first, the titles you are most confident about, the geography where your name means something. Your best fitting contacts were at the top, whether you sorted them that way deliberately or not. So month one was never a representative sample of your list. You were emailing the best part of it. The decline you are seeing is partly the campaign getting worse and partly the list reverting to its actual average.
This is why my position on list size is the least glamorous one I hold: the first list should be at least twenty thousand contacts. Volume by itself wins nothing. The reason for the number is that anything smaller runs dry before you have learned anything from it, and you spend the next quarter rewriting an email that was never at fault.
What to do. Count how many genuinely new contacts entered the sequence in each of the last four months. If that number is falling, you have your answer and nothing else on this page applies yet. Build the next tranche before you touch a single line of the email.
One warning on the refill. Small, hand-picked campaigns almost always report better reply rates than big ones, and that gets quoted as a reason to send less. It is the same effect measured from the other end. A short list is the top of a long one. Send to your five hundred best-fitting contacts and you will beat your own average, then watch the number fall as you work down the file. Shrink the list and you only reach the bottom of it sooner.
Cause two: the domains quietly lost reputation
Second in likelihood and first in how fast it hurts.
Sending reputation is a running score. It drifts, and it drifts downward whenever your list quality drops, which is precisely what happens in month three when you start scraping the bottom of the tranche. Bad addresses bounce. Irrelevant contacts mark you as spam. The provider notices, and your mail starts landing in a folder nobody opens, where it is recorded as an unopened, unanswered message exactly like a bad email would be.
The tell is a drop in replies with no drop in deliverability that you can see, because you are measuring delivery and not placement. Delivered and inboxed are different states.
Three things to check, in this order. Start with your bounce rate month by month, rather than the number your dashboard shows for the campaign overall, which hides the trend. Two percent is where most published advice lands, including one sending platform's 2026 read of its own campaigns. I run client campaigns to a tighter figure than that, under half a percent, checked before anything goes out. Stale rows and unverified addresses are what push it up.
Second, whether you added mailboxes mid campaign and skipped the warm up. New accounts need ten to fourteen days minimum before they carry cold volume. A rushed mailbox fails, and it takes the reputation of the domain it sits on down with it.
Third, your per mailbox volume. Vendors publish safe numbers for a warmed mailbox anywhere from 25 a day to 150, each describing their own setup. Whichever number you picked, doubling it after a good month one is how you end up paying for it in month three.
The full diagnosis for this one lives in why cold emails land in spam, and the data side of it in auditing a list before it burns a domain.
Cause three: the market really is saturated
Last, and genuinely rare, though it is the explanation founders reach for first because it is the only one that is not their fault.
Real saturation looks specific. You have contacted most of the addressable companies in a defined market more than once. The replies change in kind rather than in volume: people telling you they already spoke to you, or that they are already working with someone who sounds a lot like you. A falling reply rate by itself is not saturation. It is almost always cause one wearing a costume.
If it is real, the fix is a wider definition of who you sell to, or a second market entirely. No email solves that, and that is a separate conversation from this post.
How to tell which one you have, in an afternoon
Pull three numbers, month by month, for the whole life of the campaign.
New contacts entered into the sequence. Bounce rate. Replies.
Falling new contacts with a flat bounce rate is cause one. Flat new contacts with a rising bounce rate, or a volume increase you can point at, is cause two. Steady new contacts, clean bounces, and replies that tell you they have heard from you before is cause three.
Most of the time it turns out to be cause one with a bit of cause two on top, because the two feed each other. The list thins, the data quality drops, the bounces rise, the reputation slips, and the whole thing looks like a copy problem from the inside.
The reason it reads as a copy problem is that copy is the only variable most teams can see. Everything underneath it stays invisible unless somebody goes and counts it. Rebuilding lists and sending infrastructure so that campaigns stop dying in month four is most of what I get hired for, and the counting part you can do yourself this afternoon.
Before you rewrite anything, go and find out whether you still have people left to send to. The argument for why that comes first is in the list, not the email.