Illingworth says config rebuild took a client 0.9% to 23% replies
One operator's before and after, with the copy and the list held constant.
Richard Illingworth says a client whose cold email reply rate was stuck at 0.9% ended up at 23% after he rebuilt the sequencer and the sending configuration, and changed nothing else.
By his account the client was ready to scrap the entire offer. The read was that the market had gone cold and the whole thing was dead in the water. Illingworth says he looked under the hood instead and found the sends were being cooked long before they reached an inbox, with the sequencer set up wrong and the sending config hurting the client from the very first email. He rebuilt the setup and left the copy and the list exactly as they were.
What the post does and does not say
The numbers are the two endpoints, 0.9% and 23%. Illingworth does not publish the volume sent, the time period, the sequencer in question, or which settings he changed. That matters, because a 0.9% reply rate and a 23% reply rate on the same list are not usually the same experiment unless the sample is big enough to rule out luck, and the post gives no way to check. Treat it as one sender's before and after, not a benchmark you can expect to reproduce.
The diagnosis is still the useful part. If the carrier is broken, the words never get tested, so the rewrite that follows a bad campaign is often work spent on the one variable that was not the problem. His argument is that copy is the part that feels controllable, which is why it absorbs the blame.
For people sending
The practical version of this is an audit order, not a tactic. Before another rewrite, check what the sequencer is actually doing on send one, because that is where Illingworth says the damage started. Sending settings, ramp, and sequence structure are testable without touching a single line of copy, and holding the copy and the list constant is the only way to know whether the setup was the thing dragging the number down. If your reply rate sits near 1% on a list and an offer you previously trusted, the cheaper hypothesis is infrastructure.
Until the specific changes are published, the claim is a prompt to go looking, and nothing more than that.
your copy can be flawless and STILL rot at 0.9% if the thing carrying it is broken.
so i rebuilt the setup and left his copy and his list exactly as they were.