Sequence

The useful vocabulary is small and it matters that it includes work a person does. A sequence that can only send email models a mailing, not a sales motion:

• Automatic email — sent by the engine through your sending mailbox.
• Manual email — drafts a task instead of sending, for the messages worth writing by hand.
• Call, and other non-email touches — a task with a due date, on the same timeline as everything else.
• Delay — an explicit wait, so the gaps are visible in the design rather than implied.
• Condition — a branch on a field, so one sequence can serve two situations.

The published benchmarks are consistent on two points. Single-touch outbound underperforms badly, and the returns fall off after roughly eight to ten touches — the tail generates complaints faster than replies. Somewhere in between is the whole game.

The other consistent finding is that the middle of a sequence outperforms its opening. The first message is easy to ignore; the third arrives when the name is faintly familiar. Teams that judge a sequence on the first step's reply rate usually delete the part that was working.

Every step must be able to stand alone — a follow-up that only makes sense if the previous one was read is wasted on the majority who did not read it. Each should also add something: a different angle, a specific piece of evidence, a smaller ask. Bumping this to the top of your inbox is not a step.

And the sequence has to stop the moment someone replies. A person who answered on Tuesday and receives step four on Wednesday has learned that nobody was reading.

A SalesShift sequence supports automatic email, manual email, call, task, delay and condition steps, with per-step delays in days and hours. Step one fires on enrollment; the engine advances from there. Sending obeys the sequence's window, days and timezone, and per-step analytics show where enrollments actually stop.