M1 Find your market
Most companies guess an ICP and spend the year defending the guess. The first campaign should be an instrument, not a bet: small volume, several personas at once, and the replies decide which one is real.
Before the ICP is chosen
Every dot is a contact in the addressable list. Nothing is grouped yet, because nothing has been decided yet. The run tests several candidate personas at once and watches which one pulls itself into focus.
How the run works
A first campaign that guesses a single persona spends the whole list defending the guess. This one runs several ICPs at once, small, so the market gets to answer before you commit to one.
Every ICP gets its own list, cross-referenced the same way any Empra database is, carrying the variables the engine writes from. Built once, and it is yours whichever persona wins.
Several angles per ICP, several variants per angle, live at the same time. Every message still passes a human before it sends, whichever ICP it belongs to.
Not opens, not a hunch about who should care. Every reply reaches the reply agent within 30 minutes, so the read on what is working stays current while the test is still running.
The persona and the message that earned real replies get the budget next. Everything else was the cost of finding out, not the campaign.
Why it starts small
Scale behind a guessed persona and every address you touch with the wrong message is spent. The list does not reset, and neither does its opinion of you.
Testing several ICPs at a small volume costs a few weeks. Guessing wrong at full volume costs the market, before you have even found out who was in it.
Other missions
EMPRA