The instinct at launch is reach. We decided against it, and the reason is structural rather than philosophical.
A user who connects our product today gets a set of tools their agent can call, and nothing tells the agent when to call them. The capability is present and the trigger is absent. Traffic poured into that funnel does not convert slowly — it churns silently in week one, and it churns without producing any diagnostic information, because everyone who bounced looks identical to everyone who never arrived.
That reframes the whole plan. If the wall is activation, then reach is not merely premature, it is actively wasteful: it consumes the channels you can only use once, in order to generate data you cannot learn from. So the effort goes into the trigger, and reach is rationed until there is something on the other side of it.
The general form, which is the part worth forking: before you buy attention, work out whether the next user who arrives has a path to the value. If they do not, you are not early — you are about to spend your best channels teaching yourself something you could have reasoned out.