# Traction

Intent to outcome is the measure that matters. Beautiful does not yet measure it. This page
shows what is actually observed today, names what is not and never prints a missing measurement as
a zero.

## How to read this

A number here is one of four things. They are never collapsed into each other:

- measured: a service answered. This is what it said. A zero here is a real zero.
- unavailable: a service exists and did not answer. The number is unknown and is shown as
  "unavailable" rather than as any figure.
- not yet measured: nothing in the estate records this. There is no number to know.
- no history: the current value is known, but no series has been retained to plot.

Nothing on this page is estimated, interpolated or projected. No target is presented as
traction.

## Reps (measured)

Counts from the rep registry, read from beautiful-provisioning over a service binding. Aggregates
only: no identifier, no origin, no principal and nothing about any individual rep. Sample: every rep
in the registry. Freshness is stamped from Beautiful Time, not from a machine clock.

- Total reps and reps born in the last 7 and 30 days.
- By lifecycle: hatched (an identity exists, not prepared to act), ready (prepared and permitted to
  act), suspended (stopped).
- By role, which is a separate classification. Lifecycle and role are never added together.

A provisioned rep is not an active human. Lifecycle describes where a rep is in its own life. It
says nothing about whether a person engaged with it, asked it for anything or read what it sent.
Automated news or wake-ups do not, by themselves, establish a daily active human.

There is no time series yet. The registry answers with the counts as they are now. Nothing retains a
daily observation, so there is no history to plot. Back-filling the days before today
would be inventing it. History begins at the first retained snapshot. This page will state that
date.

## Not yet measured

Five of the six things worth knowing about this business are not instrumented. They are listed
rather than drawn as empty charts, because five blank graphs would claim we measure five things and
they are all zero, which is a different and much worse claim.

- Outcome delivery: accepted intents, achieved outcomes, completion rate. Task and turn states exist
  per rep; nothing aggregates them across reps. No record marks a journey abandoned as distinct
  from cancelled.
- Retained relationships: activation, retention, humans receiving useful representation. The per-rep
  engagement signals are recorded; no route publishes an aggregate of them.
- Cost profiles: average cost per user, cost per successful outcome. No shipped service records
  token usage, model cost or infrastructure cost.
- Payment signals: paying subscribers, revenue, ARPU. No subscription, payment, invoice or revenue
  record exists in shipped code.
- Cooperation: shared outcomes, referrals leading to activation. Who introduced whom is recorded;
  whether it led to activation is not, because activation itself is not yet measured.

## Not a hockey stick. Yet.

Beautiful has only just opened. This page shows an honest beginning rather than manufactured
momentum. Everything above is either something a service told us or a plain statement that it did
not.