A move that cannot be counted is a picture

Answers Derivation · Epistemology

Every legal move between journey positions is triggered by at least one event that an instrument observes, and that instrument is a numerator or denominator of a reading a KPI carries, or an input of the formula that computes one. The fact that enters a condition is the denominator of the rate that measures leaving it, so a state-entry fact is counted once as the move in and once as the population the next move leaves from. A move triggered by a window closing counts over its journey's window: every journey names the interval of inactivity after which it presumes a person has stepped back.

What goes wrong without it

A journey that names its moves and cannot count them is a funnel with arrows: it can be drawn and not measured, which is the failure a state machine was adopted to end. Customer-migration models make the point exactly. A transition rate is the moves out of a condition over the people in it, and the people in it are the fact that put them there; so the events added to give every move a trigger are not extra instrumentation but the missing denominators of rates the model already carried and could not derive. A law that every move has a trigger, satisfied by facts nothing observes, would be the same picture with more arrows.

What proves it — 1

Each is a gate the build runs: a check inside a ring, a whole ring, or an eval beside them. A build stops at the first that refuses.

• registry integritymove propagation
  Every legal move between journey positions is triggered by an event that an instrument observes, and that instrument counts a reading a KPI carries, directly or through the formula that computes it, so a journey that names its moves can also count them and is a model rather than a funnel with arrows; and a move triggered by a window closing belongs to a journey that names the window it counts inactivity over.

Derived from — 4

• Pfeifer and Carraway, Journal of Interactive MarketingModeling customer relationships as Markov chains
• Kimball GroupDeclaring the Grain
• Basili, Caldiera and Rombach, University of MarylandThe Goal Question Metric Approach
• Fader, Hardie and Lee, Marketing Science 24(2), 2005Counting Your Customers the Easy Way: An Alternative to the Pareto/NBD Model

Kinds it licenses

Relationships it licenses

triggered by · observes event · counts in numerator · counts in denominator · takes input · has metric · times out after

Where this model stands

Every move's trigger is observed by an instrument that counts a reading, including the fact that a window closed with nothing done, the fact that a model scored someone, and the fact of a referral.

Leaving is read where staying is: a lapse rate and a likely-to-lapse rate stand beside the retention readings as the same population read from the other side, and both improve by falling.

An arrival is a move from the journey's beginning, triggered by the fact that first made the person known, and the mode of arrival is read from where that fact was captured, so arrival conducts as a move rather than standing beside the machine as a condition.

A scored move rests on a model of whether a relationship is still alive, and a timed move on the window after which it is presumed not to be; both are readings a probability model makes of inactivity, not observations of a person acting, and the instrument that observes them is the model or the clock, named as such.

Each journey names the window it counts inactivity over, so a move triggered by a window closing says which clock ran out: the evaluation window for a person who stepped back before committing, the activation window for one who never started, the lapse window for a relationship that went quiet, the advocacy window for an advocate who brought no one in. The window's length is the operator's to set and is not asserted here; that a journey has one, and which, is.

A stay is a move. An account holder who converts again moves from the established position to itself, the diagonal of the customer-migration matrix, so staying is counted by the same fact as arriving and a retention reading has a numerator that is an act rather than an absence. A state machine allows a transition to its own state, and the twin declares it as one.

What it does not claim

Whether a KPI that counts a move also serves a goal that names the move's rung is reported, not required: an efficiency goal legitimately names no rung.

RUL-025