Metric Formulas

A metric formula is a versioned calculation method for a Metric, naming its inputs, grain, evaluation rules and validity. It is not the Metric it computes, and not the evidence either.

Its metric formulas— 1

• Reactivation rate formulaexemplarA versioned calculation method: fulfilled conversions by lapsed consumers over lapsed consumers in the window, with the grain, exclusions and evaluation rules stated once.

Why it matters

What goes wrong without it, as the commitments that license this kind say it, each standing on its sources.

• Make outcome quality, calculation method and attribution inspectableAn outcome number is only as good as the inputs, exclusions and window that nobody wrote down. Media measurement standards exist because outcome claims circulate detached from their method, and a figure with no stated method cannot be challenged, reproduced or compared against anyone else's. Keeping what a measure means apart from how it is computed is what makes either one reviewable.
  Outcomes and Data Quality Standards · Declaring the Grain · Measuring Visibility in the AI Era · Data-driven multi-touch attribution models · Causally motivated attribution for online advertising · A Comparison of Approaches to Advertising Measurement: Evidence from Big Field Experiments at Facebook · Why Three Exposures May Be Enough · Digital Audience-Based Measurement Standards
• Scope activation and calculation explicitly in time and geographyTwo reports can describe the same activity and disagree entirely because one covers a different window, or a different market, than the other. Free-text dates and place names put that scope beyond the reach of any check, so the two compare cleanly and wrongly. Time and place are dimensions with a boundary and a basis, not labels on a record.
  Time Ontology in OWL · GeoSPARQL 1.1
• An exemplar shows the shape a record takesA model with no instances cannot be evaluated, and a model with invented instances cannot be trusted. Treating an example as documentation of a concept rather than an assertion about the world resolves the bind: one marked exemplar shows what a record of the kind looks like, while the claim that any such record exists stays unmade.
  SKOS Simple Knowledge Organization System Reference · Toward Principles for the Design of Ontologies Used for Knowledge Sharing
• A move that cannot be counted is a pictureA 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.
  Modeling customer relationships as Markov chains · Declaring the Grain · The Goal Question Metric Approach · Counting Your Customers the Easy Way: An Alternative to the Pareto/NBD Model