Public measurement standard · Version 1.0 · 2 October 2026
The LSG Evidence Standard defines how Low Season Growth labels commercial evidence from weak-date hotel work. Its purpose is to prevent early signals, confirmed reservations and completed commercial outcomes from being presented as if they carried the same evidentiary weight.
E0 — Hypothesis / assumption
A commercial belief or planning assumption before the result is observed. Examples include an expected booking pace, a target segment, a proposed offer or an estimated baseline. E0 is useful for planning but is not evidence that an intervention worked.
E1 — Observed commercial signal
An intermediate response such as a qualified reply, direct enquiry, proposal request, booking attempt or reactivated past guest. E1 can guide the next action, but it is not a booking and not revenue.
E2 — Confirmed booking
A reservation has been confirmed and can be reconciled to the relevant property and stay period. It may still cancel, change or fail to complete. E2 can be reported as a confirmed booking when the attribution rule is stated, but it should not be reported as realized room revenue.
E3 — Completed stay
The stay has completed and the relevant room nights can be reconciled. Realized room revenue may be reported where it is available and authorized. E3 is stronger than a confirmed booking because cancellation risk for that stay has matured.
E4 — Mature case evidence
The measurement window is mature, relevant cancellations and completed stays have been reconciled, material price/inventory/channel changes are documented, limitations are stated and the public version has property approval if the property is named.
E4 supports a public case study describing what happened in that specific case. It does not automatically prove causal incremental lift.
E5 — Incremental / causal evidence
A result supported by a credible comparison design such as a control, holdout, randomized test, matched comparison or other defensible counterfactual. The design and limitations must be stated. LSG should not use E5 language merely because bookings increased after an intervention.
Attribution confidence
Separately from the evidence level, LSG records how confidently an observed outcome can be associated with an activity:
- Low: temporal association only or weak tracking.
- Medium: traceable campaign, sales or booking path with known limitations.
- High: reconciled identifiers and a clear attribution rule, while still distinguishing attribution from causation.
Public reporting rules
- Do not describe E1 signals as bookings.
- Do not describe E2 confirmed bookings as realized stays or realized revenue.
- Do not describe E3/E4 outcomes as incremental lift unless the comparison design supports E5.
- Always state the measurement date for provisional results.
- Record cancellations and material changes in price, inventory, restrictions or distribution.
- Negative and inconclusive tests remain useful evidence and should not be discarded from the private learning dataset.
- Named public case studies require property approval.
How this connects to LSG Intelligence
The evidence level is separate from the recommendation itself. LSG Intelligence can learn from structured outcomes only when the target used for learning has reached the maturity required for that target. Early signals can support operations, but they must not leak into recommendation-time model features as if they were final outcomes.
LSG Weak-Date Case Study Protocol → · Evidence Ledger → · Method Registry →