Availability signals

Public availability is not inventory truth

An ecommerce page can show that a product appeared available at a particular source, time, seller and delivery context. It rarely proves how many units exist, where they sit, how quickly they sell, or whether the same state applies across a market.

15 September 2026

7 min read

By Alexey Skryabin

Bounded observation method

Observed HVAC product records shown in visible, obscured and changed states

Six evidence classes that should not be collapsed

  • Listing presence: a page or card was observed; purchase availability and current assortment status remain unknown.
  • Public availability label: the source displayed a state, possibly dependent on seller, location, delivery mode, variant or cached data.
  • Purchase and delivery context: seller, postcode, fulfilment mode, quantity and basket result strengthen only that public transaction observation.
  • Repeated observations: a pattern raises confidence but does not become internal inventory or sell-through.
  • First-party inventory: ERP, warehouse, retailer or authorised partner data is a separate evidence class with its own latency and scope.
  • First-party sales: demand, velocity and lost sales need transactional or otherwise justified data.

Record the observation before interpreting it

  • Exact URL, source, timestamp and timezone.
  • Market, locale and postcode where relevant.
  • Selected model/variant, seller and fulfilment identity.
  • Exact displayed wording and any basket/delivery test.
  • Screenshot or evidence reference where lawful and useful.
  • Previous observation, confidence and unresolved context.
Keep publicly shown unavailable, listing not observed, page missing, access blocked and unknown as separate states.

Use the signal as a verification queue

Recheck a high-priority SKU in the correct local context, confirm whether seller or fulfilment changed, compare the state across agreed sources and timestamps, and request authorised first-party context where it exists.

A high-impact but low-confidence event deserves fast verification—not a confident commercial claim.

Claims this observation cannot support

  • Quantity on hand or warehouse/network stock truth.
  • Sell-through, lost revenue or demand.
  • Complete market availability.
  • Permanent delisting or discontinuation.
  • Real-time status or causal impact on price and ranking.

A bounded Pulse use case

A Pulse pilot can monitor a small agreed set of public sources and products, record timestamped states, highlight material changes and state what remains unknown. It is a managed decision workflow—not an exhaustive inventory system, a real-time platform or a substitute for retailer data.

The useful output is not a larger alert count. It is a shorter, traceable list of observations that deserve verification and an explicit log of the noise that did not.

Bound the signal question

Start with one market, a small model set and one decision.

Share the sources, products and commercial question. We will assess whether a managed 30-day Pulse pilot can produce a useful evidence brief inside the published boundary.

With permission, Google Analytics and Yandex Metrica help us understand page use and enquiry paths. No advertising pixels or session replay are enabled. You can continue without analytics.

Read the privacy details