Zero-Result Field Note: One Store, One Measured Window
This is a field note, not a benchmark report: a single measured deployment, not an aggregate across merchants. It exists because the underlying numbers are real and worth sharing honestly — not because one store is enough to say what's "normal" for ecommerce search.
Why this isn't called a benchmark
A real benchmark needs data aggregated across many stores, verticals, and catalog sizes — enough to say what a "typical" or "healthy" number looks like. This field note reflects one specialty retailer over one three-month window. It is a real, verified data point, not a statistical claim about ecommerce search in general.
The numbers
39% of Wine House searches returned zero results, measured monthly across April through July 2026 — a consistent share over the period, not a one-week spike. Recovered zero-result queries went on to generate an estimated $5,700 in cart value and accounted for 26.5% of the store's revenue in the period measured.
Methodology
Figures were read from Semantix analytics for Wine House (pulled July 26, 2026, UTC). A search counts as zero-result when the recovered product click is tagged with the zero-result recovery source, or when the query is recorded in the zero-result discovery log. Add-to-carts are matched to a recovered search by the same session and exact query text within a short time window. Cart values are estimated from current catalog prices of the recovered item, since cart events do not store price at the time of the event. Catalog products, SKUs, and per-item pricing are excluded from this note.
What this does — and doesn't — support
It supports a specific, verified claim: at this one store, in this window, zero-result search was a large and consistent share of traffic, and recovering it produced a measurable share of revenue. It does not support a claim like "most ecommerce stores have a 39% zero-result rate" — that would require the aggregate data this note doesn't have.
What's next
As more deployments are measured the same way, this is the kind of field note that can eventually be aggregated into a real benchmark report — at that point with an honest sample size and methodology stated up front, the same way it is here for one store.