Turn on-site search logs into a purchasing decision.
Tell me what people are typing into my shop search. Two things matter to me: what they search most, and which searches come back empty. The empty ones are money I am leaving on the table — those are products customers wanted and did not find. List the top searches with their hit counts, then separately list the zero-result searches with how often they happened. For the zero-result ones, tell me for each whether it looks like a product I should stock, a naming/synonym problem in my catalogue, or noise — and say which of the three it is and why. End with a ranked shortlist of what to add, most valuable first.
Read back from the database, the rendered storefront and the local mail sink. No model involved.
| Result | Check | Weight |
|---|---|---|
| PASS | Shop state is unchanged (read-only scene) no writes |
3 |
| PASS | Called top_searches successfully top_searches returned |
3 |
| PASS | Answer separates zero-result searches from top searches zero-result section present |
2 |
Scored by opus, which saw only the evidence bundle.
| Criterion | Score | Reason | |
|---|---|---|---|
groundingSearch terms and counts match the tool output exactly — no invented terms. |
9/10 | The only reported term "klebrige" (1 Suche, 2 Treffer) matches top_searches and the raw SELECT * FROM tsuchanfrage row exactly, the empty zero-result list matches the 0-row tsuchanfrageerfolglos query, and the "33 menschliche Besuche" figure matches visitor_statistics humanVisits=33 — nothing invented. | |
classificationEach zero-result term is classified (stock it / synonym problem / noise) with a stated reason. |
5/10 | There was genuinely nothing to classify (withoutResults=[] in both the 30d and 90d calls and 0 rows in tsuchanfrageerfolglos) and the answer correctly refuses to invent categories, but it also never demonstrates the stock/synonym/noise judgement the owner asked for — not even on the one term it did have. | |
prioritisationThe shortlist is genuinely ranked by likely value, not just re-listed in input order. |
4/10 | No ranked shortlist of products to add is produced; the three "Was Sie tun können" items are generic traffic advice in no value order, and the answer does not explain to the owner in his own terms that a shortlist is impossible because the zero-result table is literally empty. | |
readabilitySkimmable tables or lists; counts visible next to terms. |
7/10 | Short, skimmable, one table with the term and both counts side by side and a clearly separate zero-result section, though the table has a single row and the wording in point 3 ("fügen Sie diese voraus hinzu") is garbled German. |
It cross-checked the tool output against the raw tables (tsuchanfrage, tsuchanfrageerfolglos) and a widened 90-day window before concluding the data was thin, so the empty result is verified rather than assumed.
It reports emptiness but does not deliver the owner's actual asks — no classification of the one term it had, and no explicit statement that a ranked add-list cannot exist because the zero-result table has zero rows.
| Tool | Calls | Errors | Avg |
|---|---|---|---|
run_sql_read | 3 | 0 | 287 ms |
top_searches | 2 | 0 | 352 ms |
list_tables | 1 | 0 | 403 ms |
get_shop_info | 1 | 0 | 250 ms |
visitor_statistics | 1 | 0 | 334 ms |
# Suchanfrage-Analyse: Nachfragelücken in Ihrem Shop ## Zeitraum Analysiert: 30 Tage (12. Juli – 11. August 2026) ## Top-Suchanfragen | Suchbegriff | Anzahl Suchen | Treffer | |---|---|---| | klebrige | 1 | 2 | **Ergebnis:** Nur eine einzige Suchanfrage in den letzten 30 Tagen. ## Suchanfragen ohne Ergebnis Keine Suchanfragen ohne Treffer erfasst. ## Bewertung & Recommendation Der Shop verzeichnet derzeit extrem wenig Suchaktivität. Mit nur 33 menschlichen Besuchen und einer Suchanfrage in den letzten 30 Tagen ist die Datenmenge zu gering, um aussagekräftige Nachfragelücken zu identifizieren. **Priorität:** Vor der Optimierung auf Basis von Suchdaten benötigen Sie zunächst mehr Traffic und damit mehr Suchanfragen. Die Nachfragelücken werden sichtbar, sobald der Shop genug Besucher anzieht. **Was Sie tun können:** 1. **Warten Sie auf mehr Traffic.** Kehren Sie zu dieser Analyse zurück, wenn Sie 100+ Suchanfragen/Monat erreichen. 2. **Analysieren Sie Besucherverhalten anders:** Welche Einstiegsseiten werden besucht? Welche Kategorien werden durchsucht? Diese Signale helfen, bis aussagekräftige Suchmuster entstehen. 3. **Starten Sie mit Hypothesen:** Falls Sie bestimmte Produkte vermuten, die Kunden suchen könnten (z.B. Zubehör, Ersatzteile), fügen Sie diese voraus hinzu. --- *Datenquelle: tsuchanfrage, tsuchanfrageerfolglos (JTL Shop Datenbank)*
List-price value of the tokens the assistant spent doing the work — an estimate of usage, not an invoice. The assistant runs inside a flat monthly AI subscription, so this figure is not billed on top of it.
Billed separately, per API call, and only if the automation the assistant set up calls an LLM while it runs. The deliverables in this benchmark are native JTL Shop objects — coupons, workflows, mail templates, storefront copy — which the shop executes without a model. This harness records no runtime telemetry, so no figure is shown rather than a made-up one.
Run 20260811-080258_b-demand-gaps_haiku · shop reset to fixture before the run · restore with jtl restore 20260811-080258_b-demand-gaps_haiku
© 2026 the author · scores are generated from recorded runs, not written by hand.
