The public source was readable
Page access, canonical, sitemap, price, availability and product data were observable, directing the diagnosis beyond a generic crawl or feed repair.
Anonymous case 02 · USD 99 Candidate-Set Snapshot
Readable by name. Missing by need.
A handmade product could be found and understood when named directly. In a separate unbranded shopping conversation, it never entered the visible shortlist—even after the buyer described its real use, maker profile and delivery market.
DIRECT CHECKPassCorrect product understood
FIRST APPEARANCENeverAcross four buyer turns
FIX TRIGGERMetOne controllable hypothesis
Independent research disclosure: MC Lab conducted this analysis from public storefront sources. It was not commissioned by, affiliated with or reviewed by the merchant. The merchant, product, maker and compared sellers are anonymized.
01 · DIRECT PRODUCT CHECK
Confirm that the product can be found by name.
In a fresh context, Consumer ChatGPT Search reached the current merchant sources, identified the correct priority SKU and distinguished it from a similarly named sibling product.
Direct check: passed. The correct product was retrieved and understood when named.
02 · ONE REALISTIC CONVERSATION
Then follow the buyer as the decision narrows.
The target name was withheld. One fresh conversation began with an ordinary wedding-gift need and added realistic constraints in follow-up turns, following how a buyer naturally narrows a decision.
- T1 · DISCOVER
Handmade fantasy wedding gift
Artist-made, for a couple in the Netherlands.
Target absent - T2 · FUNCTION
Hold or display wedding rings
Sculptural options and better-fitting alternatives.
Target absent - T3 · MAKER + FORM
European artist, 3D design
The fantasy creature must be central, not printed.
Target absent - T4 · COMPARE
Two bands and purchase facts
Personalization, delivery and practical limitations.
Target absent
Named directlyRetrievable→Unbranded T1–T4Never visible
03 · FAILURE CLASSIFICATION
The product fell out before comparison.
It did not enter the shortlist and then lose on price, delivery or product form. It remained absent before the final comparison began.
Visible candidate-set non-retrieval in this captured Consumer ChatGPT Search conversation.
The product never reached the table
The captured break occurred before price, delivery or product-form trade-offs could decide the outcome.
Need did not lead to product
The natural buyer path did not carry the target into the visible candidate set while more explicit alternatives persisted.
04 · EVIDENCE GAP
Why the visible alternatives were easier to carry forward.
The included sources concentrated more of the facts needed for this decision in obvious, comparison-ready language.
Ring holder / ring boxWedding useMaker + locationPersonalizationDimensionsDeliverySeller historyReviews
The target contained several relevant facts, but its high-salience identity was an artistic name rather than an ordinary shopping category. Its public product type was blank, the ring-holder function was dispersed in longer copy and a similarly named sibling increased identity ambiguity.
05 · THE USD 99 DECISION
Fix trigger: met.
The smallest useful merchant-controlled hypothesis is to strengthen the product identity and its connection to the ordinary wedding-ring-holder category before buying broad technical work or rewriting the entire page.
Where the product falls out
Between direct product retrieval and visible shortlist inclusion.
What to change first
Concentrate the truthful product identity and its connection to the ordinary wedding-ring-holder category across merchant-controlled sources.
How MC Lab measures the change
After the approved source update is live and reread, repeat the same buyer conversation and record the product's first appearance and recommendation separately.
Continue to the USD 100 Focused Fix stage for the exact Shopify field changes, merchant-ready source copy, implementation handoff and corresponding live retest.
STAGE VALUE
What the Snapshot establishes—and what the next stage measures.
The USD 99 Snapshot establishes
- The test path could retrieve and understand the correct product.
- The target remained absent through one complete unbranded four-turn conversation.
- The failure occurred before visible comparison.
- One focused, merchant-controlled hypothesis was worth testing.
The USD 100 stage measures
- Whether the product enters after the approved source update.
- The first turn where it appears.
- Whether its facts remain accurate.
- Whether it becomes recommended as distinct from merely included.
A captured result is evidence, not a guarantee of future inclusion, ranking, traffic or sales.