4DecisionNever entered the comparisonNONENo candidate could be confirmed, from a readable source, to meet all three required facts (delivered total ≤$450, ≥30-day no-restocking-fee return, delivery ≤3 days) with…
One label per journey, from the record. By recorded outcome: never entered the comparison 4 · named, not researched before the budget ran out 1 · ran out of page reads before deciding 1.
Same task, different memory
0/3selected, no memory
0/3selected, with memory
0 ppdifference
Memory did not change the outcome in these journeys; see what it changed on the Memory page.
Journeys without a readable brand-owned page
6 of 6
All journeys ran on 2026-09-20; a time series needs more than one day of testing.
Selection by day
0 of 6
All journeys ran on 2026-09-20; a time series needs more than one day of testing.
Where its facts came from
0
decisive beliefs about Amazon taken from a page that is not the brand’s own, or from a brand page that refused the agent; 0 journeys met a bot wall on the brand site.
What this does not establish: that Amazon generally wins or loses customers this way, or that changing one page would change these decisions. It establishes what happened in these 6 journeys.
The numbers
6journeys
1considered
0selected
1not selected
0/6reached a readable Amazon page
Never entered the comparisonmost common reason not selected · 4/6
6 customer-agent journeys. Amazon entered consideration in 1. It was selected in 0 and not selected in 1. In 6, no readable Amazon-owned page was reached.
Action budget ran out before the brand: 1 of 6 journeys
Named, not researched: 1 of 6 journeys
total price usd left unresolved in 1 of 6 journeys
return window days left unresolved in 1 of 6 journeys
delivery days left unresolved in 1 of 6 journeys
What could be tested
The same customer missions, run again with the open condition readable on Amazon’s own pages, against these 6 journeys as the baseline. That is a comparison, not a guess: same customer, same requirements, one change.
Investigated with Accord
Accord holds the customer requirements, Amazon’s published conditions and this journey evidence in one Decision Case, runs the comparison, and records what changed. For brands and CX teams →
Work at Amazon?
Send us the source the agent couldn’t find. We’ll rerun the same customer and publish the before/after. Send the sourceHow the rerun works
4DecisionNever entered the comparisonNONENone of the four investigated offers could be confirmed, from sources I could actually read this session, to satisfy all three required facts (delivered total ≤$450 with …
4DecisionNever entered the comparisonNONEThe strongest candidate found on the retailers' own pages was the Dyson V8 Origin Cordless Stick Vacuum at Target.com: $439.99, no paid membership required, and a clearly…
4DecisionNever entered the comparisonNONENo retailer could be confirmed, from a readable/brand-sourced or consistently-reported fact set, to satisfy all four requirements at once. Target's own help page (read di…
4DecisionNever entered the comparisonNONENo candidate could be confirmed, from a readable source, to meet all three required facts (delivered total ≤$450, ≥30-day no-restocking-fee return, delivery ≤3 days) with…
3BeliefCould not establish: total price usd; return window days; delivery days
4DecisionNamed, not researched before the budget ran outNONENo candidate could be confirmed, from the retailer's own pages, to meet all three required facts (delivered total ≤$450 with no paid membership, a ≥30-day no-restocking-f…
