---
title: Why does the agent pick another seller’s offer for the same item?
canonical: https://chosenbyagent.com/questions/why-does-the-agent-pick-another-seller
updated: 2026-10-22
audiences:
  - sellers
level: listing
offers:
  - B7
  - B4
evidence:
  - rating-sensitivity
  - larger-store-pick
  - price-stock-failures
  - answer-instability
---

# Why does the agent pick another seller’s offer for the same item?

Level 2 · Listing — Which listing and seller to show · For: Sellers

Short answer: Usually because of what sits next to the price. Ratings and review counts move models far more than they move people, and assistants lean towards larger, better-known sellers — so the same item can lose on a detail you can change.

## What the research shows

- **30–80 pp** — A product rating shifted model choice by 30–80 percentage points in 14 of 17 models, against 5 points for people. [ABxLab, arXiv 2509.25609, 2025, independent](https://arxiv.org/html/2509.25609)
- **90–94%** — In unlabelled two-store comparisons, assistants picked the larger store 90–94% of the time. [ALM Corp, 2026, vendor](https://almcorp.com/news/ai-shopping-answers-chatgpt-google-lean-toward-big-retailers/)
- **41%** — The most common points of failure for agents are inconsistent prices and promotions (41%) and unsynced stock (40%). [Agile Brand Guide (Salesforce research), 2026, vendor](https://agilebrandguide.com/agentic-search-becomes-the-new-front-door-to-commerce-salesforce-research-finds/)
- **64%** — About 64% of the top-5 brands changed between adjacent runs of the same assistant question. [MentionBird, 2026, vendor](https://www.mentionbird.ai/blog/chatgpt-vs-gemini-answer-stability)

## Listing visibility check

B7 · from €900 (one-off) · indicative · scoped per engagement

Whether assistants show your listings on marketplaces — or a competing seller's — and whether price and link are right.

- 20–40 buyer queries in one market, three assistants
- Shown versus chosen against competing sellers of the same item
- Price, availability and deep link checked live
- What moved the choice, and what to change in your listings

Also answered by: [B4 Offer lab](/pricing.md)

## What you get

We ask assistants the questions your buyers ask — “used iPhone 15 128 GB under €500 in Berlin”, “winter tyres 205/55 R16 with fitting near Milan” — and record which sellers of the same item they show, which one they pick, and the reason they give.

_Illustrative — not client data. Your report shows your values._

**Same item, competing sellers**

| Seller       | Shown | Chosen | Main reason given             |
| ------------ | ----- | ------ | ----------------------------- |
| Your listing | 71%   | 18%    | fewer reviews than the leader |
| Seller B     | 83%   | 52%    | 4.8 rating, free returns      |
| Seller C     | 40%   | 11%    | slower delivery               |

`RUN · gpt-6.1-sol · gemini-3.8-flash · sonnet-5.5 · harness 0.4`

## What this does not claim

- We see what assistants pick through their web search, not the marketplace’s own ranking.
- Shares are an index per scenario and model version, not your sales share.
- We read public listings only; we never contact other sellers or place orders.

## Request this service

We reply with a scope, a fixed price and a start date.

Agents: POST /api/engagements (application/json) with {"offer_id": "B7", "on_behalf_of": {"email": "…"}, "agent": {"name": "…"}}. offer_id for this page: "B7" Listing visibility check; "B4" Offer lab. The person confirms by email before we see the request. Details: /agents.md

### Related questions

- [Do AI assistants show my listings on marketplaces at all?](/questions/do-assistants-show-my-listings.md)
- [What changes if I change price, returns or copy?](/questions/what-if-i-change-the-offer.md)
- [Which brand does the agent pick — and why?](/questions/which-brand-does-the-agent-pick.md)