Direct questions · No. 02 / 31
2. Why does ChatGPT recommend our competitors but not us?
Last updated
Treat the omission as a diagnostic question, not proof that your company has an authority problem. First establish whether the exclusion recurs under comparable conditions and whether the question accurately describes a customer you serve.
Repeated-query research from SparkToro and Gumshoe found substantial variation in recommendation lists, their order and their length. That makes a single screenshot weak evidence for commissioning a remediation programme. (sparktoro.com)
Investigate these possibilities separately:
| Possible explanation—not an established cause | What to inspect | Sensible next action |
|---|---|---|
| Your product does not satisfy the request | Required integrations, company size, geography, budget and deployment model | Confirm product fit before changing content |
| The question describes the wrong category | Customer language versus internal positioning | Rewrite the test around validated buying needs |
| Relevant information is inaccessible | Important pages, indexing status, crawler access and security controls | Commission a technical check |
| Product evidence is missing or unclear | Pricing, integration, migration and security documentation | Improve the specific missing explanation |
| External sources contain incomplete information | Exact cited pages and their publication dates | Request factual corrections where appropriate |
| The result is unstable | Repeated answers across dates and fresh sessions | Collect more evidence before spending |
A useful diagnosis might be: “We recur in general category answers but disappear when the buyer requires a particular integration.” That is more actionable than “our GEO score is low”.
Sources cited on this page
- sparktoro.com sparktoro.com/blog/new-research-ais-are-highly-inconsistent-when-recommending-br