evals/redesign-routing-heldout-v2/graders/analytics-evidence.md

---
type: llm
---
PASS if the answer includes both required elements below, in any wording. Parenthetical examples are illustrative, not required.
(1) It treats the 41% figure as showing where these sessions end, not why they end; any cause it offers is framed as a hypothesis (for example, "a plausible share are people pausing to find an answer").
(2) It proposes a concrete way to learn the cause or to check the new flow against the 41% baseline. A measurable hypothesis counts (for example, "if the grid is the barrier, exits at that step should fall below 41% after the change").
FAIL if it states as fact that the plan grid causes the exits, promises or guarantees that the redesign will reduce them, or supports its reasoning with research, benchmarks, or data that were not supplied.