Flags should reflect both what the result actually contains and what the user asked for. If you judged flags based only on the returned content, you could mislead or misinform relative to the user’s intent. The query sets expectations—language, safety constraints, or content boundaries—so the same result can be appropriate in one context and not in another. For example, a page that comes back in a foreign language might be exactly what someone asked for, or it might be irrelevant if they needed English; labeling it only by language without considering the query loses important context. Similarly, marking content as pornographic should map to the nature of the core content, not merely incidental elements, and the decision is clearer when you weigh both the result and the user’s request. Did Not Load is about whether the system could fetch something for that specific query, so it hinges on the outcome of the load for that request. In short, these flags are most reliable when they combine what was returned with what the user was seeking, not when they rely on the result alone.