SpyderBot Documentation

Troubleshooting Overview

Use this page when a value, report, evidence item, integration, permission, or account action does not behave as expected and you are not sure which product owns the issue.


Start with the symptom

What you seeCheck firstContinue to
A Brand report value looks wrongScope, eligible population, valid responses, estimate label, and evidenceBrand Insights Troubleshooting
A response or citation is missingPrompt/system scope, execution state, time, evidence permission, and source availabilityPrompt Intelligence
Traffic is zero or suddenly dropsMeasurement Health, source state, freshness, filters, and true-zero versus unavailableData Availability
A report differs from the live productReport source, data cutoff, generated date, filters, and live updatesReports
Two periods cannot be comparedPrompt/system/source/method changes, timezone, incomplete periods, and compatibilityThe owning product's comparison guide
Evidence cannot be openedAccount access, report visibility, plan, source retention, and evidence restrictionAdministration
An integration is connected but emptyConnected/no-data state, test event, required capability, delay, and permissionsIntegrations
An estimate conflicts with first-party dataProvider/method scope versus connected source, period, geography, and freshnessReliability
An account, credit, or billing action is unavailableActive account, plan, and billing authorityAdministration
An API request failsApproved access, authentication, scope, version, and request IDEnterprise API Overview

Standard diagnostic checklist

  1. Confirm the active account, project, brand, context, or website.
  2. Confirm the selected date range and timezone.
  3. Review market, language, AI systems, prompts, competitors, connected sources, and filters.
  4. Identify whether the result is Observed, Calculated, Estimated, Verified, Classified, Inferred, Unavailable, or Insufficient.
  5. Check execution, report, integration, or Measurement Health status.
  6. Check data freshness and whether the current period is incomplete.
  7. Distinguish a true measured zero from missing, delayed, insufficient, or not-connected data.
  8. Open the closest available evidence or provenance record.
  9. Confirm comparison compatibility.
  10. Retry only when the product provides a safe retry path.

Common states

  • Processing or queued — work has not finished; do not treat incomplete output as final.
  • Partial — some eligible data is available, but coverage is incomplete.
  • Delayed or stale — latest expected data has not arrived within the approved freshness range.
  • No eligible data — data may exist, but none qualifies; this is not necessarily zero.
  • Insufficient — data exists but cannot support the requested interpretation.
  • True measured zero — required source and population are available, and the result is zero.
  • Not connected — required source is not connected.
  • Connected, no data — authorization exists, but qualifying events have not arrived.
  • Incompatible comparison — material scope differences prevent a safe trend conclusion.

Report a data or evidence issue

Use In-app Support with workspace or project, product page URL, identifiers, metric name, date range, filters, expected versus observed result, and a safe screenshot when needed.

Do not send passwords, API keys, session cookies, full payment-card information, or unnecessary personal data.