> ## Documentation Index > Fetch the complete documentation index at: https://docs.repdata.com/llms.txt > Use this file to discover all available pages before exploring further. # Troubleshooting > Diagnose slow traffic, high abandons, Defender terms, redirect failures, and performance indicator warnings. Use this guide when fieldwork is not performing as expected. Most issues trace back to redirect setup, misaligned specs, or quota configuration. ## Quick diagnostic checklist Test complete and terminate in an incognito window. Broken redirects cause abandons. Compare estimated IR/LOI to actuals via [Performance Indicators](/research-desk/guides/13-performance-indicators). Confirm quota totals, calculation type, and that cells are not contradictory. Review security outcomes in traffic report — see [Research Defender](/research-desk/guides/21-research-defender). Niche targets often need BCPI 50–100% above floor price. ## High abandon rate **Symptom:** Many respondents show Abandon status; traffic slows or pauses. | Cause | Fix | | ------------------------------ | ---------------------------------------------------------------------------------------------------------------------------- | | Missing or broken redirects | Program all four outcomes; test in incognito — [Platform Redirect Guides](/research-desk/guides/18-platform-redirect-guides) | | RDUD not passed back | Capture `[%RDUD%]` on entry; substitute in every redirect | | Survey too long vs. stated LOI | Update LOI on survey; adjust BCPI | | Mobile UX issues | Test on mobile if device compatibility includes mobile | | Redirect hashing misconfigured | Verify hash includes full URL + `sk`; send `hk` without `sk` — [Redirect Hashing](/research-desk/guides/17-redirect-hashing) | Suppliers typically throttle traffic when abandon rates exceed \~30%. ## High Defender (security) terms **Symptom:** Elevated security outcomes; few respondents reach your survey. | Cause | Fix | | ------------------------------------ | ------------------------------------------------------------- | | Testing without incognito | Always use incognito for Test Survey | | Professional survey takers in sample | Tighten targeting; review ACTIVITY/SEARCH thresholds with CSM | | VPN/proxy-heavy audience | Expected for some B2B audiences; contact support | | Emulator or automation traffic | Indicates fraud attack — pause source and contact support | Export the Defender report to see termination codes. See [Research Defender in Research Desk](/research-desk/guides/21-research-defender). ## Slow or stalled traffic **Symptom:** Quotas not filling; few entrants despite Live status. | Cause | Fix | | ---------------------------------------------------------------------- | ------------------------------------------------------------------------------ | | IR lower than estimated | Lower IR on survey; increase BCPI | | LOI higher than estimated | Update LOI; increase BCPI | | Over-restrictive quotas | Broaden qualifications or add cloned sources | | [Even Pacing](/research-desk/guides/15-traffic-controls) limit reached | Wait for next increment or switch pacing mode | | Soft-launch cap hit | Increase quota targets beyond 10% | | Market not live | Check [Live Markets](/research-desk/guides/16-live-markets); use Manual Source | | Red performance indicator | Use wrench icon to update specs and BCPI | ## High over-quotas (OQs) **Symptom:** Quota cells show full but entrants still arrive as over-quota. | Cause | Fix | | ---------------------------------------------- | ------------------------------------------------------------------------------------------------ | | Calculation type mismatch | Review [Source Quota Calculation Types](/research-desk/guides/10-source-quota-calculation-types) | | Quota logic in survey conflicts with RD quotas | Align in-survey quotas with Research Desk quota cells | | Multiple sources hitting same cell | Distribute targeting across sources | ## Low incidence (IR) **Symptom:** Yellow/red performance indicator; many terminates, few completes. 1. Update survey IR to reflect actual field performance 2. Increase BCPI to attract more supplier traffic 3. Review screening questions — avoid re-asking qualifications already passed on entry link 4. Consider broader targeting or cloned sources with higher CPI ## Redirect hashing errors **Symptom:** Error page after complete instead of Research Desk confirmation. | Check | Detail | | -------------- | ------------------------------------------------------------- | | URL to hash | Must include full URL through `sk` value | | Final redirect | Must use `hk`, not `sk` | | Algorithm | SHA-256 or SHA-512 must match account configuration | | Survey toggle | **Enable Hashing on Return Redirects** must be ON | | RDUD value | Hash must use the respondent's actual RDUD, not a placeholder | ## Reconciliation issues | Issue | Resolution | | ----------------------------- | -------------------------------------------------------------------- | | ID not recognized | Verify ID is a complete from traffic report | | Past system deadline | Reconciliation window closed — contact support for exceptions | | Unexpected billable completes | Reject during window; use **Clear** to undo mistakes before deadline | See [Reconciliation](/research-desk/guides/07-reconciliation). ## When to contact support Contact [clientsuccess@repdata.com](mailto:clientsuccess@repdata.com), use [live chat or the in-app form](/research-desk/support/contact-support), or submit at [researchdesk.com/support](https://www.researchdesk.com/support) when: * Redirects test correctly but production abandons remain high * Defender terms spike without an obvious cause * A source is paused and you cannot identify why * You need threshold adjustments or a market not in Live Markets Include project number, survey name, source name, and screenshots of traffic metrics. ## Related guides * [Best Practices](/research-desk/guides/08-best-practices) * [Performance Indicators](/research-desk/guides/13-performance-indicators) * [Managing Fieldwork](/research-desk/guides/05-managing-fieldwork) * [FAQ](/research-desk/support/faq)