First isolate client, network or platform
Record the time, exact steps and complete message or screenshot before reinstalling. Test whether the same account can sign in and submit through the web console. If the web works but the client fails, inspect local environment, version and permissions. If both fail, account state, maintenance or an upstream outage is more likely. Isolating the side prevents repeated blind steps.
Sign-in and connection failures
For timeouts, certificate warnings and proxy errors, confirm that the official site opens, then inspect system proxy, VPN and corporate firewall rules. Compare with a phone hotspot to identify local-network policy. Verify credentials and second-factor freshness and whether simultaneous remote sign-ins triggered a security restriction. Upgrade an outdated client from the official source because interface changes can break handshakes. If the message says unavailable, check maintenance notices before retrying.
Import failures and stalled jobs
Follow documented file rules for one record per line, numeric notation and country codes. Hidden characters, spreadsheet scientific notation and whitespace can break parsing. For a job stuck in queued or processing, verify balance, quota and per-job limits. Submit a small sample to separate bad records from service congestion. Record the task ID and start time and ask support after the documented SLA rather than repeatedly cancelling and resubmitting.
Empty, mismatched or abnormal results
Compare submitted, successful and failed counts with billing before downloading. An empty file may indicate an unfinished job, overly restrictive rules or the wrong batch. Read status definitions so unknown is not interpreted as disconnected. If one number changes across runs, compare the elapsed time, portability or suspension events, and detection channel or rule version. Retain source and both outputs for a reproducible comparison.
Local environment and support handoff
For crashes or garbled UI, test required OS and runtime versions, permissions and conflicts with endpoint-security scanning; clear only the documented client cache before signing in again. If unresolved, send client and OS versions, network type, task ID, screenshots and a minimal reproducible sample such as ten test records. A complete handoff shortens diagnosis and repeated questions.
