Files
QR-master/tasks/lessons.md
2026-05-27 14:33:58 +02:00

974 B

Lesson: For lead scraping, do not rely only on pre-enrichment dedupe. Website crawling can assign the same corporate email to multiple locations, so exports must be deduplicated again by final email before use.

Lesson: Large API scraping runs should write incremental output or use smaller controlled batches. A long Overpass workflow can hang or rate-limit without producing files, making it hard to recover useful partial results.

Lesson: For bulk email pre-validation, avoid synchronous per-domain PowerShell DNS checks because slow domains can stall the whole run. Use a concurrent DNS checker with explicit per-query timeouts and write separate valid, unknown, and invalid reports.

Lesson: DNS/MX-valid is not enough for Smartlead-quality lead uploads. Calibrate high-confidence exports against Smartlead feedback by domain; in the first two batches, gmail.com-style domains were far more reliable than large hotel-chain domains even when both had valid MX records.