Case study: a dedup scrubber cut rejected leads from 21% to 6%

Case study: a dedup scrubber cut rejected leads from 21% to 6%

What it is: A lead seller kept getting batches rejected by the broker for duplicates against its existing user base. A pre-delivery scrubber checked leads against a hashed suppression list before sending.

Best for: Lead aggregators delivering to brokers that reject and don't pay for known users.

What was done:
— Built a hashed email+phone suppression list from past delivered leads (62,000 records)
— Ran every new batch through the scrubber before API delivery
— Added the broker's returned-rejection feed back into the suppression list weekly

Outcome:
— Broker rejection rate fell from 21% to 6% over 5 weeks
— Effective paid-lead yield per batch rose ~19%
— Stopped a near-miss account suspension from a high duplicate ratio

Pros:
— Hashing kept PII out of the suppression store
— Feeding rejections back made the list self-improving

Cons:
— Cross-broker dedup is impossible without their data; this is per-broker
— Phone-format normalization bugs caused early misses until standardized to E.164

Who should skip this: Sellers with one-time exclusive leads and no recurring delivery relationship.

Verdict: Scrub before you send.