Background
The Setup
Before filtrat, the stack was MillionVerifier for first-line email verification and Scrubby for catch-all resolution as a second pass. The system worked but Scrubby was slow (1.5 hour queue) and expensive.
The test: run filtrat against real sending data with known bounce outcomes.
Methodology
The Tests
What was tested across 2,000+ real emails:
Catch-all emails
334 emails — head-to-head vs Scrubby
Known hard bounces
814 emails — with known bounce outcomes (ground truth)
General inbox
500 general-inbox emails
Personal B2B
800 emails — vs MillionVerifier
Head-to-Head
Catch-All Comparison
| Metric | Filtrat | Scrubby |
|---|---|---|
| Bounce rate on "valid" verdicts | 3.64% | 3.69% |
| Processing time | Minutes | 1.5 hours |
| Per-email reasoning | Yes | Binary only |
| Price | Fraction | Full plan |
Matched
Scrubby's catch-all accuracy
0x faster
Minutes vs 1.5 hour queue
~0% more
Sendable leads recovered
Fraction
Of Scrubby's cost
The Switch
Before / After
Before
Scrubby
1.5 hour queue
Full plan price
Binary verdicts only
After
Filtrat
Minutes
Fraction of the cost
Per-email reasoning
Result
~7% more
Sendable leads recovered
Lower total cost
Subscribed. In production.
Results
Key Numbers
| Emails tested | 2,000+ |
| Catch-all accuracy vs Scrubby | Matched (3.64% vs 3.69%) |
| Speed improvement | 20x |
| Additional sendable leads | ~7% |
| Per-email reasoning | Yes |
| Outcome | Subscribed. In production. |
Takeaway
The Bottom Line
Filtrat matched Scrubby's catch-all accuracy at 20x the speed and a fraction of the cost. A high-volume sender tested it on 2,000+ real emails with actual bounce data, put it into production, and subscribed.

