The problem with filter UIs
Most B2B email finders (Apollo, Hunter, ZoomInfo, Lusha, Cognism) ship a filter sidebar with 6-12 dropdowns: industry → sub-industry → company size → role → seniority → location → funding stage → tech stack → headcount growth → and so on. To build a list, you click through every one.
In practice, this means:
- you have to know each tool's category taxonomy
- "industry" lists in different tools don't match (one calls it "Mining & Metals", another "Materials")
- you can't describe a niche cleanly ("VPs of operations at copper mines in Chile" doesn't map to any standard filter)
- every list you build takes 5-10 minutes of clicking before you even hit search
What plain-English search changes
I built findmemail.io because I was tired of the click-through workflow. The core idea: type the ICP in one line, get verified emails back.
Examples that work today:
- "VPs of growth at Series A SaaS in NYC"
- "founders of seed-stage climate tech in Berlin"
- "CMOs at consumer SaaS with 50-200 employees"
- "agency owners doing performance marketing in India"
The query gets parsed into structured filters server-side, matched against the company graph, then SMTP-verified emails are returned.
What I learned shipping it
A few things were surprising:
1. Most queries don't need a filter UI at all. When users could type freely, ~80% of queries fell into 6-7 patterns: founder/CEO + stage + industry, VP/Director + function + region, etc. The filter UI was solving for a long tail of edge cases that almost nobody hit.
2. SMTP verification matters more than database size. A list of 1,000 verified emails outperforms 10,000 catch-alls — bounce rate kills sender reputation, and once you're flagged, every campaign suffers. We index 32k+ companies but the value is the per-email verification.
3. Pricing models are the real differentiator. Most tools sell credits + monthly subscriptions. Indie founders hate subscriptions and want predictable cost. We ship a $200 lifetime tier with 7-day refund — no recurring billing, no per-seat. Conversion on this tier outperformed monthly 3:1 in early months.
Stack notes for builders
- semantic parsing of the query: small fine-tuned model + structured filter extraction
- company graph: combined from public registries, LinkedIn-style sources, and crawled domains
- email finding: pattern enumeration + MX/SMTP probe per pattern
- verification: throwaway sender + RCPT TO check, no DNS-only "valid format" checks
Why the build-in-public version
I'm sharing the architecture and lessons because most "B2B leadgen" content online is recycled SaaS marketing. Indie founders shipping in this space need actual signal, not best-of lists.
If you're shipping a leadgen workflow and want to try plain-English search instead of filters: findmemail.io — free tier on signup, 50 credits to play with.
Happy to answer architecture questions in the comments.
Top comments (0)