# Cold Outreach Email Drip: Pulpie Web Content Extraction

## Email 1: Hook
**Subject:** One thing tanking your data pipeline costs

Hey [Name],

We just helped a data eng team cut their web extraction spend from $8K to $400/month.

Turns out most extraction APIs are built backwards: they generate output token-by-token, forcing you to read the entire model for each token. It's like paying per-keystroke instead of per-word.

Our encoder-based approach runs one forward pass, labels what's content vs. boilerplate, done. 20x cheaper. Same quality.

If you're processing thousands of webpages monthly, we should talk.

[View pricing]

---

## Email 2: Social Proof
**Subject:** How we got cost down 95%

Hi [Name],

Quick follow-up on that extraction cost problem.

The industry standard (Dripper) charges $159K to extract 1 billion pages. We do it for $7.9K.

The math:
- Decoder models: read full model per token = memory-bound bottleneck
- Pulpie: read full HTML once, classify blocks = compute-bound + parallelizable

Result: SOTA extraction quality at 1/20th the cost.

Free API key included with sign-up.

[Start free]

---

## Email 3: Urgency
**Subject:** Your extraction budget can be cut in half

Hey [Name],

We're expanding capacity this month only.

If you process >100K pages/month, you qualify for our volume tier. We're locking in founder rates through end of July.

One customer went from $8K/month to $380. Another from $15K to $650.

The gap closes after this window.

[Reserve your rate]

---

## Email 4: Specificity
**Subject:** Built for your stack

Hi [Name],

Pulpie works with:
- Airflow DAGs (direct HTTP calls)
- Snowflake pipelines (batch ingestion)
- Real-time firehose (our queue handles spikes)

No refactor needed. Drop-in replacement for existing extraction.

We support HTML-to-Markdown (your format), HTML passthrough, and structured JSON output.

Technical docs: [link]

---

## Email 5: Close
**Subject:** One 15-min call?

Hey [Name],

If you're serious about cutting extraction costs, let's lock 15 min this week.

I'll show you the ROI on your specific volume, answer architecture questions, and we can get you live on day 1.

[Book time here]

Or just spin up a free API key and run 1000 test URLs. We handle the proof.

Thanks,
Feyn Team