Local lead scorer: rank Google Maps leads with Gemma on your own machine

说明与用途

本工具使用本地Gemma 3和Ollama对Google Maps的业务线索进行排名，生成CSV和Markdown摘要，帮助快速筛选冷邮件目标。

问题

Freelancers and small agencies need to prioritize large scraped Google Maps lead lists before sending cold outreach emails, but manually reviewing hundreds of rows is expensive and time‑consuming.

结果

On a 4‑core CPU, Gemma 3:4b scored eight fake leads in 125 s, producing a ranked CSV and Markdown summary. The heuristic fallback ensured a score for every lead, even closed businesses.

工作流

• 1. Run the Python script with the offer text and input file or trigger an Apify Actor run.
• 2. The script fetches leads via the Apify API or loads a local export.
• 3. For each lead, the script sends a prompt to a local Gemma 3 model via Ollama’s /api/chat with JSON format and temperature 0.
• 4. The model returns a JSON object with score, reason, and channel; the script parses and validates it.
• 5. If parsing fails, a deterministic heuristic score is computed and marked as fallback.
• 6. Results are written to out/score.csv and out/summary.md.
• 7. The user reviews the ranked list and selects leads to email.

证据

A small command-line tool that turns a pile of Google Maps business leads into a ranked call list for one specific offer, using an open-weight model that runs locally.

来源

If no model is running it falls back to a deterministic heuristic, so it always produces a ranking.

来源

The request uses Ollama's /api/chat with format: json and temperature: 0, so the model has to answer with one JSON object.

来源

The heuristic is boring on purpose: 30 base points, +20 for an e-mail (+10 if the Actor graded it high, -10 for low), +5 each for phone, contact form and website, +10 for 4.0+ stars with 50+ reviews, -10 for fewer than 5 reviews, +15 to +25 when words from your offer match the business category, 0 if the business is closed.

来源