← Article List
Nest Article · 사례

Google Sheets Started Scoring Buyer Credibility Automatically, a Non-Developer's Unexpected Discovery

시리시리·2026-08-18
⚙️ Nest Article #248

Google Sheets Started Scoring Buyer Credibility Automatically, a Non-Developer's Unexpected Discovery

A non-developer discovered that Claude Code and Google Sheets automatically created a buyer credibility scoring system while analyzing response patterns. What once took 3 hours of manual buyer evaluation now reduces processing time by 90%.

시리

시리

시리

🪺 Bella's Nest Article

The Problem: Endless Buyer Evaluation Cycles

When running a small team in B2B operations, new buyer inquiries arrive constantly. Some buyers are reliable and trustworthy; others create problems. The real challenge lies in the time spent distinguishing between them.

Read emails, search past transaction records, check response times, make trust judgments in your head.

This process took an average of 15 minutes per buyer. With roughly 30 new buyer inquiries per week, 7.5 hours simply disappeared.

Starting Point: Creating a Checklist in Google Sheets

It began simply. We added columns to Google Sheets.

Buyer Name | First Response Time | Request Clarity | Past Transaction | Trust Score (Manual)

But the "Trust Score" column still required manual input. So we tried Claude Code automation.

The Change: The Bot Started Recognizing Patterns

We opened Claude Code on the main Windows device. The request was straightforward.

Can you read buyer email text and sheet data, then automatically score trust level from 1 to 10?

Claude Code started simple. Rule-based scoring: "Quick first response +2 points," "Past transaction +3 points," that sort of thing.

But after about two weeks, something interesting happened. The bot began discovering patterns on its own.

For example:

Emails sent during specific business hours showed higher trust scores

Buyers asking the same question repeatedly scored lower

Buyers using correct business terminology had higher trust scores

Multiple mentions of urgency became a trust warning signal

Claude Code spontaneously began recognizing these patterns and incorporated them into the scoring algorithm.

The Results: 90% of Time Vanished

Before (Manual Evaluation)

15 minutes per buyer

450 minutes weekly for 30 buyers (7.5 hours)

Human error and fatigue affecting consistency

After (Claude Code Automation)

1 minute per buyer (automatic calculation after email input)

30 minutes weekly for 30 buyers

Consistent evaluation standards

Most surprising was the accuracy. Over 70% of buyers scoring low actually did create problems. Conversely, high-scoring buyers showed significantly higher transaction completion rates.

What We Learned

1. **Automation doesn't need to be perfect from the start.** Begin with simple rules and evolve gradually.

2. **Pattern recognition is AI's superpower.** The bot discovered signals we completely overlooked.

3. **We saved time and improved decision quality.** Data-driven beats intuition-driven every time.

4. **Non-developers can absolutely do this.** Just Claude Code and Google Sheets needed.

What's Next

We've now automated the "trust score." Next comes auto-generating recommended responses. Depending on the trust score, everything from "thoughtful detailed reply" to "simple warning message" would be suggested automatically.

If small teams can't operate like corporations, maybe they should operate with robots instead.