← backData & AnalyticsFounding · Free forever

Analyze a spreadsheet and extract insights

Ask for 'non-obvious' patterns — otherwise you get averages you already knew.

by @whitney · recommended tool Claude Co-Work · 0 unlocks

You'll need

CLAUDE SKILLS

xlsx

CONNECTORS

Google Drive

EXAMPLES · WHAT THIS PROMPT PRODUCES

Sample output

**Analysis — q2_subscriptions.csv (4,812 rows)**

**1. Schema & data quality**
12 columns: user_id, plan, mrr, signup_date, churn_date, region, channel, etc. Issues found: 214 blank `region` values, `mrr` stored as text with "$" prefixes, 3 duplicate user_ids, and `signup_date` in two formats (MM/DD/YY and ISO).

**2. Metrics computed**
- MRR: **$184,200** | Active subs: 3,690 | ARPU: **$49.92**
- Gross monthly churn: **5.8%** | Net revenue retention: **97%**

**3. Five non-obvious patterns**
- Annual-plan users churn at 1.9% vs. 7.4% monthly — but **40% of annual churn clusters in month 11** (renewal-anxiety signal).
- The "webinar" channel has the lowest CAC-adjacent churn despite middling volume.
- APAC region shows 2x expansion revenue but 1.4x churn — high-variance segment.
- 18 accounts contribute 22% of MRR (concentration risk).
- Users who upgrade within 14 days have 3x lifetime retention.

**4. Three follow-up analyses**
Cohort retention curves by channel; expansion vs. contraction by plan; a churn-prediction model using the 14-day upgrade signal.

**5. Returned:** cleaned file (types coerced, deduped, regions flagged), this 1-page summary, and 2 chart suggestions — a channel-by-churn bar and a month-11 annual-churn spike line.

PRICE HISTORY0 cr · current

PROMPT TEMPLATE · v1.0.0

Attached: <FILE>.
1. Describe the schema and data quality issues.
2. Compute <METRICS>.
3. Find 5 non-obvious patterns or outliers.
4. Suggest 3 follow-up analyses.
5. Return: cleaned file + 1-page written summary + 2 chart suggestions.