Data & Analytics · Analyze & Recommend
Model the decision before you make it.
Ask "what happened last time we raised prices?" Doe pulls historical pricing and churn data from your database, segments the impact by customer type, and builds a forward model you can inspect and adjust. You make the decision with actual data instead of guesses in a spreadsheet.
Works acrossPlanetScale
What you get.
Doe helps you model business decisions by querying your database for historical patterns. Ask what happened last time you raised prices, expanded into a segment, or deprecated a feature. Doe finds the historical data, segments the impact by customer type, and builds a forward model you can inspect and adjust. Connects to Neon, Supabase, and PlanetScale. All queries and model code visible.
The pricing meeting is Thursday. The model is a spreadsheet held together with guesses.
The CEO wants to raise prices 15%. Before the decision, she needs to know: how will this affect churn? Net revenue? Expansion? Finance builds a spreadsheet with three scenarios, each based on assumptions typed into yellow-highlighted cells. Churn estimates come from a conversation with the CS lead, not from actual data. The model takes three days and nobody trusts it.
Your database has every pricing change you've made, every churn event, every expansion with exact timing. The data to answer "what actually happened last time" is there. The problem is turning that into a forward-looking analysis within the decision timeline.
What changes.
- 01Model inputsBefore · Assumptions typed into yellow cells from stakeholder conversationsWith Doe · Historical patterns from your actual customer data
- 02SegmentationBefore · One model for the whole customer baseWith Doe · Impact segmented by customer type, because different segments respond differently
- 03AuditabilityBefore · A spreadsheet where nobody can trace how the numbers were derivedWith Doe · Model code, source queries, and historical data all visible
- 04Time to answerBefore · Days to weeks building and iterating on the spreadsheetWith Doe · Hours. Ask the question, review the model, iterate on assumptions
How Doe models business scenarios
- 01Parses the question and identifies relevant historical dataDoeDoe maps the question to past pricing changes, churn events, segment attributes, and contract values in your database.
- 02Pulls historical pricing and churn patternsPlanetScaleWhich customers saw increases, by how much, and what happened next — renewed, churned, downgraded, or expanded. Segmented by customer type.
- 03Builds a forward model from historical ratesDoeDoe applies historical churn rates by segment to your current customer base at the proposed 15% increase. If multiple past changes exist, it shows the range of outcomes. Model code fully visible and editable.
- 04Delivers the projected impact to your channelSlackDoe posts the segment-by-segment projection with the historical basis shown. Example: "Mid-market churn increased X points after the last 10% raise — applying that rate at 15% gives this range." Source queries and model code attached.
Up and running in under ten minutes.
- 01Connect your toolsOne-click OAuth for each integration. No API keys, no engineering.
- 02Describe what you need“We are considering raising prices 15% on the Pro plan. Model the impact on MRR, churn, and upgrade rate using our last 12 months of billing data. Show best-case, base-case, and worst-case.”
- 03It runs on scheduleOn demand. Ask a new question whenever a decision needs modeling and the scenario lands in minutes.
Before you delegate.
- 01What kinds of scenarios can Doe model?Any question where your database contains relevant history. Pricing changes (what happened last time?), feature deprecation (which customers use this and what's their churn risk?), market expansion (how do customers in adjacent segments behave?). The quality of the model depends on how much relevant history you have.
- 02Can I see the model code?Yes. The model runs as Python code in a sandbox. Every line is visible: how historical rates were computed, what assumptions were applied, how the projection was built. You can edit the code, override assumptions, and rerun.
- 03What if we've never made this kind of change before?Doe is honest about it. If there's no relevant history in your data, the model says so rather than fabricating projections. It can still show you the current state (who would be affected, by how much) even without historical analogues.
- 04How reliable are the projections?They're only as good as the historical data they're based on. If you've changed pricing three times with clear churn data after each, the model has a reasonable basis. If you've done it once, it's one data point and the model will say that. Doe shows you the data, not magic.
- 05Can I adjust assumptions and rerun?"Rerun but assume mid-market churn is higher than historical." Doe updates the model and delivers revised projections in minutes. You can iterate as many times as you need.