# Forecast the next few months from historical data in Excel

Builds a simple, explainable forecast from your past numbers, with trend and seasonality, a range, and the Excel steps to reproduce it.

- Topic: Data & Analysis
- Framework: Few-Shot
- Level: Advanced
- Tags: Data & Analytics, Plan, Explain & Teach, Fill-in Template
- Source: https://www.thinkwithprompt.com/prompts/forecast-the-next-few-months-from-historical-data-in-excel

## Prompt

```text
Help me forecast the next few periods from my historical data, using a method I can build and explain in Excel.

**What I'm forecasting:** [e.g. "monthly revenue", "weekly ticket volume", "units sold by product"]
**History:** [PASTE date and value columns - ideally 24+ months for monthly data]
**Forecast horizon:** [e.g. "next 6 months"]
**Known events:** [one-offs in the past (promotions, outages, COVID, lost client) and planned changes ahead (price rise, new product)]
**What the forecast is used for:** [e.g. "staffing plan", "budget", "inventory order"]

Steps:
1. Inspect the history: trend, seasonality, unusual spikes or dips, missing periods, and structural breaks. Say which past points should be adjusted or excluded, and why.
2. Choose the simplest method that fits, and say why: average of recent periods, linear trend (FORECAST.LINEAR / TREND), seasonal method (FORECAST.ETS or Forecast Sheet), or seasonal indexes applied to a trend.
3. Back-test: forecast the last 3-6 known periods as if they were unknown and report the error (MAPE or average % miss).
4. Produce a range, not just one number: low / expected / high, with the confidence interval if using FORECAST.ETS.
5. Layer in known future events separately, so the baseline and adjustments are visible.

Return:
- Observations about the history.
- The forecast table: Period | Baseline | Adjustments | Forecast | Low | High.
- Back-test accuracy.
- The exact Excel formulas or Forecast Sheet settings to reproduce it.
- Assumptions and the main risks to the forecast.

Use only my data. If the history is too short or noisy for a reliable forecast, say so plainly and suggest what to use instead.
```

## What to fill in

- e.g. "next 6 months"
- e.g. "staffing plan", "budget", "inventory order"

Free to copy and use. From ThinkWithPrompt (https://www.thinkwithprompt.com/), a library of 2,663 curated AI prompts.