Build a simple forecast in a spreadsheet from past monthly numbers

What I am forecasting: {{measure}}
App and version: {{app_and_version}}
How far ahead: {{horizon}}
What the forecast is for: {{purpose}}
Known changes coming (price change, new site, campaign, policy): {{known_changes}}

The history, one row per period with the date and the value:
"""
{{history}}
"""

Task:
Build a simple forecast of {{measure}} for {{horizon}} in {{app_and_version}}, for {{purpose}}. I need a method I can explain and update each month, and an honest range rather than one number.

Format:
1. Read the history first: trend (rising, falling or flat), seasonality (the same months high or low each year) and unusual periods that should not drive the forecast, quoting the values that show each.
2. Method: the simplest that fits, for example a moving average, a straight-line trend (FORECAST.LINEAR or TREND) or a seasonal method (FORECAST.ETS in Excel), with two sentences on why. If the full history is less than two years of monthly data, do not use a seasonal method, and say so.
3. Formulas in code blocks, with the cells they go in, and one plain sentence each.
4. Forecast table: Period, Forecast, Low, High, with a sentence on how the range was set, for example from the size of past errors.
5. Back-test: hide the last three periods, forecast them with the same method and formulas, and compare with the actual values, so I can see how far off the method would have been.
6. Known changes: as a separate, stated adjustment on top of the forecast, never hidden inside the formula. If I gave no size for a change, list it as an unquantified risk to the forecast rather than inventing an amount.

Requirements:
- Under 450 words, plus formulas and the table.
- Use only the history I gave. Do not add outside data or assume growth rates.
- Round forecasts sensibly for the measure, and note that a forecast is an estimate, not a target.

Example of one back-test line's form only: "March: forecast 1,240, actual 1,310, off by 70 (5%)."

If the history has gaps, mixed periods (weeks and months) or a change in how the measure was counted, say so before forecasting and ask whether to leave out the affected periods.