FORECAST.ETS is the most under-used forecasting function in Excel. The pattern is: give it 24+ months of history, tell it the seasonal cycle, and it returns a point forecast and a confidence interval. The senior uses FORECAST.ETS on every monthly sales, demand, and revenue forecast. The function is reliable, fast, and built into Excel 365 - no add-in, no library.
The two functions
FORECAST.ETS handles seasonality. The pattern is: =FORECAST.ETS(target_date, values, timeline, [seasonality], [data_completion], [aggregation]). FORECAST.LINEAR fits a straight line. The pattern is: =FORECAST.LINEAR(target_x, known_ys, known_xs). The senior uses FORECAST.ETS for monthly and quarterly data with seasonality; FORECAST.LINEAR for annual data or any series without a seasonal cycle.
The seasonality parameter
The seasonality parameter is the length of the cycle. 12 for monthly data (cycle repeats every 12 months), 4 for quarterly data, 7 for weekly data, 1 for annual data. The senior always sets the seasonality explicitly - never leaves it blank. Excel will infer the seasonality from the data, but the inference is unreliable on short series. Explicit is safer.
How to validate the forecast
The senior validates every forecast against a hold-out sample. The pattern: hold out the last 3 months of history, run FORECAST.ETS on the remaining months, compare the forecast to the actual. If the forecast is within 5-10% of the actual, the model is reliable. If not, the senior investigates - structural break, new product launch, market change - and adjusts. The validation step is the difference between a forecast that informs decisions and a forecast that wastes the reader's time.
What the senior-built forecast looks like
The senior-built forecast has three elements: a chart of the historical data, a shaded region for the forecast horizon, and a band for the 95% confidence interval. The chart tells the story - what happened, what is expected, and how confident we are. The forecast is a worksheet, not a slide. The reader opens the file, sees the chart, and trusts the numbers. The pattern ships in one day and replaces the manual spreadsheet-tidying that consumes every FP&A team.
