What is the revenue forecast?

The revenue forecast extends your monthly net-revenue history three months into the future using a simple, inspectable trend projection. It is deliberately a trend line, not a demand model: no seasonality decomposition, no external signals — the same 12 monthly bars you can see, extended by the method you chose.

Formula

inputs   = last 12 complete calendar months of net revenue
forecast = trend(inputs), projected 3 months, floored at zero

Worked example

Twelve months of net revenue climb irregularly from €82,000 to €110,000. The linear method fits one straight trend through all twelve and lands the next three months around €112–118k. Switching the widget to a 3-month moving average makes the projection lean on the recent quarter instead — better after a step change (a big launch, a price rise), worse when one outlier month sits in that window. The gap between the two projections is itself information about how stable the trend is.

How Saldo Metrics computes it

The widget sums line_revenue_base from v_order_lines per calendar month over the last 12 complete months — the current partial month is excluded, since its low total would read as a crash — counting net sales only. Four methods are available per widget: linear (least-squares trend over all 12 months, the default), 3- and 6-month moving average (trend of the smoothed series, weighting recency), and exponential smoothing. All project three months and clamp at zero — a declining trend line may cross zero mathematically, but a negative revenue forecast would be noise. The chart shows actuals and the projection as one continuous series.

Why it matters

Inventory purchasing, cash planning and hiring all need a revenue number for next quarter, and the honest baseline is your own trend. A transparent projection you can recompute by eye is more useful for those decisions than a black-box model — when reality diverges from it, the divergence is meaningful rather than a modeling artifact.

Common mistakes

• Trusting it across seasonality. A linear fit through October–December happily projects the Christmas ramp into February. None of the methods know what December is; correct by judgment.
• Reading the projection as a target. It is inertia, not ambition — the number the business lands on if nothing changes.
• Switching methods until the forecast looks good. Pick the method that matches your trend's shape (stable → linear, recent step change → short moving average) and keep it, or the forecast becomes an opinion.

Where you see this in the app

The Revenue Forecast dashboard widget, with the projection method selectable per widget.