Forecasting
Fload projects your app's future performance using historical trends, seasonality, and automatic model selection. See where your metrics are headed — not just where they've been.
How it works
The forecasting engine analyzes your historical data to detect:
- Trend direction — growing, declining, or stable
- Seasonality — weekly or annual patterns
- Volatility — how stable the metric is
- Mean reversion — whether the metric tends to return to a baseline
Based on these characteristics, it automatically selects the best statistical model for your specific data and explains why that model was chosen.
What you can forecast
| Metric | Description |
|---|---|
| Net Revenue | Proceeds after store cut |
| Gross Revenue | Total customer spend |
| Total Downloads | First-time installs |
| Active Subscriptions | Current subscriber count |
| New Trials | Trial starts |
Ad spend metrics are intentionally excluded — campaign budgets are human decisions, not organic trends that can be modeled.
Forecast output
Each forecast includes:
- Historical data — your actual past values
- Projected values — where the metric is heading
- Confidence interval — upper and lower bounds (95% confidence)
- Trend direction — up, down, or stable
- Trend strength — 0–100 score indicating momentum
- Model explanation — which algorithm was selected and why
Default: 52 weeks (1 year) forward. The forecast window is configurable.
Reading the results
The confidence interval shows the range Fload expects your metric to land in. A wide band means high uncertainty (volatile data); a narrow band means high confidence.
The trend strength score:
- 0–20: Effectively flat
- 20–50: Mild trend
- 50–80: Strong trend
- 80–100: Steep sustained trend
How to use it
- Go to Forecasting in the sidebar
- Select an app
- Choose a metric to forecast
- Adjust the forecast window if needed
- Review the chart and model explanation
You can also ask Chat: "What does my revenue forecast look like for Q4?"
Requirements
Forecasting requires at least 90 days of historical data. Newer apps or recently connected data sources may not have forecasts available yet.