Historical and forecast sales
Adjusted sales: net sales plus partial refunds, excluding tax, tips, and fees.
1. Upload Square files
Upload your Square Transactions and Items exports. CSV, XLSX, and XLS files are supported.
Forecast assumptions
Executive view
Sales dashboard
Forecast interpretation
Monthly actuals + forecast
Weekday pattern
Top categories
Scenario forecast
Monthly actuals
Best and worst periods
Monthly forecast
Weekly forecast
Next 30 sales days
Top categories
Top items
Data quality review
Outlier days
Refunds
Advanced forecast method
The expected forecast is an ensemble forecast. The dashboard backtests several methods on the most recent observed sales days, weights each method by accuracy, and then projects future sales using known operating rules, holiday/shopping events, and City Creek / Temple Square tourism assumptions.
- Adjusted sales = Net Sales + Partial Refunds.
- Usable revenue estimate = Adjusted Sales + Fees, excluding sales tax and tips.
- Closed days = Sundays, Thanksgiving Day, and Christmas Day by default.
- Outliers are winsorized for model fitting so one-time spikes do not dominate the forecast.
- Model selection uses backtested WAPE, MAE, RMSE, and bias.
- Conservative and optimistic ranges use the recent backtest error distribution, widened by the scenario settings when needed.
Formula concept
Expected sales = weighted average of:
robust weekday/hour model
+ exponentially weighted demand level
+ ridge trend regression
+ recent moving average
+ simple open-day baseline
Each model is adjusted for open hours, known closures,
tourism seasonality, and holiday/shopping-event factors.Because the store opened January 2, 2026, the dashboard still cannot learn true prior-year holiday seasonality until older comparable history exists. The added sophistication improves structure and error handling, but it does not eliminate the uncertainty around the first Black Friday and Christmas season.