What will next quarter's sales be?
Forecast from your own file
Spreadsheet trend lines miss seasonality and trend breaks. Drop in your monthly or weekly sales and see a forecast with its likely range, and how it did on the most recent months, so you can judge it before you plan on it.
To see where your own measure is heading, with a likely range, try the free forecasting tool.
Three questions this answers
- What will sales be next quarter, and how wide is the range?
- Is there a seasonal pattern, and how big is it?
- How wrong has the forecast been on recent months?
What the forecast tells you
Most businesses forecast by extending a trend line. That works until it doesn't.
Seasonality
When your history has seasonal cycles, the forecast carries the pattern forward, so a strong December stays strong in the forecast. A straight trend line does not.
A likely range
Every forecast comes with 80% and 95% prediction intervals. Plan hiring, inventory and cash for the middle, and look at what the edges would mean.
Checked on recent months
The report shows how the method did on the most recent months of your own history, so you can see how far off it has been before you rely on it.
Upload. Forecast. Plan.
Three steps from historical data to actionable projections
Export Your Revenue Data
Export your sales history from your accounting software, store or CRM. Include a date column and a column of sales. The more history, the more the seasonal method has to work with.
Upload Your CSV
Drop the file in, or try the example file first. The tool picks the method your history supports: seasonal smoothing, trend smoothing or a simple ARIMA model.
Get Your Forecast
Receive the forecast with its likely range and how it did on recent months, with the method named and the R code.
The methods it runs
One is picked for your history, and the report says which
Holt-Winters seasonal smoothing
Used when your history has seasonal cycles. Tracks the level, the trend and an additive seasonal pattern, and projects all three forward.
Holt trend smoothing
Used when there is no seasonal pattern to follow. Tracks the level and the trend and projects them forward.
ARIMA(1,1,1)
The fallback when the smoothing methods do not apply. A simple autoregressive model of the changes in your series.
Sales Forecasting FAQ
How much history do I need?
The seasonal method needs enough history to see the seasonal pattern repeat. With less, it fits a trend-only method. If a model cannot be fitted, it uses ARIMA(1,1,1), and as a last resort a straight trend line. More history gives a narrower, more reliable range, and the report shows how the forecast did on recent months so you can judge it.
Which forecasting method does it use?
When your history has seasonal cycles it uses Holt-Winters additive seasonal smoothing. Without seasonal cycles it uses Holt's level-plus-trend smoothing, and if that does not apply it uses an ARIMA(1,1,1) model. Every report names the method that ran and shows the R code.
What is the likely range around a forecast?
A forecast is a best guess, and the range shows how far actual sales could plausibly land from it: the report gives an 80% and a 95% prediction interval. A wider range means more uncertainty. Plan for the middle, and check what happens to your plan at the edges.
What should my file look like?
One row per week or month, with a date column and a column of sales. The example file shows the layout. Forecast one series at a time, such as total sales or one product.
How is this different from forecasting in a spreadsheet?
Spreadsheets also offer smoothing-based forecasts. What this adds is the method named in every report, the R code that ran, and a check of the forecast against recent months, so you can see how wrong it has been before you plan on it.
Forecasting & Time Series Resources
Guides and examples on forecasting
Tutorials
Example Reports
Ready to Forecast with Confidence?
Try the example file first, then drop in your own sales history.
Where to start
One email with your data and your question. We take it from there.
MCP Analytics
Data science by email
Send us your data and your question, and we will reply.
Try it on your own data
30 days free.
Answers, analytics and custom analytics. No card, no call, nothing to install.
30 days free. →