Skip to main content

Sales Forecasting

Sales forecasting gives you the revenue projection for your deals in the pipeline: it estimates the revenue you will generate from won deals in the next months. It is based on historical deals data, and uses machine learning algorithms to provide you with accurate predictions.

Request a Sales Forecasting prediction​

To request a sales forecast you need to have a collection with historical data for deals in the desired pipeline, enabled on Sales Forecasting. You can create a collection and upload your data following instructions in the Collection section.

Once you have successfully uploaded your deals data, you can request a sales forecasting by sending a POST request to the invocation endpoint, with sales-forecasting as invocationType. See more at Invoke Sales Forecasting.

You can select the desired horizon for your forecast, by specifying in the request body:

  • horizon: the number of months for which you want to predict the revenue. Should be an integer between 1 and 12.

You can also specify the month from which you want to start the forecast, by adding the optional parameter:

  • first_month_to_predict: a month in YYYY-MM format indicating when the forecast should begin. If not specified, forecasting will automatically start from the month immediately following the last month in your data.

Interpret the Sales Forecasting prediction results​

The results of a sales forecasting request will provide you with the estimated revenue generated from won deals in your pipeline, for the horizon you selected.

The result will have n predictions, where n is the number of months in the horizon you selected. Each prediction will include the month (represented by the first day in the month, in the format YYYY-MM-DD) and the corresponding estimated revenue.

Example​

You requested a sales forecasting prediction with horizon 5, and no specified first_month_to_predict. The output, if the last complete month in data is e.g. 2021-12, will have 5 predictions starting from 2022-01-01. In json format, it will be:

[
  {
    "predictedDate": "2022-01-01",
    "predictedValue": 10000
  },
  {
    "predictedDate": "2022-02-01",
    "predictedValue": 12000
  },
  {
    "predictedDate": "2022-03-01",
    "predictedValue": 11000
  },
  {
    "predictedDate": "2022-04-01",
    "predictedValue": 13000
  },
  {
    "predictedDate": "2022-05-01",
    "predictedValue": 14000
  }
]