Back to Home

How to Use ForecastFlow

A guide to understanding your project forecasts.

1. Preparing Your Throughput Data
The quality of your forecast depends entirely on the quality of your historical data.

Throughput is a measure of how many work items your team completes in a given period. For this tool, you need to provide your team's throughput on a **daily** basis.

Your data must be in a single-column CSV file. The first row can optionally be a header, but every subsequent row must contain a number representing the quantity of items (stories, tasks, etc.) completed on a given day.

Example CSV Data:

items_completed
2
0
3
1
1
0
4
...

Each number represents the total items completed in one day. Include days with zero throughput (like weekends or holidays) as they are a natural part of your process. A good dataset has at least 20-30 data points.

2. Choosing a Forecast Type
ForecastFlow offers two ways to look into the future.

How many items to forecast?

Use this when you have a fixed scope (e.g., a project with 50 stories) and you want to know **when** it might be done. The simulation will tell you how many days it is likely to take.

How much can we get done by?

Use this when you have a fixed deadline and you want to know **how much** work you can likely complete. You provide a target date, and the simulation tells you how many items you are likely to finish.

3. Understanding the Outputs
The results are based on running 10,000 simulations using your historical data.

Completion Date Forecast

This calendar view appears when you run a "How many items" forecast. Each day on the calendar is color-coded to show the cumulative probability of finishing the work *by that day*. Greener days mean a higher certainty of completion. This helps you visualize the likely range of delivery dates.

Forecast Certainty

This is the core of the forecast. It breaks down the probabilities into specific numbers. For example:

  • An 85% chance means that in 8,500 out of 10,000 simulations, the outcome was this number or better. This is often a good level of confidence for planning.
  • A 50% chance represents the average or most likely outcome, but it's as likely to take more time as it is to take less.

Simulation Results (Histogram)

This bar chart shows the distribution of all 10,000 simulation results. The x-axis shows the outcome (days or items), and the y-axis shows how many simulations ended with that outcome. A tall bar means that outcome occurred frequently in the simulations. This chart helps you see the spread of all possible outcomes and identify the most common ones.