Five prediction methods are available to help you build forecasts based on the characteristics of your data and your forecasting goals. The right method depends on the patterns in your historical data, such as trends, growth, short-term fluctuations, or seasonality.
Available Prediction Methods
Trend Tracker
Best for: Data with a consistent trend.
Trend Tracker is designed for data that follows a relatively steady upward or downward direction over time. Use it when your historical data shows a clear and consistent trend without significant changes in the rate of growth.
Example: Use Trend Tracker to forecast sales that have been increasing at a relatively consistent rate over several periods.
Growth Accelerator
Best for: Data with an increasing growth rate.
Growth Accelerator is suitable when your data is not only growing but the rate of growth is also increasing over time. It can help capture accelerating growth patterns in your historical data.
Example: Use Growth Accelerator when your monthly sales are growing more quickly each quarter.
Smooth Forecaster
Best for: Data with short-term fluctuations.
Smooth Forecaster helps reduce the impact of short-term fluctuations in your data to make the underlying trend easier to identify.
Use it when your data has temporary ups and downs that may not represent the longer-term direction of your business.
Example: Use Smooth Forecaster when monthly sales fluctuate significantly but follow a relatively stable overall trend.
Season Sense
Best for: Data with recurring seasonal patterns.
Season Sense is designed for data that follows predictable patterns that repeat over time.
Use it when your business experiences regular changes during particular months, quarters, or other periods.
Example: An e-commerce business may use Season Sense when sales consistently increase during the holiday season and decrease afterward.
Smart Predictor
Best for: Complex time series.
Smart Predictor is an advanced method designed to handle more complex patterns in time-series data.
Use it when your data contains multiple patterns or when it is difficult to identify a single clear trend, growth pattern, or seasonal pattern.
Example: Use Smart Predictor when your historical data combines changing trends, fluctuations, and other complex patterns.
How to Choose a Method
Use the characteristics of your historical data as a starting point:
If your data... | Consider |
|---|---|
Follows a consistent upward or downward trend | Trend Tracker |
Has an increasing growth rate | Growth Accelerator |
Has frequent short-term fluctuations | Smooth Forecaster |
Shows recurring seasonal patterns | Season Sense |
Contains complex or multiple time-series patterns | Smart Predictor |
These guidelines are a starting point rather than strict rules. The most appropriate method can vary depending on your dataset and forecasting goal.
Tips and Best Practices
- Review your historical data first. Look for clear trends, growth patterns, fluctuations, or seasonality.
- Match the method to your data. Choose the method that best reflects the pattern you want to analyze.
- Don't assume the most advanced method is always the best. A simpler method may be more appropriate when your data has a clear, consistent pattern.
- Review your method when your data changes. A method that works well for one dataset may not be the best fit as your business patterns change.
Tip: Start with the method that best matches the most visible pattern in your data.