Using level, trend and seasonality combined with smoothing parameters, the Holt-Winters algorithm creates a simple but effective model for forecasting with seasonality. It works particularly well when a clear seasonal element is present in the timeseries.
State Space Exponential Smoothing
State space models are a continuation of the Holt-Winters algorithm but allow for a variety of parameter combinations (level, trend and seasonality). In many cases, the results may be identical to the Holt-Winters forecast. However, where there are differences, the results from the state space models tend to be more accurate.
Autoregressive Integrated Moving Average models, or ARIMA for short, behave a little differently from the other two models. They can be thought of as a filter that separates the underlying signal from the noise. This is achieved by combining both an autoregressive and moving average model. ARIMA works by examining the difference in consecutive values in the timeseries, rather than the actual data values themselves.
Our tool provides a simple interface to complex forecasting algorithms for a non-technical audience. We encourage you to explore the topic of forecasting further to build models tailored for your specific use cases.
Privacy and Usage Policy
We do not store any data relating to any Google account that connects to this tool.
In order to carry out the forecasting calculations, your data must pass from Google Analytics, through our server and finally to your browser via an encrypted HTTPS connection.
We only request read-only access to Google Analytics data. Our tool is not able to access any other information connected to your Google account. We are not able to make any changes to Google Analytics data. We are only able to request data, which we then use for the purpose of forecasting.
We do not request, receive or store any refresh tokens from Google that would enable us to access your Google Analytics account at any time. As such, once your account is authenticated, we can only request data within a short timeframe of approximately one hour. Once the authentication token has expired, re-authentication is required to continue using the tool.
We do use Google Analytics to track interaction with this tool. We do this to better understand how we can improve the tool for the end user. No personally identifiable information is recorded.
Email addresses provided during early-access signup (not emails provided for Google authentication) may be used for marketing purposes. You may unsubscribe from these emails at any time.
THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE.
Thank you for taking the time to check out Forecaster, RocketMill's new tool to simplify forecasting.
You can find out more about Forecaster and how it works here. We hope you enjoy using the tool and look forward to hearing your feedback.