Did the launch beat the trend?
This isn’t in the product yet.
30-second version
Draw the trend before the launch, extend it, and see if reality jumped off that line. The jump, with a range, is the effect. The feather on the sample chart is only pointing at the launch date.
Do it
- This is a method page, not a button. You cannot run this study in the product today.
What you'll see
The sample chart draws a pre-period gel line, a dashed continuation, and a gel line that leaves it after the launch tick.
Under the hood
An interrupted time series fits the pre-period trend and compares the post-period level to that projection. It assumes the trend would have continued, and that nothing else important happened on the launch date. Short series and seasonal businesses need more care than a default fit.
When not to trust it
- A launch in the same week as a holiday sale is a bad week to call.
- This is not shipped.