A planner before you spend
Estimate the minimum detectable effect and the duration a test needs from your own baseline, before turning anything off.
Incrementality
Find out what your marketing actually caused
Plan holdout and geo experiments with the right duration, then read lift with a p-value, an interval and a plain three-way verdict.
- Minimum detectable effect and duration planner
- Geo tests by session country
- Unified view with attribution and the mix model
Testing capabilities
Results with a verdict
Lift is reported with a p-value and an interval, and summarised as positive, negative or inconclusive — so a noisy result is not mistaken for a win.
Geo experiments
Use session country to compare regions where a channel ran against regions where it was held back.
No return figure without a real change
If spend did not actually move, or the effect sits inside the margin of error, Deltryx does not display a return figure.
One unified measurement view
Attribution, the mix model and test results side by side for each channel, with an agreement score that shows where the methods disagree.
Know before you buy
What to know
- Tests are run in your ad platforms; Deltryx plans and reads them but does not pause campaigns for you.
- Synthetic-control methods are not available yet.
- Small tests often end inconclusive — the planner is there to tell you that before you start.
Questions, answered plainly
Do I have to stop a channel completely?
No. A holdout can be a share of regions or a reduced budget; the planner shows how the size of the change affects the duration you need.
How does this relate to attribution?
Attribution tells you which touchpoints preceded orders; a test tells you what would have happened without the channel. The unified view puts both next to each other.
See your own numbers, labelled honestly
Start a free trial and connect a store and an ad account, or book a walkthrough with our team.