How Advisionr measures the effect of a change
Every executed action gets a before and after measurement. Advisionr accounts for conversion delay, season, market and other changes so the outcome is fair.
- How long a measurement takes
- Why a campaign rests during a measurement
- How season and market are filtered out
Why this matters
A change is only valuable if you can see what it delivered. Advisionr therefore compares the period before an action with the period after it.
Because many things change at the same time, the platform uses a few fixed rules to keep chance and noise out of the outcome.
Measurement window and conversion delay
Customers do not always buy right after a click. Advisionr looks at how long conversions take to come in for your account and extends the window accordingly, up to 30 days.
Rest per campaign
While a measurement runs on a campaign, another type of action on that campaign waits until it is finished. That way two changes do not spoil each other’s result. Protective actions, such as exclusions or pausing, always go ahead.
Season, market and two comparisons
Each measurement is compared with other campaigns in the same account that were not changed. That way Black Friday does not count as success and a quiet January not as failure. The outcome is only better or worse when both the short and the long comparison agree.
After 7 days there can already be a preliminary result. A conclusion is only drawn that early when the effect is very clear.
Step by step
Follow this order to avoid drawing conclusions from incomplete data or mixing several changes at the same time.
- Apply an action via One-Click or AutoPilot.
- Check the measurement status in the action history.
- Afterwards, read the outcome and the explanation under Results.
How to get the best result
Advisionr performs best when data, goals and responsibilities are clear. Use this working method to improve the quality of analyses and decisions.
- Avoid big manual changes on the same campaign during a measurement.
- Let measurements finish; the last days contain late conversions.
- Use the explanation with an outcome to see why something does or does not count.
Common mistakes
- Drawing conclusions after three days.
- Thinking every action has an immediately measurable effect.
Frequently asked questions
Why does a measurement sometimes not count for learning?
If there was another change at the same time, a tracking issue or too little data, Advisionr deliberately does not learn from it.
Can I roll back an action during the measurement?
Yes, rolling back is always possible.