how come for percentage of alert rule

[pigpig]

I created the rule for checking some number which is anomaly. the alerte show this message 

Daily Visits was 4,015 which was 3,494% above expected. This is statistically significant. The filters applied were : xxx

i would like to know how to calulate the percentage from?

attached for raw as well.

3/11/2023	0
3/12/2023	0
3/13/2023	0
3/14/2023	0
3/15/2023	0
3/16/2023	0
3/17/2023	0
3/18/2023	0
3/19/2023	0
3/20/2023	2
3/21/2023	81
3/22/2023	83
3/23/2023	65
3/24/2023	113
3/25/2023	90
3/26/2023	68
3/27/2023	90
3/28/2023	108
3/29/2023	95
3/30/2023	111
3/31/2023	177
4/1/2023	149
4/2/2023	133
4/3/2023	119
4/4/2023	114
4/5/2023	159
4/6/2023	140
4/7/2023	111
4/8/2023	120
4/9/2023	105
4/10/2023	4015
4/11/2023	3767

thanks for support

[Krishna_Musku · Accepted solution]

It depends on the condition and threshold amount you select in the alert builder. Also, if you select the one of the anomaly condition and the % of threshold in the alert builder, then the alert you get is depend on the anomaly and the daily visits for that specific day.

[kayawalton]

Hey @pigpig - if you are asking as to where the 3494% comes from, it looks like it is the percent change for the daily visits that day (4,015) from the average in the last 20 days (which is what you may have set your alert granularity to). 

As @Krishna_Musku said, the anomalies are triggered based on the threshold amount (whether it is by standard deviation, a specific percent increase, or a set value) and the granularity for the training period.

You can find additional resources here: 

https://experienceleague.adobe.com/docs/analytics/analyze/reports-analytics/alerts.html?lang=en

https://experienceleague.adobe.com/docs/analytics/analyze/analysis-workspace/virtual-analyst/intelli...