Expand my Community achievements bar.

Join us at Adobe Summit 2024 for the Coffee Break Q&A Live series, a unique opportunity to network with and learn from expert users, the Adobe product team, and Adobe partners in a small group, 30 minute AMA conversations.
SOLVED

Adobe Analytics Alerts

Avatar

Level 2

Hi,

 

I set up Adobe Analytics alerts with the following configuration:

Visits + "anomanly below expected" + "99.75% threshold"

 

Now I am receiving ten emails per day saying that my visits were 30% below expected.

In my opinion, these are false positives.

However, I only want to be notified when the visits are 99.75% below expected.

 

Can you help or clarify how anomaly and threshold play together?

Thank you!

1 Accepted Solution

Avatar

Correct answer by
Community Advisor

You should read this blog on how to stop false alarms during off peak hours: https://experienceleaguecommunities.adobe.com/t5/adobe-analytics-blogs/woken-up-at-3am-by-false-hour...

 

I can't help you more with the anomaly detection... to be honest, I don't use that feature since I don't really trust how it calculates anomalies.... I create my own alert thresholds... 

View solution in original post

8 Replies

Avatar

Community Advisor

Dear Adam19,

Anomalies are based on Standard Deviations.

You need to understand how Adobe calculates anomaly so that you can understand the threshold. If you have an alert for hourly granularity, the 24-hour cycle, as well as the weekend/weekday cycle is validated. So, if the numbers change based on the threshold, you should receive an alert.

If you think you have more false alerts, try some other conditions or check whether you want to stop the alerts during the non-peak hours (Like 12:00 AM to 4:00 AM). There are some workarounds for that, check google.

Thank You, Pratheep Arun Raj B (Arun) | NextRow DigitalTerryn Winter Analytics

Avatar

Level 2

Thank you, @PratheepArunRaj !

If you want to use anomalies it seems to be the best the turn off the alerts during non-peak hours.

Avatar

Correct answer by
Community Advisor

You should read this blog on how to stop false alarms during off peak hours: https://experienceleaguecommunities.adobe.com/t5/adobe-analytics-blogs/woken-up-at-3am-by-false-hour...

 

I can't help you more with the anomaly detection... to be honest, I don't use that feature since I don't really trust how it calculates anomalies.... I create my own alert thresholds... 

Avatar

Level 2

Thank you, @Jennifer_Dungan !

I have to say that I also do not trust this a lot.

The article is very helpful and interesting!!

Avatar

Level 1

 

Hello @Jennifer_Dungan 


I have a follow-up question. I am trying to set alerts for our business KPIs and do not understand the calculations either.
Can you offer any guidance on how to create effective alerts? I've attempted one method, which involves examining the average daily visits from the past six months and setting an alert if the visits decrease by 10%. Is this method reasonable, or would you suggest a better approach?

I appreciate any help you can provide.

Avatar

Community Advisor

Hi @samk10 that does seem reasonable on the surface... I would just want to be sure that the calculated "average daily visits over past six month" works properly... using the average (or "mean") function in Calculated Metrics relies on seeing the Day dimension in the rows on the freeform table (it averages the rows)... unless maybe you are using a custom date range (like last 180 days) and a static divisor of "180" to get your average that way?

 

Then using a threshold on that should be a decent way to trigger alerts.

 

Sounds like a nice creative solution

Avatar

Level 1

 

Yes, I calculated the average manually and then used it to set the alerts. I'll try the advanced function. Thank you very much. I would really appreciate any other recommendations or tips for setting up alerts. 

 

 

Avatar

Community Advisor

I would be happy to consult on ideas for alerts. Suggesting new ones can be tricky without being familiar with your implementation, or what is important to your stakeholders