More relevant Alerts (3-hourly/6-hourly Alerts, Alerts only if HIGHER or LOWER plus Significance) | Community
Skip to main content
luekl
Level 3
Jan 12, 2018 8:47 AM UTC
New

More relevant Alerts (3-hourly/6-hourly Alerts, Alerts only if HIGHER or LOWER plus Significance)

We are heavy Alert Users.

The challenge: We want *relevant alerts*. Adobe's Anomaly detection system is ok, but 95% of the *Hourly Alerts* it spits out are completely irrelevant as traffic simply oscillates too much from one hour to the next, AND because Adobe seems to take only the last 24 hours into account when looking for Anomalies. So we had to deactivate all our Hourly Alerts.

1. However, Daily Alerts are too late for something really critical, so Adobe needs to either improve the Hourly Alert logic, or offer something in between, like _3-hour-periods or 6-hour-periods_ which don't jump up and down all the time like hourly data.

2. Alerts should be configurable to fire only "only if a value is _HIGHER/LOWER_ than expected" (and a certain Significance Level. We don't care so much if the Conversion Rate of Google Shopping is greater than expected, we want to be ALERTED when it goes DOWN because then we have to take immediate action.

5 replies

Community Manager
Oct 5, 2022 7:04 PM UTC
No text available
Translate
jantzen_b
Employee
Oct 27, 2020 10:53 PM UTC
No text available
Translate
New Participant
Mar 25, 2019 9:37 AM UTC

This is something I would be interested in.  I am relatively new to alerts, but want to be able to see issues as soon as possible.  zero bookings in a 3 hour period would be great.

Hourly alerts mean I can get up to 10 a day for some of our sites and it's difficult to know if there are issues, especially out of hours (which is when most issues seem to happen).

Translate
Asheesh_Pandey
New Participant
Mar 16, 2019 11:58 AM UTC

Have more ideas to improve it further here Enhanced Analytics Intelligent Alerts in 2019

Translate
danf46525780
New Participant
Feb 28, 2018 10:11 PM UTC

We also love alerts, but face the same issues as luekl where trends are different by hour and day of week. And also when campaign activity adjusts our 'norm', it sends people into a frenzy when revenue is down.

Translate