Plan requirement
| Subscription | UC-Premium (Collaboration 7), x-bees SuperBee |
Have a KPI tell you when something changes, instead of remembering to open a report. Three kinds of alert, set up from CDR-View.
Where alerts live
Go to Analytics and open CDR-View 2.1. Select the KPI you want to watch, hover over it, and choose the Monitor icon. The pop-up then asks which of the three kinds of alert you want.
If you do not see Analytics at all, the account is missing the right to see it. Ask us to enable analytics access for that user.
Regular KPI updates
A scheduled alert: the figure arrives hourly, daily, weekly or monthly whether or not anything has changed.
- Select Regular KPI updates, then Next.
- Give the alert a name. This is optional, but an inbox full of unnamed alerts is hard to sort out later.
- Under Send every, choose hourly, daily, weekly or monthly.
- Add a message for the recipients if the number needs context.
- You are already a recipient. Add colleagues by typing their names.
- Select Create alert.
When a KPI crosses a limit
This one only fires when something is actually wrong, which is usually the more useful kind.
- Select KPI crosses a set limit, then Next.
- Choose the condition: greater than, greater than or equal to, less than, less than or equal to, equal to, or not equal to.
- Enter the threshold value.
- The interval at which the condition is checked is shown under the condition. Select Change if it does not suit.
- Add a message and any extra recipients, then select Create alert.
When one attribute crosses a limit
The same idea, narrowed to particular values of an attribute such as a region, a team or a product group. Useful when the total looks healthy but one part of it is not.
- Select Values of an attribute crosses a set limit, then Next.
- Choose the attribute, and tick the values you care about.
- Set the condition and the threshold value as above.
- Add a message and any extra recipients, then select Create alert.
Set thresholds people believe
An alert that fires every afternoon becomes background noise within a week, and then a real problem passes unnoticed. Set the threshold at the point where somebody would genuinely change what they are doing. If nobody would act on it, make it a scheduled update instead.
Review your alerts after a month. Anything that fired repeatedly without anyone acting is set too tight. Anything that went wrong without an alert needs one.
Comments
0 comments
Article is closed for comments.