Notification Agents: A Hidden Treasure (3 of 4)

noviembre 16, 2018

This is the third in a series of four blog posts on TARGIT® Notification Agents. For more on basic Notification Agents, we recommend reading blog post No. 1 and blog post No. 2.

In this blog post, we’ll examine how to utilize TARGIT® Notifications to:
  • Look for missing registrations and let users know they need to register (Use Case No. 6)
  • Check for N/A (Unknown) Values and monitor if such values occur and/or check if we succeed in bringing down these values (Use Case No. 7)
  • Check for extreme values to catch errors in registrations and suspected fraud (Use Case No. 8)
  • How to monitor a value that you have calculated inside TARGIT® (Use Case No. 9)
  • How to monitor user engagement in TARGIT® (Use Case No. 10)

Use Case No. 6 – Missing Registrations

In this case, we have a list of employees who are supposed to log their project hours on a monthly basis. However, they tend to not register their hours on time, so we are going to use TARGIT to improve the data discipline and, through that, the quality of the data available for analysis.

(Tip: Normal behavior in a TARGIT Data Object is to only show members with values. In this case, we right-clicked the table, selected data from the right-click menu and chose include all members).

From this table, we create a Notification Agent by right-clicking the Total amount – just like in many of the previous Use Cases in this series. Note: This use case covers only select dialogue tabs. For a more detailed look, please see blog post No. 1

Here, we see the Condition tab, which we change to condition is less than 1:

We also make sure to monitor each member of Employee. Finally, we set the Schedule to once a month and decide who should receive the email. Now we have an effective tool to work with data discipline in the organization.

Use Case No. 7 – Word with Missing Values (N/A or Unknowns)

Almost any organization has data quality issues in its data warehouse. These issues could result from:
  • People forgetting to make complete registrations
  • Data being deleted by accident
  • Data conversion gone wrong
  • New numbering systems being introduced

We could go on, but the point is data quality issues are everywhere, so it makes sense to use the tool at hand (TARGIT Notification Agents) to solve these issues or at least keep them at an acceptable level. 

Below is a crosstab with a filter of Previous Month showing Invoice Quantity per Product. We seem to have a product called N/A (unknown or not applicable). The name N/A was added automatically when we updated our Data Warehouse, which means we have product numbers (at least one) that don’t exist in our product table but still appear on invoices. N/A products appear on 100 Invoices this month, which translates to product(s) we can’t track in our inventory and price(s) we can’t analyze.

Now we start investigating the N/A problem. Our goal is to reduce the number of Invoices with N/A products by 10% every month. So we right-click the N/A number (100) and choose to monitor this value. The condition we create looks like this:
On the Criteria tab, we notice that we inherited the Previous Month filter. And since we right-clicked the value at the N/A product, we also inherit the N/A filter:
So now we just choose Schedule (once a month) and recipient(s) on the Delivery tab, and we're done. Now the process of improving the Data Quality of our Invoices can begin – based on solid data.

Use Case No. 8 – Extreme Values/Fraud Detection

In this use case, we want to be notified if something out of the ordinary occurs. It could be a mistake, but it could also be a sign that someone in the company is doing something suspicious. It could be a credit note with a high amount, a discount of 90%, a registration of an unrealistic number of days/hours – anything that seems strange and needs a little investigation. It could be a simple mistake (which are important to correct), or it could be actual fraud taking place.

In this crosstab, we look at Purchase Credit Memos. Using the filter of Previous Month, we see that we only have two, although one of them is a quite substantial amount ($56,688):

In the future, we would like to be notified whenever Purchase Credit Memo Amount exceeds $10,000. We right-click the total and choose to "Monitor this value". We make sure we are Monitoring each member of Purchase Credit Memo. We set the condition to Is greater than 10,000:
Now we set a frequency of once a month and make sure the right people are notified – and that’s it!

Use Case No. 9 – Calculated Value from Crosstab

The Crosstab below shows a Comparison of two periods and the Growth in Revenue per Country from one period to the other. Some show a negative growth, which we would like to monitor if this trend continues. We want to receive a notification on any negative growth encountered, and it should be checked every two weeks:

Object Notification Agent

However, when we right-click the growth total number we get a surprise. Monitor this value is grayed out. That’s because you can’t monitor a calculated value. However, you can use "Monitor this object". So let’s choose that option instead (referred to as an Object Notification Agent):

Now we move to the left-hand side (Properties) and get the options shown below. In our case, we want to monitor Growth Percent Revenue:

That selection brings up a new dialogue (also on the left-hand side), where we decide on condition. We choose "Negative Values":

We click "Apply Agent" to finish the Condition.

Now we’re back in a guide that looks like the “ordinary” Notification Agents (with the exception of the Condition tab, which we have already taken care of). Note: You can’t have an automatic link back to the Analysis from which this Notification was created.
Now you can add Criteria in the Criteria tab, add a name to the Notification on the General tab, set a frequency in the Schedule tab and finally choose recipient(s) in the Delivery tab. You should at least add a suitable name in the General tab to simplify administration of all your agents.

Tip: If you use the Anywhere Client in TARGIT, every Analysis can be reached with a deep link. That link could be (manually) inserted in the email and look something like this.

Use Case No. 10 – Monitoring User Engagement in TARGIT®

In this use case, we take on the role of a TARGIT Administrator. We are using the TARGIT Log through which we can monitor the number of users and types of reports. This simple crosstab shows the number of requests per user. It’s a pretty precise measurement of the level of activity per user in the TARGIT solution. 

With a filter of Previous Month on the crosstab, we create a Notification Agent that monitors whether any of the users decrease their TARGIT usage. 

We are back using "Monitor this value":

We have to monitor each member of the dimension User and set up a condition saying Decrease by more than 0:
Now we receive a detailed email telling us who has decreased their TARGIT usage since last month.

We can now contact each of these users to determine a reason for the decrease and what we can do to increase their use of TARGIT. This is a great way to ensure users are happy and getting real value from TARGIT.

In our next blog post, we’ll dive deeper into the administrative angle of Notification Agents and draw final conclusions on when to utilize Notifications in your TARGIT environment.

Niels Thomsen

Solution Architect
I’m eager to study the trends of the BI market and compare them to the actual implementations that take place worldwide. My position of Solution Architect at TARGIT has me covering all aspects of turning data into insight, from technical setup to end-user training. I create data warehouses and user friendly analyses and reports that enable our custom..
Continue Reading...

We use cookies to improve your site experience, but they also provide us with information on your use of our website.
To find out more about the cookies we use and how to delete them, see our Privacy Policy. By continuing to browse the site, you are consenting to our use of cookies.