Are you seeing one of these two error messages?

You may have encountered one of the following two error messages while manipulating metric cards or dashboard tiles in Atrium:

• “Negative values removed”
• “Split graph by” settings changed

This article will explain what is happening in each case, as well as provide workaround options.

 

 

“Negative values removed” error message

How does this error message appear?

What is happening here?

What should you do?

“Split graph by” settings changed error message

How does this error message appear?

What is happening here?

What should you do?

 

“Negative values removed” error message

How does this error message appear?

This message reads:

“One or more data points in this graph have negative values. As this graph can not display negative values, all such data points will show up with a value of '0'. You can click the ’View Data’ button to see the raw numbers, including negative values.”

Here is what it looks like:

 

What is happening here?

In any given Atrium graph, Atrium displays 1 to many data points for any number of people or teams.  For example, the graph below shows the total bookings value for two different people, per month, for the last 6 months:

 

But imagine that one or more of the data points is a negative number.  In this case, Atrium does not currently allow you to display negative numbers in a graph, and as such automatically displays all negative numbers as “0” values.

 

 

What should you do?

Unfortunately, there is not currently an ability to visualize negative numbers in Atrium.  However, there are a few mitigation tactics you can take:

1. Switch to the “View details” mode of the graph to see the negative values
   • Click the view details icon at the top of the graph, as you can see here:
     
     
     
   • You can see that the negative value will still show on the back of the card:
2. Expand the data set until your total is not negative

• It is typically quite rare that any data point in Atrium would be negative, and it generally only happens when you have narrowed down the data set to a small number of potential data points (e.g. you are looking at the Bookings amount for only “Contract Adjustment” opportunities in a small period of time like a week).
• In these cases, if you loosen your data filters a bit (e.g. remove the opp type constraint), or increase the grouping time period (e.g. group by month instead of week), you typically will pull in more data points, and the sum of all the data points will now be above zero, thus removing Atrium’s need to alter the data

 

“Split graph by” settings changed error message

How does this error message appear?

This message reads:

“The selected [Data Filter Name] data filter in the ’Split graph by’ control would have resulted in a negative value for one or more bar chart subdivisions. We have therefore automatically changed the setting to ’No Subdivision’”

 

Here is what it looks like:

 

What is happening here?

In Atrium, when viewing data as a bar chart, your data can be split up using the “Split Graph By” function, which you can find on the left side of any metrics card.

 

If you select “No Subdivision”, your data will not be split up, and look like this:

 

But if you decide to split the graph, you can do so by selecting any of the available data filters for the graph.  In the mockup below, the graph is being split by the Opp Source value, which creates multiple subdivisions per bar chart:

 

However, imagine what would happen if the values for one of your subdivisions were negative!  

For example, what if the only Bookings in January with the “Channel Partner” opportunity source were negative value opportunities (this might be possible, for example, if there were no new business channel partner deals in January, but only deals of opportunity type “Contract Adjustment”).

It is not possible to display a bar chart subdividing by a negative amount.  For example, imagine you had 3 opportunities in January as follows:

• Opp Source = Renewal
  • Opp #1 = $5K
• Opp Source = Email Marketing
  • Opp #2 = $5K
• Opp Source = Channel Partner.
  • Opp #3 = -$3K

The total value for your January bookings is $7K, as you booked a total of $5K + $5K - $3K in the time period.

But if we want to subdivide the January bookings by Opp Source, we need to split it into 3 distinct subdivisions:

• Renewal = $5K
• Email Marketing = $5K
• Channel Partner = $-3K

If you really want to see each of those subdivisions, you would need to use a negative Y axis for the $-3K, like this:

 

But the above graph makes it appear like you did a total of $10K in bookings in January.  To make this more clear, imagine you tried to compare your January bookings with your February bookings, where you sold a single deal worth $7K:

 

The above graph makes it look like January had more total bookings than February, but that is not true.  

The implication of this is that we can not both show you the negative subdivision AND make it easy to compare totals between different people or teams.

Given that the system can not effectively subdivide the bar chart with a negative amount, the system throws an error message.

 

What should you do?

1. Remove sub divisions
1. If you change the “Split graph by” functionality to select “No Subdivision”, while you will lose the ability to split up the graph by the currently selected data filter, your graph will now be able to display.
2. Note that this is what the system chooses to do for you automatically
2. Use a different sub division
1. You can select a different data filter by which to split the graph, as long as the new data filter does not result in a negative subdivision
3. Expand the data set and then try to subdivide again
1. It is typically quite rare that any data point in Atrium would be negative, let alone for the sum of all the data points for a given subdivision to be negative.  This generally only happens when you have narrowed down the data set to a small number of potential data points (e.g. you are looking at the Bookings amount for only “Contract Adjustment” opportunities in a small period of time like a week).
2. In these cases, if you loosen your data filters a bit (e.g. remove the opp type constraint), or increase the grouping time period (e.g. group by month instead of week), you typically will pull in more data points, and the sum of all the data points for the given subdivision will now be above zero, thus enabling you to subdivide
3. If you want to take this approach, alter your data set configurations while you have “split graph by” set to “no subdivision”, and when you have finished, only then change “split graph by” to subdivide by your desired data filter