By month avg + calculation

[LukaszR]

Hello. 

I think I have a problem with by month avg+ time series.

I have a widget (list / scorecard) showing the monthly results for a metric. It is based on the formula indicator that counts the % of SLA met items vs all the items. However, the result shown seems not to be correct.

For example, the result for Dec'24 is 7 divided by 11 ( = 72.73%), but my widget is showing 70.69%. The funny story is that when I go into details of the formula indicator, it refers to 2 automated indicators where the count is 7 & 11, so the base numbers are correct and the result of the calculation should be 72.73%. It also happens for the previous months, the result is close but not the correct one.

Any ideas?

[Community Alums]

Hi @LukaszR ,

The issue you're encountering with your monthly average and time series calculation in ServiceNow could be related to how the formula is being calculated or aggregated at the widget level, rather than an issue with the individual indicator data. One key factor to investigate is the potential difference in how the formula indicator and the widget calculate the percentage. The formula indicator might be correct when viewed in isolation, but the widget may apply additional aggregation or rounding, which could explain the discrepancy in the results. For instance, when using time series data, the widget might be averaging percentages for each day within the month rather than calculating a simple percentage based on the total SLA met and total items for the month. This aggregation could distort the results.

Additionally, it's important to verify the widget’s aggregation logic. The widget might be calculating the average percentage for each day in the month and then aggregating it, which can yield a result different from simply summing the total SLA met and total items for the month and calculating the percentage. Another thing to check is if any rounding is happening either in the formula indicator or the widget. Even small differences in rounding precision could lead to inconsistencies in the final result, so it’s important to ensure that both are using the same decimal precision.

You should also ensure that the base data is consistent between the indicator and the widget. Double-check that the data being pulled into the widget is filtered correctly and includes the right records for each month. Sometimes, discrepancies can arise if the data is not filtered correctly or if there's a delay in data updates between the time the indicator runs and when the widget is querying the data. It might also help to try creating a report based on the same formula indicator to compare the results and confirm if the issue is specific to the widget or the underlying data.

Lastly, reviewing the formula logic itself is crucial. Make sure the formula correctly calculates the SLA percentage as expected, without any unintended transformations or logic that could affect the results. If the formula is based on custom scripting or business rules, it's worth revisiting these elements to ensure that they are functioning as intended. By carefully reviewing the aggregation, rounding, and formula logic, you should be able to pinpoint the cause of the discrepancy and adjust the widget or formula accordingly

[Tai Vu]

Hi @LukaszR 

It sounds like most probably because of the Apply time series checkbox.

The Apply time series to result property controls how a Time series is applied. When using time series analysis with a formula, it is applied to every individual member/component of the formula. However, you can apply the time series calculations to just the final result of the formula calculations, by selecting Apply time series to result option. The differences are subtle, and the choice depends on objectives and personal preferences.

Example: 7d running AVG of Incident Priority 1

When Apply time series to result is checked

When Apply time series to result is unchecked

Cheers,

Tai Vu

[JudeHolt]

Thank you so much for the suggestion.