Sections in this category

Interpreting a Time Comparison KDA

  • Updated

When investigating a metric, understanding what’s driving the performance and which subgroups might be driving this KPI is key. For a Time Comparison Key Driver Analysis (KDA), Sisu compares two timeframes that you define, then returns key statistics for the metric selected, as well as min, max, median, average, sum, the number of rows to help validate the quality of the data used. The Analysis results also include detailed information about facts that impact the timeframe comparison and performance of the chosen metric.

info_icon.png For more explanation on Time Comparison Analyses in Sisu, please refer to Understanding Key Driver Analyses.

 

tip_icon_-_small.png

At any point, you can Run the Analysis again to refresh these results, such as when new data becomes available, or when you have made adjustments to the Analysis settings or underlying query.

In addition, you can set up your Analysis to run automatically at an interval you can define. Refer to Scheduling KDA Auto-Runs for more information.

 

Related Articles:



General Layout of the Analysis Results

The following image provides an overview of the information and actions you can take from a Time Comparison Analysis results screen.

TC_Overview.png

 

The Key Statistics Cards

These cards provide a quick overview of key statistics for all records that meet the criteria set in the configuration of the Analysis that Sisu uncovered as a result of the Analysis.

In the example, this Time Comparison KDA based on the metric “Average Order Value - Digital” analyzed the average order value for all orders generated from the Order Channel “digital_mobile” in the data (this was set as a filter in the configuration), and compared the results between two time periods.

 

TC_Key_Card.png

 

The two cards on the left display the following stats for each time period defined:

  • The average of the metric’s column within the data
    (in this example, average digital order value, since that is the metric selected for the Analysis)
  • The minimum, maximum, median values of the metric’s column within the data
  • The total sum of the metric’s column within the data
  • The number of rows included (records in the data - in this case, it’s the number of digital orders) 

The card on the right displays graphs of the data for each defined time period. Note that on the graph, you can turn on the Smoothing option. Short-term fluctuations in time series plots can often obscure the overall trends. Sisu's Smoothing feature automatically makes time series as smooth as possible while preserving long-term deviations, by preserving the kurtosis measure of the original plot (which in turn preserves long-term deviations).This allows for an improved user perception of deviations in time series.

TC_Smoothing.png


The Top Drivers Subtab

This subtab is displayed by default, and helps you identify what is meaningfully impacting metric performance in each defined time period so that you can take action with confidence. Each line of the table represents a “Subgroup”, or a subset of rows in your data determined by the factor, that had a statistically significant impact on the metric. 

Top_Drivers.png

The following table describes each of the columns in this view.

tip_icon_-_small.png Some Analyses include many, many rows. You can use the filter icons in each column to filter the rows using properties applicable to that column, if desired.

 

Subgroup

A subgroup is defined by a set of factors. A factor is a column-value pair (such as order_store_city = new york). Some subgroups are determined by multiple factors, referred to as "2nd order facts" or “3rd order facts”.

TC_expand_collpse.png

The name of each subgroup describes which column and value pair it refers to.

For example, “ORDER_STORE_CITY = New York” refers to order records that were in the New York store only.

Use the expand/collapse icon to view/hide details about certain subgroups:

In this example, Sisu identified additional information about sales in New York stores that impacted the metric’s performance, such as time of day, gender, and coupon information. Refer to Understanding Facts & How They Are Grouped.

Use the column’s filter icon to include only orders in a certain city or SKU, for example. Refer to Sorting & Filtering the Fact Table.

Click on a Subgroup name to display more details. Refer to Exploring & Drilling Down Into a Fact.

Subgroup size

Describes how big each subgroup of the data for each time period is as a percentage of the total. For example, if there are 100 rows in the data and the subgroup size is 40%, it means that 40 rows out of the 100 match that criteria.

In Time Comparison Analyses, the sizes are displayed as follows:

<first time period> vs <second time period> 

TC_comparison_vs_2.png

The percentage shown highlighted in grey is the difference between the two. In the example above, the subgroup’s size during the second time period increased over the first time period by 5.5%. (Negative percentages indicate a decrease in size.).

Use the column’s filter icon to include only subgroups that are over or under a specified size. Refer to Sorting & Filtering the Fact Table.

Subgroup metric


Note: this column reflects the calculation type that was used for the metric.

Describes the average (or sum or count) for each subgroup or subset of rows in the data for each time period defined. This column will reflect the calculation type that was selected, so it could be a Subgroup sum, average, or count in General Performance Analyses.

In Time Comparison Analyses, the values are displayed as follows:

<first time period> vs <second time period> 
TC_comparison_vs.png

The percentage shown highlighted in grey is the difference between the two. In the example above, the subgroup’s metric during the second time period increased over the first time period by 8.8%. (Negative percentages indicate a decrease.)

Use the column’s filter icon to include only subgroups that are over or under a certain average value. Refer to Sorting & Filtering the Fact Table.

Impact column

Describes each subgroup’s impact on the overall metric performance for each defined time period.

Subgroups may overlap. Refer to Understanding Impact for details about impact and how it is determined.

Use the column’s filter icon to include only subgroups that have an impact value over or under a certain number. Refer to Sorting & Filtering the Fact Table.

 

There are several actions you can perform for each fact in the table. 

Expand facts containing further details.

Some facts have an expand icon to the left of the subgroup name, indicating that Sisu identified additional information about the fact.

TC_expand_collpse.png

In this example, Sisu identified a related fact (time of day) that impacts the original fact (store is in New York), and provides additional statistics about that combination (morning sales in New York stores).

Click on a fact to view its detail page

Refer to Exploring & Drilling Down Into a Fact.

Drill down into a fact

Use the drill down icon at the end of each row to display further information about the fact. 

drilldown_icon.png

Refer to Exploring & Drilling Down Into a Fact.

Add a comment

Use the additional options menu at the end of each row (hover to display) to hide a fact for everyone.

Add_Comment_to_Fact.png

Refer to Adding a Comment.

Hide a fact

Use the additional options menu at the end of each row (hover to display) to hide a fact for everyone.

Hide_Fact.png

This is a toggle selection also allowing you to show hidden facts.

Refer to Hiding a Fact.

info_icon.png For more details on exploring facts, refer to Exploring & Drilling Down Into a Fact.

 

Finally, the default view of the Top Drivers tab is the table view. You can choose to view the same information in “natural language view”.

Natural_Language_view_-_general.png

This option is a toggle, so you can easily switch between the two views.

info_icon.png For more details on the natural language view, refer to Fact Table: Natural Language & Table Views.

 

 

The Waterfall Subtab

This subtab displays data for the analysis in the form of a waterfall plot.

Waterfall_TC_Analysis.png

info_icon.png Refer to Using Smart Waterfall Plots to Visualize Key Data for a description of how to interpret this screen.



The Query Preview Subtab

tip_icon_-_small.png This subtab is only visible if the underlying data is based on a query. If the underlying data is a table with no query applied, the Table Preview subtab will be used instead.

This subtab in the Analysis results allows you to explore a preview of the underlying query used to generate the analysis.

Query_Preview_Tab.png

 

info_icon.png For more details, refer to Previewing the Underlying Data Table.

 

 

The Query Subtab

tip_icon_-_small.png This subtab is only visible if the underlying data is based on a query. If the underlying data is a table with no query applied, the Table Preview Subtab will be used instead.

This subtab in the Analysis results allows you to view the actual underlying query that defines the metric used within this analysis.

Query_Subtab.png

info_icon.png For more details, refer to Creating & Managing Queries.

 

 

The Table Preview Subtab

tip_icon_-_small.png This subtab is only visible if the underlying data is not based on a query. If a query is used for the analysis, the Query Preview Subtab and Query Subtab are displayed instead.

This subtab in the Analysis results allows you to view the data table used to generate the analysis.

info_icon.png For more details, refer to Previewing the Underlying Data Table.