Skip to main content

Understanding key driver analyses

Sisu provides three types of analyses so you can explore your data and gain valuable insights into your business.

Understanding Sisu Analysis in General

All Sisu analysis types described in this article use our algorithms to explore your data and provide you with valuable insights into your data that you would not otherwise be able to easily discover on your own.

The insights Sisu provides can lead you to impactful decisions, enabling you to take advantage of hidden opportunities and mitigate risks to your organization.

To do this, Sisu identifies subgroups that drive your key metric and calculates relevant statistics for each. Sisu then  uses our algorithm to determine which subgroups are statistically significant.

In all analysis types, Sisu will show you:

  • The size of each subgroup, as a percentage of your total dataset
  • The metric’s value in each subgroup
  • The relative impact of each subgroup
  • And much more


General Performance Analyses

A General Performance Analysis provides an “overall” analysis of your data, providing insights into overperforming or underperforming subgroups in your data and much more. This type of analysis is useful for understanding your key metric’s performance and identifying top drivers that affect that metric.

For example, let’s say you have identified “Average Order Value” as your key metric, because your business objective is to increase that value. Sisu can show you which product categories or SKUs are affecting your average order value the most, either negatively or positively. The Analysis presents these subgroups as “facts”.

Furthermore, Sisu can show you which combinations of factors are significantly affecting your order value, such as sales for peanut butter granola bars on Mondays in the New York location were very high. (Insights like this might encourage you to feature that product on that particular day.)

Time Comparison Analyses

A Time Comparison Analysis compares data in two different periods of time against each other. This type of analysis is helpful in identifying top drivers that explain the change in performance from one period of time to another.

For example, if your key metric is “Average Order Value”, Sisu can show you that your average order value decreased this quarter over last quarter, as well as identify what factors and combination of factors led to that decrease, such as location, age groups, or coupon codes.

As another example, with this type of analysis you might discover that a particular marketing campaign had twice the conversion rate this quarter compared to last quarter. The Analysis presents this timeframe difference for that marketing campaign as a “fact”.

Group Comparison Analyses

Group Comparison Analysis compares the data between two groups against each other. This type of analysis can be useful for identifying top drivers that led to a difference in performance between two groups.

With this type of analysis you might discover that, for instance, a marketing campaign had twice the conversion rate within the mid-market segment (Group A) compared to its conversion rate within the enterprise segment (Group B).

Groups are mutually exclusive subgroups of data. Common examples include:

  • Sales from TX versus Sales from CA
  • Completed orders versus uncompleted orders
  • Sessions that converted versus sessions that did not convert
  • A control group versus an experiment group in A/B testing

For example, if your key metric is “Average Order Value”, Sisu can show you that the difference in average order value between CA (Group A) and NY (Group B) is higher among certain product categories. The Analysis presents this subgroup as a “fact”.