Within each metric, Sisu offers two different analyses types: general performance and comparison.
1. General performance analyses will surface facts describing subpopulations that either overperform or underperform for a particular KPI or metric, across the entire data set.
2. Comparison analyses will surface facts describing subpopulations that either overperform or underperform relative to how the subpopulation did in either a previous time period or in a different group.
To better understand the facts, click on any fact to access the fact detail page. This displays the statistics on the table view, and also gives greater explanation as to what each statistic means.
Here are a few terms to help understand the statistics behind Sisu:
A subgroup or subpopulation describes a smaller group within the data set.
e.g. tenure_in_months = 13+ months OR household_income = 51-100K, tenure_in_months = 3-7 months
|Average / sum / rate / count||
This is the KPI you are trying to analyze, which can be summarized as either an average, a sum, a count, or a rate.
e.g. Find out what drives sum OR average session_duration, or the rate of converted = true.
Next to the measure, we include a relative difference of this subpopulation’s average. e.g. In the case below, the subpopulation tenure_in_months = 13+ months had an average session duration of 37.3 minutes, which is 34.4% lower than the rest of the population. By clicking on the text "Subgroup average", you can now order from the greatest to the least average.
|Size||The percent of total rows that a subpopulation represents. This enables you to understand how big a population is, or how much it may have grown from one period to the other.|
Impact is the contribution from a subpopulation on the overall KPI. To have large impact, i.e. to be able to drag the overall KPI up or down, a subpopulation must be relatively large in size or have a KPI that is significantly different from the rest of the population.
General Performance: Impact shows how the subpopulation affected the KPI of the overall population.
Time Comparison: Impact shows how the subpopulation affected the change in the KPI for the overall population.
Group Comparison: Impact shows how the subpopulation affected the difference in the KPI between two groups.
|Combination||Number of unique factor-value pairs that Sisu has analyzed, e.g. 1) tenure_in_months = 13+ months 2) region = Canada 3) tenure_in_months AND region = Canada constitute three unique combinations.|
Finally, because Sisu has scanned through every possible combination, we break down the same statistics for each constituent group, and allow you to compare the selected subpopulation across all other possible values.
In this case, we see that the average for the combination where household_income = 50-100K and tenure_in_months = 3-7 months is 61.2, which is greater than each individual factor.
For other values of either household_income or tenure_in_months, we see that this combination is also one of the highest.