Survival analysis delivers some really powerful insights about your business. It let’s you predict the likelihood that an employee will reach a particular tenure milestone, based on your entire history of employee turnover.
Considering the statistical methods being used to calculate the survival rate from attrition and tenure (see our Explanation Article), we have chosen to restrict the cross-filtering function of this page to ensure the most accurate representation of the data. The reason for this is that survival analysis requires us to consider an employee over their entire career at an organisation, however filtering requires us to categorise the data based on its attributes at a discrete point in time.
Even though we can’t use cross-filters like in our other analytics pages, survival analysis still provides an important representation of your organisation’s retention health, to identify any time based attrition patterns and measure improvements over time.
Why isn't this page filterable like other analytics?
Consider the following situation where we might want to filter survival by supervisor (this is an oversimplified example to explain the principle. It gets a lot more complicated when considering multiple filters):
Because filtering by Supervisor A will return the employees who report to that supervisor at the effective date of the report, we are susceptible to excluding data from employees that now reported to them in the past but report a new supervisor today. In the case above, if Supervisor A was filtered after the employee moved teams, the survival rate of this team at the 1 and 2 year milestones would be reduced, as the analysis would no longer include the employee that ‘survived’ their first two years of tenure. Similarly, the survival rate of that employee’s new supervisor would be inflated, as it would ‘receive’ the extra survival that should actually be attributed to Supervisor A.
What can we still filter?
There are certain attributes that can still be filtered on this page. We have selected these on the basis that a) they won’t affect the accuracy of the output, and b) they are important dimensions to filter for survival analysis.
- Gender: The gender filter allows you to detect key differences in retention between genders. This is an important dimension to consider for diversity reporting as well as identifying potential opportunity areas in your business. As an employee’s gender is (very) unlikely to change during their career, filtering by gender will not affect the validity of the survival calculation.
- Work Type/Class: It’s useful to be able to filter by work type and class, as often you may want to remove transient work types (casual, contract, etc) from your analysis, as you would expect that these may negatively skew your survival rates. Although it is possible that employees may move between work types throughout their career (eg casual to permanent), and this may slightly impact the accuracy of the calculations, this is a relatively rare change, compared to other attributes.
If you have any further questions about using your Survival Analysis page, please contact Customer Success.