Our Key Driver Analysis is an advanced statistical analysis that identifies which elements of a survey’s results have the most impact on the primary outcome that the survey is intended to achieve. The Key Driver Analysis then recommends areas for you to prioritize so that you know where to invest your time and energy for the most return on your investment.
Want to learn about how to use our Key Driver Analysis? Click here!
Want a quick overview of some of our Key Driver Analysis best practices? Click here!
Why is Key Driver Analysis important?
- It helps you invest in the right things. Each decision your organization makes to improve engagement is an investment. Since these decisions take time, effort, and money, you would expect to see a return from them. Our Key Driver Analysis helps you invest in the areas that will potentially give you the highest return on your investment.
- It presents important takeaways - in real time! Just seeing high and low scores on a survey tells you what your strengths and weaknesses are, but it may not help you pinpoint the noteworthy results. Our Key Driver Analysis does this for you as soon as you have data for it to analyze.
- It saves you time and money on analysis. With our Key Driver Analysis-supported surveys, you no longer need to pay hundreds of thousands of dollars for an external consultant or an in-house Data Science team. With the Key Driver Analysis, you can confidently present your data-driven proposals to your leadership team and colleagues. Your organization can move quickly with instant actionable insights and focus its time on action planning rather than analyzing data.
How does the Key Driver Analysis work?
In our Content Library, there is a list of TINYpulse Templates that have been scientifically designed and statistically validated to be supported with the Key Driver Analysis. These templates have a Key Driver Analysis supported label next to them. You can also find them in the Templates list when creating a survey to be sent out from the Manage Surveys page.
Each template has Outcome questions that specifically measure the primary topic that the survey is about. For example, our Engagement Survey has Engagement Outcome questions and our Benefits Satisfaction Survey has Benefits Satisfaction Outcome questions. These Outcome questions represent your users’ feelings and perceptions around a particular topic.
Each template also has Driver questions that measure more defined areas within the survey’s topic that can potentially impact the overall Outcome. Driver questions are actionable and give you exact items to focus on.
When is the Key Driver Analysis available?
While we recommend that you send Key Driver Analysis-supported surveys to your entire organization so that your initiatives will have a company-wide impact, you can still send a Key Driver Analysis-supported template survey to any group of your TINYpulse users as long as that group has at least 30employees.
Our Key Driver Analysis works the best with a larger audience - the bigger the audience, the stronger the analysis will be.
On the reporting side, anyone who has Super Admin or Admin permissions for Engage will have access to Key Driver Analysis. And, if you send out a survey to Attributes or Saved Groups with more than 30 users in them and collect at least 30 Responses, you will also be able to view Key Driver Analyses that are Attribute or Custom Group-specific that Admins and the Group Admins or Viewers of these Groups can view, too.
How is Key Driver Analysis calculated?
- The Outcome Score is calculated using the average of the Outcome questions.
The Driver Impact is calculated using Kendall’s tau-c, a statistical technique that measures the strength and direction of association that exists between two variables. Kendall’s tau-c evaluates the relationship between the outcome index and every other driver question in the survey. We then rank all the evaluations calculated. This technique involves looking at all comparisons between the outcome and each of the drivers for each of the Responses then comparing the data trending with every other Response's trending. Though the statistics are intuitive and simple, it is an intense calculation and evaluation with this many comparisons to find the trend.
- To determine the strength of a Driver Impact on the Outcome (x), we use the rules below from Kendall's tau-c calculation:
Extremely High: x > 0.6
Very High: 0.5 < x <= 0.6
High: 0.4 < x <= 0.5
Medium: 0.3 < x <= 0.4
Low: 0.2 < x <= 0.3
Very Low: x <= 0.2
Non-Significant: anything with p-value > 0.05
We use the average for all Driver Scores as the Score reference. Plotting all the Drivers on the Key Driver Quadrant will give you the Recommendation as below.