FAQ: Impact circles on drivers

This article contains frequently asked questions about the impact circles that display on driver cards in the main dashboard.

The impact circles are not the same feature as the impact metrics report, which enables you to upload external metrics for comparison to Peakon question scores. See Measuring engagement impact on business metrics for more.

General questions

What are the impact circles?

Each driver has some impact on the main score of its associated question set.

Impact circles display on standard question sets. Example: Impact circles on the Non-discrimination driver refer to the driver's impact on the overall Diversity & Inclusion score.

Impact circles have 3 classifications: low, medium, and high, and they display as a venn diagram. Example: A high overlap of the circles indicates high impact.

These classifications are in relation to other drivers within the same question set. We recommend focusing on improving priority drivers with high impact, where available.

How does Peakon calculate driver impact?

For each organization and segment we use statistical modeling to draw inference on all drivers and subdrivers, extracting the drivers and subdrivers that have most influence on the main score.

How can we assess whether our actions affected the driver's impact? Is it possible to improve a driver's impact?

Driver impact isn't something you should target or seek to improve, as it will naturally change over time. Improving a driver score won't reduce its impact on the main score, as driver impact doesn't follow a decrease or increase in scores. Both priority and strength drivers can have an impact on the main score.

Instead, we recommend assessing the effect of actions on improvements in scores, and in priority drivers being deprioritized.

Is it possible to see the impact circles over time?

We don't display historical classifications of impact circles in the user interface, however this can be provided as a paid service by our data science team.

How sensitive are the impact circles to small changes in employee feedback? 

It's possible for the impact circles to update without a noticeable change in driver scores. This can be due to a significant change in the pattern of scores (the correlation structure between drivers) that affects driver impact but not the average score.

Why would the impact circles change on the overall dashboard when a smaller survey takes place?

Impact circles only recalculate if both conditions are true:

• The aggregated participation rate for the dashboard is more than 25% since the last update of strengths and priorities.
• At least 30% of the segment's employees were in a survey since the last update of strengths and priorities.

Do survey scores from small segments contribute to impact calculations elsewhere?

Yes. Even though Peakon doesn't calculate driver impact for segments with less than 30 employees, the scoring patterns of those employees still contribute to impact calculations in the larger segments that they’re part of and at company level overall.

Why are the impact circles sometimes missing from the Drivers page?

If all drivers have low impact, the impact circles don’t display for the specific question set within the Drivers page.

In segments with more than 30 employees, the impact circles display:

• On driver cards on the main dashboard for each question set, regardless of the impact classifications.
• In the Drivers page of each question set, if there are drivers with impact levels other than low.

Do you provide the raw data behind the impact circles or is it possible to export it through API?

We don't display the figures behind impact circles and it’s not possible to export it using API. If, for example, you have 3 priority drivers with high impact, they should be considered equally impactful.

Are there any driver combinations that have more impact?

Example: Do high scores for Autonomy and Recognition lead to higher overall engagement than high scores for Strategy and Management Support?

We don’t see any driver combinations that have a higher (and statistically significant) impact on the main scores - it varies by company and segment.

Situational questions

Is it normal that there's not much diversity in our impact levels? All drivers have low or medium impact.

It's common for organizations to only have 2 impact levels for drivers.

How can the impact circles have a big shift despite no major company initiatives?

The correlation structure between the drivers in the model can change based on changing employee scoring patterns, without any apparent large change in scores.

How can 2 drivers have the same benchmark difference and impact, but only one is a highlighted driver?

Example: 2 drivers (whether on the same or different question set) are 0.2 above benchmark and both have high impact. Only 1 of these drivers is a priority driver.

The impact circles and difference to benchmark use rounding in the user interface, however the prioritization calculation uses raw data. Referring to the example, it's likely that 1 of the 2 drivers had a higher unrounded difference to benchmark or higher impact compared to the other driver.

How can 2 driver cards have a different impact classification if they're using the same question?

Example: An organization has deactivated the Meaningful work core driver question, but has activated the Significance question, which is a subdriver of Meaningful work. The score is the same for the 2 driver cards due to having the same question, but impact is different.

In situations where only a subdriver question is active and the subdriver and overall driver score are equivalent, the impact calculation only applies to the driver.

We believe that every driver and subdriver has some impact on the main score. We’ve therefore made a design decision to display a low impact classification on the subdriver in this edge case, even though the subdriver wasn’t part of the impact calculation.