Overview

Focus points detection is based on detecting deviations from a company’s normal functioning or trends of its key indicators (metrics) over time. Identifying deviations is not just a technical capability, it's a strategic asset. It helps solve critical pain points related to operational efficiency and employee satisfaction. By leveraging outlier detection, businesses can enhance their responsiveness to changes, protect assets, optimize operations, and ultimately deliver a superior employee experience.

Outlier detection is a powerful tool for deeper analysis, offering insights beyond traditional metrics. It is a process that identifies data points, events, or observations that deviate significantly from the usual behaviour and indicates important organizational issues or opportunities.

Access the Focus points analysis

Currently, the feature is implemented in the metrics that have a timeline view.

The set of currently supported metrics includes:

Headcount
Number of Starters
Number of Leavers
Number of Open Positions
Unplanned Leaves
New Hires Rate
Cost of Unplanned Leavers

☝️ Image 1: Example of focus points detected for the Number of Leavers metric on the platform

We use the Median Absolute Deviation method (see more below), which is a statistical method that helps to identify focus points. The median is selected as a more robust measure than average and less affected by extreme outliers in the data set. Identifying focus points is tied to the time range you are looking at and your preference to look at absolute numbers or percentages. When a focus point is detected on a certain metric, the point representations will include:

a highlighted visual
the score of difference, explaining how many times a metric value (point) is higher or lower than expected at some time point (date)

How focus points are detected

Median Absolute Deviation (MAD) Method

Introduction

The Median Absolute Deviation (MAD) method is a statistical tool used to understand the spread or variability of a dataset. In this documentation, we explore how we calculate focus points (outliers) using the MAD method and what those points represent in simple terms.

What are Focus points?

Focus points are data points that are significantly different from other data points in a dataset. They can either be unusually high or low values that don't fit the pattern of the rest of the data.

💡 In order to identify a data point as significant (aka Focus Point), it will deviate significantly from the median compared to other data points within a chosen time frame. For instance, if the average difference from the median for all data points is represented by a number 'K' (where K is, therefore, the expected difference), and the difference for a significant point is 'n * K', where 'n' is 3.5 or higher, it means that the significant point is at least 3.5 times more extreme than expected. The threshold n = 3.5 represents the maximum deviation (difference) tolerated in this statistical method known as MAD.

In the next releases, we expect to implement support for enabling the use of different detection methods and thresholds of tolerance.

Calculations using MAD:

Find the Median: Instead of calculating the mean, we start by finding the median, which is the middle value of the dataset when arranged in ascending order.
Calculate Deviation: Calculate the absolute deviation of each data point from the median. Absolute deviation means we're only interested in the distance of each point from the median, regardless of whether the value is higher (bigger) or lower (smaller).
Calculate MAD: Calculate the Median Absolute Deviation by averaging all the absolute deviations obtained in the previous step. This value represents the expected variation of data points in selected time range.
Identify: Once you have the MAD value, you can determine outliers. Generally, data points that are significantly farther away from the median than the MAD value will be considered outliers.

What Focus points represent?

Understanding those points is crucial as they can provide valuable insights into the dataset:

Data Quality Issues: Focus points might indicate errors in data collection or entry, highlighting potential data quality issues that need to be addressed.
Rare Events: In some cases, focus points represent rare events or extreme conditions that can either already be known to the analyst or are worth investigating further.
Important Variability: Focus points can also indicate true high variability (differences) within the dataset, suggesting that there might be underlying patterns or factors influencing the data. This might indicate a previously unknown cause that needs to be investigated.

Conclusion

There are numerous benefits of having this feature:

It enhances early problem identification by indicating potential problems, such as an unexpected attrition increase, allowing quick intervention.
Also, Focus Points can identify exceptionally performing segments of the workforce, offering insights into best practices and strategies that can be replicated.
By clearly understanding data patterns and outliers, managers and executives can make more informed decisions about training, resource allocation, and strategic planning.

The feature could be very useful in situations such as checking the seasonality of unplanned leaves or investigating the steadiness of attrition rates within your organization. The focus points feature helps you check if you have any specific periods with more unplanned leaves than usual or whether there is significant seasonality to your leavers.

💬 Need further help? Just reply with "Ask a Person" in a Support Conversation to speak with a Product Support Specialist.