Frequently asked questions

The most common reasons for and ways to resolve login issues include:

• Your username is not your email address. Instead usually it's your first name; make sure to check your initial Welcome email to find your username.
• Your password is incorrect or has expired. Try resetting your password. In the login screen below the login details, click Forgotten password, input your email address, then and follow the email process to create a new password.
• Your account's (super) admin may have turned on SSO for your account, which means that your usual username and password login no longer work. Check with the PayAnalytics account's (super) admin to verify if this is the case and how to login from here.

Head over to Settings > User Management > Add User to invite your team members to join your PayAnalytics account. Here, you'll need to add some basic user details and assign them a user role.

Your PayAnalytics account is managed via role-based access. By default, your very first user is assigned the user role "super admin" which grants them access to all features and all data sets. You can create additional role-based access by heading to Navigation > User Roles > Add Role and completing the Detailed Access settings. Next, you can assign your newly created user role to users from Navigation > User > Add User and completing the user details with the assigned user role.<

Head over to Settings > User Management > User Roles to view existing user roles and permissions and to add new custom user roles and related permissions. For each role, you will see a long form of permissions and accesses which can be enabled or disabled depending on your needs. The two most common types of permissions are either feature-based or access-based:

• An example of feature-based permission might be providing your talent acquisition team member with access to the Compensation Assistant only, but restricting access to running pay equity analyses.
• An example of access-based permission might be providing your UK compensation team to UK data sets and analyses only, but restricting access to other locations data and analyses.

You can either create custom access permissions for a specific user, or you can create a user profile with specific access permissions and assign that role to the user. Creating a new user role is a more scalable way of replicating access across multiple users. Both individual access and user role access work in similar ways. Head over to Settings > User Management > User Roles to view existing user roles and permissions and to add new custom user roles and related permissions. Assuming none of the existing user roles cover your access needs, click "Add role", give the role a descriptive name, for example "Recruitment Comp Assistant Only", and then set the permissions in the form sections, specifically in the section "System feature availability". In this example, you would enable the "Compensation Assistant" feature and disable all other features including "Pay Equity Analysis" and "Job Evaluation". Save the user role and assign it to the relevant user(s).

You can either create custom access permission for a specific user, or you can create a user profile with specific access permissions and assign that role to the user. Creating a new user role is a more scalable way of replicating access across multiple users. Both individual access and user role access work in similar ways. Head over to Settings > User Management > User Roles to view existing user roles and permissions and to add new custom user roles and related permissions. Assuming none of the existing user roles cover your access needs, click "Add role", give the role a descriptive name, for example "Pay Equity UK Only", and then set the permissions in the form sections, with specific focus on the "Per-label access controls" section. If you have not done this already, ensure that your data sets and related analyses contain labels based on how you wish to restrict access, for example by country, function or business unit. In this specific example, you would label the relevant data sets as "UK" and set user permission for this role to data set "UK" only. Save the user role and assign it to the relevant user(s).

You can either create custom access permission for a specific user, or you can create a user profile with specific access permissions, and assign that role to the user. Creating a new user role is a more scalable way of replicating access across multiple users. Both individual access and user role access work in similar ways. Head over to Settings > User Management > User Roles to view existing user roles and permissions and to add new custom user roles and related permissions. Assuming none of the existing user roles cover your access needs, click "Add role", give the role a descriptive name, for example "Pay Equity UK Only" and set the permissions in the form sections, with specific focus on the "Dataset and analysis labels" section. If you have not done this already, ensure that your data sets and related analyses are labelled based on how you wish to restrict access, for example by country, function or business unit. In this specific example, you would label the relevant data sets as "UK" and set user permission for this role to dataset and analysis "UK" only. Save the user role and assign it to the relevant user(s).

Please reach out to your PayAnalytics admin user. They will be able to deactivate your two-factor authentication from the User Management section within PayAnalytics. Once you have logged in without 2FA, you will be able to reactivate it again within your account's system settings.

You cannot configure data deletion rules from within the platform, however we can set them on our side as a policy across your instance and account. We can set a rule stating that all datasets will be deleted once they reach [number] days/years of age. Note that deletion occurs at dataset-level and when a dataset is deleted, all related analyses and reports are also deleted. To request such policy, please reach out to our support team.

We currently do not enable bulk import of users.

You have two options to add users:
• Navigate to Admin and click "Users", click the "Add User" button and add user details and user role or set individual permissions.
• Through enabling SSO, which depending on your SSO configurations can provide permitted users access to PayAnalytics and can even automatically assign user roles (reach out to support to learn more).

Setting up SSO on your instance is a process that needs to be fully completed to have all users successfully log in. Please check out our article Setting up the SSO for detailed instructions to ensure all users can access the system.

Deactivated users will be shown in the 'Inactive' user list. To reactivate them, click 'Edit' and change their status to "Active".

• Average: you calculate it by adding all the numbers together and then dividing by the total number of values. It's a common measure of central tendency, but it can be affected by extreme values.
• Percentiles: percentiles divide the data into 100 equal parts. The most common quartiles are the 25th percentile (Q1), 50th percentile (Q2, which is also the median), and 75th percentile (Q3). These values help describe the spread and distribution of data. The rance between the 25th and 75th percentiles and shows where the middle 50% of the data lies, and the range between 10th and 90th percentiles shows where 80% of the data lies.
• Percentiles: percentiles divide the data into 100 equal parts. The most common quartiles are the 25th percentile (Q1), 50th percentile (Q2, which is also the median), and 75th percentile (Q3). These values help describe the spread and distribution of data. The rance between the 25th and 75th percentiles and shows where the middle 50% of the data lies, and the range between 10th and 90th percentiles shows where 80% of the data lies.
• Standard deviation: standard deviation is a measure of how spread out the values in a dataset are. It tells you, on average, how much each data point differs from the mean. A small standard deviation means the data points are close to the mean, while a large standard deviation means the data points are more spread out.

Statistical significance refers to the likelihood that the results of a study are not due to random chance. It is commonly measured by the p-value, and a low p-value (<0.05) means the result is statistically significant, suggesting the findings are likely meaningful. However, it does not guarantee the effect is large (the impact on compensation for example) or practically important, the significance is only telling you the significance of the effect.

Multicollinearity happens when two or more variables in a regression model are highly correlated with each other. This makes it difficult to understand the individual effect of each variable on the outcome. It can lead to unreliable results. To address it, the system might remove one of the related variables.

Linear regression is a statistical method used to model the relationship between a dependent variable (compensation) and one or more independent variables (objective factors, such as job role, grade, etc.), assuming a straight-line (linear) relationship. The goal is to find the line that best fits the data, which allows for predicting the dependent variable based on new values of the independent variables.

If you have set a maximum raise, as well as specified that all employees should be brought to a set outlier threshold, you might see employees receive a raise suggestion that is larger than the maximum raise. In order to bring all employees to the set outlier threshold, a raise larger than the set maximum raise is needed. In other words, bringing employees to the outlier threshold overwrites the maximum raise.

If you are working within a budget, and the budget is not large enough to fulfill all set requirements, no raises will be suggested.

Consider the following example, with two scenarios of remediation that only differ in size of max. raise cap (salary increase):

• Cap 5%, cost ~9M EUR, 5,400 addressed outliers
• Cap 4%, cost ~16M EUR, 10,000 addressed outliers

In the first scenario, fewer employees require raises to reach the target because the employees mostly affecting the pay gap receive higher raises (5%), resulting in an achievement of the target at a lower cost. The individuals affect the pay gap differently and, depending on the impact on the pay gap (cost-effectiveness) or the outlier percentage (fairness), the system calculates the raises in a specific order. With a maximum of 5% raise, the system manages to reach the pay gap target by giving larger raises to fewer employees. With a maximum of 4% raise, the system needs to give a raise to more employees in order to reach the pay gap target.

With the balance toward fairness in the raise configurations, the system starts by targeting the largest outliers and moves down the list. In the second scenario, by giving the higher negative outliers a smaller increase (4% instead of 5%), more employees need raises to meet the same target, resulting in higher costs.

The pay equity models work with full-time annualized compensation data. That means that any raise suggestions are based on and reflect the full-time and annualized equivalent. For part-time employees, the raise suggestions and associated costs of achieving pay equity targets can be prorated based on the part-time hours (if you have used the "Automatically scale raise components" option). Within your data, you will find three helpful fields: "Raise (automatically scaled), "Raise" and "Raise (%)". Of these three options, "Raise" will give the part-time equivalent of the raise, whereas "Raise (automatically scaled)"" is the full-time equivalent of the raise.

There are a few possible reasons for this, depending on your raise suggestion configuration settings:

• Size of the pay gap: the system is working to reduce the overall pay gap by aligning the average outlier percentage for one demographic group (such as women) with that of the reference demographic group (for example, men). Although a woman's outlier percentage is 0%, the system might be suggesting a raise because, to reduce the overall gap, some women at or near their predicted salary may need raises to bring the group average closer to that of men.
• Group fairness setting: if this is set to 75% for example, it prioritizes raises to women in larger groups with larger pay gaps. If the specific employee is in a group (such as country), where we find a significant pay gap affecting many women, they may be suggested a raise despite having actual meet predicted may, purely because this group of employees is prioritized for a raise. There are several possible reasons for this, depending on your raise suggestion configuration settings.
• Size of the pay gap and your target: if your target is zero, meaning you aim to eliminate the pay gap entirely, the system works to ensure that both women and men have an equal likelihood of having a positive or negative outlier percentage. Even if a woman's outlier percentage is positive (above 0%), the system may still recommend a raise. This is because, to close the overall gender pay gap, some women may still need raises to bring their salaries closer to those of men.
• Maximum raise limitations: the maximum raise allowed should be considered in relation to the size of the pay gap. If the pay gap is large, the system may quickly run out of women with negative outlier percentages, even after they have received the maximum raise. In such cases, the system may need to give raises to women with positive outlier percentages to bring their pay up to the same level as men.

The goal of the raise suggestions is to close the pay gap. In cases of a gender pay gap, only the underpaid gender will receive a raise. For example, if there is a pay gap favoring males, only female employees will receive a raise. Even though some male employees may be negative outliers (with a lower actual salary compared to their predicted salary), they will not receive a raise because the focus is on closing the gender pay gap. The objective is to ensure that both male and female employees have an equal distribution of outliers, meaning they are equally likely to be either a positive or negative outlier.

This could be due to various configurations of your raise suggestion, including:

• Group fairness: if an employee is part of multiple underpaid groups, the system will prioritize raises for that individual to address pay disparities across groups. As a result, employees not in underpaid groups may not receive raises.
• Defined target: the system is suggesting raises until the pay equity target has been reached. Therefore, depending on the size of the pay gap and the pay equity target, the number of raises may vary.
• Max raise defined: when a high maximum raise is set, the system will suggest fewer but larger raises to meet the pay equity target. Conversely, if the maximum raise is lower, the system will suggest more, but smaller, raises to close the gap.
• Close by group method: when closing pay gaps per group, the system treats each group separately. For example, if a group has a pay gap favoring males, the system will suggest raises for females. In another group, where the pay gap favors females, raises will be suggested for males.

Understanding how reference groups are selected is key to understanding if a group is eligible for a raise, and whether you will get raise suggestions to close the pay gap. Why? Because the reference groups themselves are not eligible for raise suggestions.

Selection of reference group.

A demographic group is eligible for a reference group if:

1. The headcount is at least 25% of the headcount of the largest demographic group.
2. The largest demographic group is always eligible for being the reference group.
3. The headcount is at least 10% of the entire population.
4. The headcount is at least the minimum group size threshold (the default number is 7 but can be adjusted in the System Parameters).

Once the eligible demographic groups are identified, the group with the highest adjusted pay is selected as the reference group.

Note: when closing the pay gap by groups, a reference group will be selected for each group.

The arguments for the requirements are:

The reference demographic group needs to be at least 25% of the largest demographic group, and at least 10% of the total population. This ensures, that the system will not suggest giving raises to the majority of the employees, due to the minority being higher paid. For example, 100 male and 5 female employees, with a pay gap in favor of females, the system will not suggest raises for 100 employees due to the 5 higher-paid employees. These scenarios require a manual review.

The lower-paid demographic group that has fewer employees than 7 will not get a raise. Note that this can be configured in the system parameters, but the default value is 7. This requirement ensures that raise suggestions to close a gender pay gap, where the gender groups are really small, will not be generated. These scenarios also require a manual review.

The system has a certain level of tolerance built in to 0.04% points within target. This is because - depending on the size of the dataset - the system might not be able to suggest raises that exactly (to the monetary value or percentage) meet the target. You have the option to set the pay gap target to 0.9%, to ensure the pay gap is no larger than 1%.

If the raise configuration is focused entirely on cost-effectiveness (100%), it does not take compensation drivers into account. Instead, raises are allocated to employees of the underpaid gender who are the least expensive to compensate, such as lower-paid employees.

Once fairness is incorporated into the raise configuration, the system begins to consider compensation drivers when making raise suggestions. However, it only accounts for the drivers included in the compensation model.

When the focus is on individual fairness, the system prioritizes the most underpaid employees based on the outlier percentage, which is determined by each individual's predicted salary. This predicted salary is calculated using all compensation drivers in the model. On the other hand, when fairness is assessed at the group level, the system evaluates whether there are any gender-based discrepancies within each compensation factor. For example, if the system detects a disparity in tenure, such as men being rewarded more for tenure than women, the system will adjust by granting raises to women based on their level of tenure. To conclude, the raise suggestions model takes compensation drivers into account once fairness is incorporated into the raise configuration.

PayAnalytics is now a part of beqom and as such integrates with beqom's compensation suite. Other integrations are on a custom basis and need to be discussed with the PayAnalytics by beqom account manager. Depending on the requirements, custom integrations may require significant time and extra costs.

You need to have registered to PayAnalytics with an email address that matches your organization's email domain. If you have registered with an external domain (e.g. name@gmail.com), you will not be able to gain access. In that case, you can either (re)register with your organizations domain email address, or route your support ticket through another user within your organization who does have access.

PayAnalytics API documentation and endpoints are available to all our customers upon request. In summary, our APIs enable three types of functions:

1. Import dataset (updates) in PayAnalytics from the HRIS;
2. Run preconfigured analysis on existing datasets within PayAnalytics;
3. Download the datasets and specific analysis results from PayAnalytics.

PayAnalytics provides access to API endpoints required to integrate with our system. As a customer, you will be responsible for:

• Creating the trigger events to run the above functions within/ from your HRIS (e.g. Workday). PayAnalytics API's are the receivers of these trigger events (cronjob).
• Consolidating all required data for import (e.g. from Workday) and submitting the data through our API's to PayAnalytics. PayAnalytics does not consolidate the relevant employee data from multiple data sources.

As part of the onboarding process, you would have run your first pay equity analysis, and have a view of which data fields to include and preset analysis configurations. API authentication is automatically available in every customer instance from your system settings as an admin user.

You can find your API key once logged into your PayAnalytics account, through the following steps:

1. In the top right corner click your user name to get to "Your Settings".
2. Click "API access" key from the list of options.
3. Click "Create an API access key" button to copy the key.

In certain cases, the Guide (Digital Assistant) might need a wake up call. Please refresh your browser and click the Guide icon (lightbulb), located in the right top corner of your screen, again to relaunch it. The full content should now be visible.

The standard session for a user that logs in through form based login and SSO is terminated after 60 minutes of inactivity.