WHEN TO USE THIS ARTICLE
Use this article when:
you have successfully uploaded your employee data into PayGap
you have set up Comparable Jobs
you want a first view of where pay gaps exist in your organization, before going deeper
Important: Base Analysis is not your final result. It gives you a baseline view of unadjusted pay gaps - where they are, how large they are, and what is likely driving them. You use this view to decide where to focus during Drill-Down Analysis and Adjusted Pay Gap analysis.
If you need help interpreting the figures you see on the dashboard (mean vs. median, quartiles, negative gaps), see Understanding the measures in Base Analysis
OUTCOME OF THIS STEP
By the end of Base Analysis you should have:
A clear read of your overall pay gap - unadjusted mean and median gaps
A view of representation — how men and women are distributed across the four salary quartiles
A shortlist of comparable jobs to investigate - categories with the largest unadjusted gaps (especially those >=5%)
A view of which pay components drive the gap - base pay, variable pay, or benefits
A primary driver hypothesis -representation-driven, within-role pay-driven, pay-component-driven, or mixed
A list of data concerns to flag or clean up before the adjusted analysis
A list of open questions to raise with PayGap, HR Admin, Managers
You do not end this step with conclusions or actions. Those come after Drill-Down and Adjusted Pay Gap analysis.
STEP-BY-STEP GUIDE
1. Open Dashboard
In PayGap, open Dashboard from the left-hand navigation. The dashboard has three main sections:
Overall pay gap figures (top tiles)
Distribution based on salary level (quartiles)
Gender pay gaps by comparable job
2. Review the overall unadjusted pay gaps
At the top of the dashboard you see three tiles:
Mean unadjusted gender pay gap — the percentage and monetary difference between the average salary of men and women
Median unadjusted gender pay gap — the percentage and monetary difference between the middle salary of men and women
Adjusted gender pay gap — the gap remaining after controlling for legitimate pay factors (shown here as a reference only)
Each tile shows both the percentage difference and the underlying men's and women's salary values, so you see the absolute amounts behind the percentage.
A meaningful difference between the mean and the median usually points to a small number of high earners pulling the mean up. To understand what the two measures mean and how to read them together, see Understanding the measures in Base Analysis.
How to read each tile (example: Mean unadjusted gender pay gap):
Men's mean salary — the average salary across all men in the dataset, regardless of role, job family or seniority
Women's mean salary — the average salary across all women in the dataset, on the same basis
Unadjusted mean gender pay gap (amount) — the absolute difference between the two averages, in your reporting currency
Difference (%) — the same gap expressed as a percentage of the men's mean salary
A positive value means women earn less than men. A negative value means men earn less than women. The same logic applies to the Median tile (using middle salaries instead of averages).
Ask yourself:
How large are the mean and median gaps, and in which direction do they go?
How far apart are the mean and median (in percentage points)?
Red flags:
Mean gap noticeably larger than median (more than ~5 percentage points apart) — a small group of high earners, usually men, is pulling the mean up
Very large figures (e.g. mean gap above 35%) — check for data quality issues (see the Data concerns flag in Step 6)
3. Review the distribution by salary level (quartiles)
Scroll to Distribution based on salary level. PayGap divides your workforce into four equally sized groups, sorted by salary:
Upper Quartile — the 25% highest paid employees
Upper Middle Quartile — from the median up to the 25% highest paid
Lower Middle Quartile — from the median down to the 25% lowest paid
Lower Quartile — the 25% lowest paid employees
Look at the gender split in each quartile. A strong concentration of men in the Upper Quartile (and women in the Lower Quartile) is one of the most common drivers of unadjusted pay gaps.
For a deeper explanation of how to read quartile distribution, see Understanding the measures in Base Analysis.
4. Review gender pay gaps by comparable job
Scroll to Gender pay gaps by comparable job. This table shows, for each comparable job:
Median Pay Gap
Mean Pay Gap
Adjusted Pay Gap
Number of men and women in the category
By default, the table shows only comparable jobs with a gap of at least the 5% threshold. This helps you focus on the jobs that will need follow-up. You can switch the filter off at any time using the toggle above the table.
Tip: You can download the comparable jobs results using the export icon next to the threshold toggle. Save the file before each new upload — this lets you compare comparable-job results across uploads and track how specific jobs evolve over time.
Use the page-size control at the bottom of the table to display more groups per page. By default, the table shows only the first 10 comparable groups
Ask yourself:
How many comparable jobs sit above the 5% threshold?
Which 2–3 comparable jobs have the largest mean gaps?
Are those jobs large (many employees) or small (few employees)?
Are there any negative gaps (women earning more)? Do you understand why?
Red flags:
A small number of comparable jobs dominating the overall gap — they become your main focus in Drill-Down
Many jobs with similar small-to-medium gaps in the same direction — the pattern is structural, not localized. Do you have a hypothesis why? For example, are certain job families dominated by women (or by men)?
Comparable jobs with very low headcount (for example, fewer than 5 men or 5 women) producing extreme gap values — these figures are statistically unreliable but should be also flagged. This often happens when one gender heavily outnumbers the other: the gap can look very large, but with so few people on one side it does not point to systematic discrimination. Before drawing conclusions, we will look at the characteristics of both groups — performance, seniority, specific skills paid at a premium — because in small groups single cases drive the result instead of averaging out.
Tip: You can download the comparable jobs results using the export icon next to the threshold toggle. Save the file before each new upload — this lets you compare comparable-job results across uploads and track how specific jobs evolve over time.
5. Filter by pay component
Use the pay component filter at the top of the dashboard to isolate how individual pay components contribute to the gap. Run the analysis for at least:
Base salary only
Variable pay only (if applicable)
Total cash (or your full reporting measure)
Ask yourself:
How does the mean gap change between base salary and variable pay?
Does a single component account for most of the overall gap?
Does the quartile distribution change when the filter changes?
Red flags:
Variable pay gap significantly larger than the base salary gap — a very common pattern, often driven by bonus or commission differences
A specific allowance or benefit with an outsized gap — can point to a legacy policy issue
Total cash gap noticeably different from base pay gap — indicates that what is reported publicly (often total cash) may tell a different story from fixed pay
6. Form your hypotheses for Drill-Down Analysis
You finish Base Analysis with a clear view of where the gap is coming from — not a final answer, but a prioritized list of hypotheses to test in Drill-Down and Adjusted Pay Gap analysis. Use the findings from Steps 2–5 to fill in the four blocks below.
Primary driver hypothesis (pick one)
Representation-driven — the gap is mainly caused by men and women being unevenly distributed across salary levels. Signal: unbalanced quartiles, mean much larger than median, comparable jobs relatively balanced.
Within-role pay-driven — men and women in similar roles are paid differently. Signal: balanced quartiles, mean close to median, comparable jobs show large gaps.
Pay-component-driven — a specific component (usually variable pay) drives the gap. Signal: large difference between the base-only and variable-only filters.
Mixed — more than one of the above contributes meaningfully.
Areas to investigate in Drill-Down
The 3–5 comparable jobs with the largest unadjusted and/or adjusted gaps (start with at least 5% gaps)
The pay component(s) with the largest gap
Any specific population, legal entity, or location you suspect from Step 3
Data concerns to flag
Comparable jobs with very low headcount producing unreliable values
Obvious outliers spotted during upload review (Case Management, Descriptive Analysis)
Pay components that were set up incorrectly or need revisiting
Any legal entity or country with unexpectedly extreme figures
Open questions for stakeholders
Specific jobs where you need HR or business-line input before the adjusted analysis
Variable pay rules that may explain a variable-pay gap
Promotion, hiring, or retention patterns that may explain the representation imbalance
Any known one-off events (restructurings, acquisitions) that could skew the numbers
Save these four blocks. They are the input for Drill-Down Analysis and the Adjusted Pay Gap analysis.
KEY INFORMATION
Base Analysis always shows unadjusted figures. No controls are applied for role, seniority, tenure, or other factors. The Adjusted Gap tile is included as a reference only; the full Adjusted analysis is a separate step.
Before finalizing your Base Analysis, make sure you have tested different grouping scenarios. See Comparable Jobs — set up and review in PayGap.
Quartile boundaries are recalculated with every new upload. Your distribution view will change as your dataset changes.
Base Analysis ends with hypotheses, not conclusions. Resist the temptation to draw final statements here.
Example 1: Representation-driven gap Mean: 32,3%, Median: 28,5%. Upper Quartile: 91% men. Comparable jobs mostly balanced. Base salary gap similar to total cash gap. → Hypothesis: representation-driven. The 28.5% gap is primarily caused by male concentration at the top of the salary scale. → Next step: Drill-Down focused on Upper Quartile composition and senior-level compensation.
Example 2: Pay-component-driven gap Mean gap (total cash): 28.5%. Base salary only: 12%. Variable only: 45%. Quartiles moderately imbalanced. → Hypothesis: pay-component-driven. Base pay differences are real but secondary — variable pay is the main driver. → Next step: investigate bonus and commission policies and payout patterns during Drill-Down and Adjusted analysis.
Example 3: Localized within-role gap Mean: 18%. Median: 16%. Quartiles roughly balanced. 2 out of 15 comparable jobs show gaps above 20%; the rest are below 5%. → Hypothesis: within-role pay-driven, but localized. The overall figure is driven by pay differences in 2 specific jobs, not a systemic pattern. → Next step: deep-dive those 2 jobs first; use the Adjusted analysis to confirm which factors explain (or fail to explain) the differences.
COMMON QUESTIONS / TROUBLESHOOTING
What does a negative pay gap mean?
A negative pay gap means women earn more than men on that measure. The sign follows the standard formula (men's salary minus women's salary, divided by men's salary). Some country regulations use the opposite convention — in that case your official annual report will follow the local rule, but the dashboard always uses this standard formula.
Why are the mean and median pay gap different?
The mean is sensitive to outliers — a few very high salaries pull it up. The median is not. A large difference between the two usually means a small number of high earners has a disproportionate effect on the average. For a full explanation with examples, see Understanding the measures in Base Analysis.
Why is my quartile distribution not 50/50 between men and women?
Quartiles group employees by salary, not by gender. Each quartile contains the same number of employees, not the same share of men and women. The split in each quartile also depends on the overall gender ratio in your organization — if women make up only 30% of the workforce, you would not expect 50% women in any quartile. What matters is whether their share is stable across the four quartiles, or rises or falls as you move from the Lower to the Upper Quartile.
Does the Adjusted Pay Gap tile on this page mean my analysis is done?
No. The adjusted figure on this page uses default settings and is shown here as a reference only. The full Adjusted Pay Gap analysis is a separate step, where you select pay factors and review the results in detail.
How confident should I be in my hypotheses at the end of Base Analysis?
Treat them as working hypotheses — the best current read of the data. Drill-Down and the Adjusted analysis are explicitly there to confirm or disprove them. It is completely normal to revise or drop a hypothesis after the adjusted analysis.
I changed my Comparable Jobs setup. Do I need to re-run Base Analysis?
No re-run is needed — Base Analysis refreshes automatically when Comparable Jobs are updated. Your numbers will change, though, so you should review the dashboard again after any change in groupings. The hypotheses you formed before the change should be re-examined.