From payroll export to a filled job architecture in minutes (without handing over your pay data)

You have set up your job architecture: roles weighed, grouped, salary bands wrapped around them. Then comes the tedious part. To carry the pay transparency duty you need to know, per role, how many men and women sit in it and what they earn on average. Typing that over by hand from your payroll system is drudgery, and a single typo undermines exactly the justification you need.

The odd thing is: those numbers are already sitting there. Your payroll system (AFAS and most other packages) exports, in a few clicks, an overview of your employees with role, sex and salary. The only question is how to get that into your job architecture without retyping it, and without sending a pile of sensitive pay data out the door.

That is what the payroll import is for. This article explains how it works and, just as important, why your pay data never leaves your own computer.

The problem with pay data: privacy clashes with convenience

Most tools that promise to "connect your payroll system" do so by sending your whole file to their server. Convenient, but a legal problem: at that moment the vendor is processing the individual salaries and names of your employees. That means you need a data processing agreement, and you fall under all the GDPR duties that come with it.

We did not want that. The premise of aycabtu is that we process no individual personal data, only role descriptions and aggregated counts. A payroll import must not break that.

The solution lies in where the calculation happens.

How it works: the maths happens in your browser

When you upload your export, the file is not sent to the server. It is read in your browser and summarised there, right away, into aggregated figures per role. Only that summary (the number of men, the number of women and the average pay per role) goes to your job architecture. The rows with individual names and salaries stay on your own computer and are never sent.

In three steps:

Export from your payroll system. In AFAS (or another package) you make a CSV export with, per employee, the role, the sex and the gross salary. More fields may be in there; we ignore them.
Upload and confirm the columns. The import card at the top of the Categories tab asks you to confirm which column is the role, which is the sex and which is the salary. That is the only manual step.
Your job architecture fills itself. Per role the head counts by sex and the average pay appear, as a read-only column next to your categories. Done.

Roles from your export are matched to the roles already in your job architecture. If the import does not recognise a role, you get a clean warning instead of a wrong match.

The built-in privacy limits

Because the average of a very small group can still be traced back to a person, two hard limits are baked in:

No minimum or maximum. We store only the average, never the highest or lowest salary. Otherwise you would effectively be keeping the exact pay of your top earner.
Averages only for groups of two or more. If a role holds just one man or one woman, the "average" is simply that one salary. In that case we do not calculate it. That keeps a figure always a group figure.

The result: your job architecture is filled with exactly the aggregated figures the law asks for, and not a single individual salary has left your system. No data processing agreement needed, because there was nothing to process.

Bonus: your information-request form fills along

Employees have the right to ask what colleagues doing work of equal value earn on average, broken down by men and women. Once you have run the payroll import, that answer is already there: the information-request form fills in the average pay per sex automatically as soon as you pick a role. When an employee asks, you have the answer in a few clicks, evidence log included.

Fill your job architecture from your own figures

Already have a job architecture? The import sits at the top of your Categories tab. No job architecture yet? Look at the worked example first, or weigh your first role for free.

View the example job architecture or weigh your first role for free

The bottom line

Your pay data is already sitting in your payroll system. The payroll import pulls out the aggregated figures your job architecture needs (head counts by sex and average pay per role), without you retyping anything and without a single individual salary leaving your computer. The maths happens in your browser, only the group figures go to your job architecture, and the small-group limits make sure an average never becomes a person.

Want to see what a job architecture looks like before you fill it? Look at the worked job architecture example. And if you are wondering what the whole thing costs, read what a job architecture costs a small business.