04:05 GLOBAL
One Shift, Three Systems In theory, payroll should receive a single, clean data set. In practice, agency shifts pass through at least three systems before they reach payroll: the roster showing what was planned, the Time and Attendance (T&A) system recording what was worked (either electronically or via manual timesheets), and the agency timesheet or invoice documenting what the agency believes occurred. As soon as start and end times differ, breaks are taken differently, or overtime is treated inconsistently, the data diverges. From that moment, payroll is no longer processing timesheets. It is reconciling competing versions of reality.
What This Looks Like For Payroll In one healthcare environment I worked with, agency‑related discrepancies occurred on almost every invoice received. Payroll officers spent hours reconciling invoices, investigating overpayments and underpayments, and moving between systems to determine which data source was most accurate. Sometimes, the internal T&A data best reflected reality. Other times, the agency data was closer. With no consistent source of truth, every discrepancy became a fresh investigation. As one payroll team member put it, “It’s a hunt every time we reconcile agency
invoices.” That constant ambiguity created frustration, exhaustion, and a sense of blame for issues payroll had not caused. Where the Data Diverges Most Across that environment, two issues caused the majority of payroll pain. The first was start and finish times. Small differences introduced through rounding rules or varying capture methods compounded quickly across hundreds of shifts. The second was breaks. In clinical settings, breaks are not always taken as planned. Internal systems may capture actual behaviour, while agency processes often assume breaks unless notified otherwise. Neither approach is inherently wrong, but payroll must reconcile the outcome. Why Does This Land So Hard On Payroll? In many healthcare
As soon as start and end times differ, breaks are taken differently, or overtime is treated inconsistently, the data diverges.
16 I 04:05
GLOBAL PAYROLL MAGAZINE ISSUE 23
Made with FlippingBook - Share PDF online