Inaccurate worker time entries made contractors distrust our system & left the ops team cleaning up payroll for 3 days
Redesigning the time tracking flow resulted in fewer errors, more accurate hrs from workers, & reduced days of manual payroll cleanup
[I]
Final Design & Outcomes
Once we released time tracking v2, everyone did less but in the best way. Workers made more accurate submissions up front, contractors barely had to touch them because they aligned with their own logs, and our ops team didn’t have to mediate every payroll cycle. The time tracking system stopped being a problem that needed constant fixing. It wasn’t perfect, but it was going to help us scale the business.
42%
Less edits made by all parties
~2,500 → ~1,450 edits/wk
15%
Workers corrected their own time
~35% to ~50%
21%
Contractors/ops had to fix less
~45% to ~24%
20%
More time approved without intervention
~55% to ~75%
The new clock-out flow adds friction and clarity; I designed a step-by-step review with location data and 15-minute rounding, so workers know exactly what gets submitted.
End-of-week flow to double-check entries before it auto-submits on Monday. Similar to the old design, but workers can edit throughout the week and don’t have to explicitly “submit“ hours.
Contractors can view location and time entry details for each worker. This helped build trust in the system, even though the main approval surface stayed unchanged aside from new brand styling.
Workers are notified when time is approved or edited so they can dispute issues directly. This removed ops from having to chase down confirmations.
[II]
Problem Framing
Our time tracking system works like a stopwatch: easy to forget. Most workers forget to clock in or out, and about 65% of these time entries go uncorrected.
Workers get a second chance to fix their hours at the start of the week but most skip it, either unaware or assuming the original entry is “good enough.“
Contractors use this web app surface to approve worker hours but usually end up editing them. They often overwrite worker entries to match time sheets from their own systems.
Workers didn’t review or correct their hours with ~65% of entries were submitted without worker edits
Contractors often ignored worker-submitted time with 45% of entries edited by contractors
Ops had to intervene constantly with only ~55% of time sheets getting auto-approved