Factorial

Factorial exposes a broad set of HR objects through its API: employees, attendance shifts, leave requests, overtime requests, contracts, project tasks, performance reviews, training sessions, and ATS candidates. This integration covers four object types — leaves, attendance shifts, overtime requests, training sessions, and project tasks — selected because they carry indicators that are directly attributable to an individual employee and resistant to post-hoc manipulation. Objects such as performance reviews were evaluated and excluded, and contracts were retained only as a source of employee identity rather than as a KPI surface. Six KPIs were retained across this scope, selected against three criteria: ability to attribute to an owner, resistance to gaming, and balance between leading and lagging indicators.

Workforce availability and time contribution

Days of Leave Taken measures the total approved leave days consumed by an employee in a given period. Worked Hours measures the total time clocked through attendance shifts. Overtime Hours Requested measures additional hours formally submitted for approval, before those hours are worked.

Reading these three indicators together reveals the actual availability profile of an employee over a period — something that no single metric can describe. A high Worked Hours figure combined with high Days of Leave Taken in the same quarter does not cancel out: it may indicate a concentrated workload compressed into fewer weeks, which is a qualitatively different situation from a steady pace across the full period. Overtime Hours Requested adds a forward-looking dimension: it is a leading signal of workload imbalance filed before the effort is expended, giving managers an intervention point that the two lagging indicators cannot provide.

The pairing of Worked Hours and Overtime Hours Requested also surfaces a structural tension. An employee with high Worked Hours and low Overtime Hours Requested may be absorbing workload informally, without triggering the administrative signal that would make that pressure visible. Conversely, high Overtime Hours Requested with moderate Worked Hours may indicate resourcing friction — work that is anticipated but structurally constrained. Neither pattern is interpretable in isolation; together, they reveal whether workload is distributed within or beyond formal capacity frameworks.

Learning and development activity

Training Sessions Completed counts the number of training assignments that an employee has finished within the period. Overdue Trainings counts the training sessions whose expiry date has passed without a completion record.

Training Sessions Completed is a lagging indicator: it records what has been done. It is structurally resistant to gaming because completion is a binary administrative event — either the training record is marked complete by the system, or it is not. This makes it a more reliable OKR target than self-assessed learning metrics or manager ratings. Overdue Trainings operates in the opposite temporal direction: it measures what has not been done despite a deadline having passed. This is a compliance pressure metric as much as a development metric, and its primary utility is as an early warning for managers before regulatory or certification risks materialise.

Reading the two indicators together distinguishes between an employee who is progressing through a development plan and one who is accumulating a backlog of deferred obligations. High Training Sessions Completed with low Overdue Trainings describes an employee who is engaging actively with assigned learning. High Overdue Trainings regardless of completed sessions may indicate that new training obligations are being added faster than they are resolved — a structural problem rather than individual disengagement.

Project output

Project Tasks Completed counts the number of tasks marked complete and attributed to the employee through Factorial's project module. This is the only KPI in this integration that measures direct work output rather than time, availability, or compliance. It is conditional on the organisation actively using Factorial Projects: teams that manage project work in a separate tool will find no signal here.

The KPI is attributable and binary — a task is either completed or it is not — which makes it structurally resistant to inflation. The main interpretive risk is attribution: Factorial allows multiple assignees on a single task, which means that the same task completion may be credited to several employees simultaneously. The connector credits all assignees, so this indicator measures contribution to completed tasks rather than sole ownership of them. It should be read accordingly: as a signal of participation in output, not as an exclusive count of individually owned deliverables.

Scope and limits of the integration

Factorial records time and administrative events, but does not measure the quality of work. Hours worked do not indicate productive output; training completion does not indicate knowledge retention; task completion does not distinguish between a well-executed deliverable and a superficially closed item. These are the inherent limits of HRIS data, and they apply to any integration built on this category of tool. Performance review scores were explicitly excluded from this KPI set because Factorial's implementation of review scores carries a high risk of gaming through reviewer alignment, and because the scoring schema varies significantly across organisations using the platform.

Beyond quality, this integration only captures activity that flows through Factorial. Informal development, unlogged overtime, verbal coaching, and cross-functional contributions that are not tracked in the tool remain invisible. The reliability of each KPI depends directly on the organisation's usage discipline: teams that clock in consistently, log leave correctly, and assign training through the platform will generate interpretable data; teams with loose practices will produce signals that are incomplete at best and misleading at worst.