Services · integration

Your largest expense, currently one line

For most companies people cost is between forty and seventy percent of the P&L, and it arrives in the ledger as a handful of summary journal lines with no department, no project, and no way to answer what a team or an engagement actually costs.

Which payroll system?

Tell us your payroll provider and how labour is currently posted. We will scope it.

1 / 3
From $6,500, fixed scopeDimensions preserved per personAccruals derived, not estimated
source chart of accountsproposed target · you approve every row6100 · Advertising6100 · Advertisingexact6110 · Ads - Google6100 · Advertisingmerged6115 · Ads – FB6100 · Advertisingmerged6200 · Contract labor6200 · SubcontractorsrenamedMisc expenseheld for reviewreviewAsk My Accountantheld for reviewreviewAI proposes the mapping · a human approves it · nothing loads unapproved

The situation

What gets connected

Per-person, per-pay-period detail with the dimensions attached, rather than a summary journal you cannot decompose.

Per-person detail

Gross, employer taxes, benefits, and deductions per person per period, which is what makes any subsequent allocation possible.

Dimensions carried through

Department, location, entity, and where time data exists, project — so labour cost lands where it was incurred rather than in a single overhead pool.

Loaded cost computed

Salary plus employer taxes, benefits, and a configured overhead allocation, per person, which is the number every margin calculation actually needs.

Accruals derived

Payroll accrual for the days between the last run and period end, computed from actual pay data rather than estimated from last month.

Time and project links

Where you run Harvest, Jira, or a timesheet system, hours are joined to cost so project margin becomes computable rather than approximate.

Sensitive by default

Individual compensation is field-level restricted. Most people who need departmental labour cost should not see individual salaries, and the model enforces that rather than trusting it.

Why payroll arrives flattened

Payroll providers post a summary journal because that is what the accounting integration was designed to do a decade ago, and because most accounting systems could not usefully receive anything richer. The result is that the single largest cost in the business is the least analysable thing in the ledger.

Everything downstream inherits that. Department P&L is an allocation guess. Project margin uses a blended rate. Headcount planning runs off a spreadsheet maintained separately from the actual payroll. None of those are hard problems once the underlying data is not flattened.

Half the P&L arrives as four journal lines, and then everyone spends the quarter allocating it back apart.

Loaded cost, per person

A blended rate applied across a team systematically misrepresents both directions: work staffed with juniors looks less profitable than it is, senior-heavy work looks better. For a services business, where staffing mix varies by engagement, the distortion is large enough to change which clients you would want.

Per-person loaded cost — salary, employer taxes, benefits, and a stated overhead basis — is more setup and it is the input every honest margin figure requires.

The accrual nobody enjoys

When a pay period does not align to month end, somebody estimates the accrual, and next month somebody reverses it and estimates again. Derived from actual per-person pay data and days elapsed, it is exact and it stops being a recurring judgement.

Sensitivity is the constraint

Payroll integration is where field-level permissions stop being a checkbox. Department heads need departmental labour cost and should not see individual salaries. Project managers need project cost and should not see rates by name.

Because exposure is enforced in the data layer rather than in the interface, that separation holds through reports, exports, and the copilot rather than only on the screen it was configured on.

Where to start

How the engagement runs

01

Review the current posting

What arrives in the ledger today, at what grain, and what dimensions exist upstream but are being discarded. Usually more than expected.

02

Map people to structure

Employees and contractors mapped to departments, entities, and where relevant projects, including the ones who genuinely split across several.

03

Build loaded cost

Employer taxes, benefits, and the overhead basis agreed and applied. The basis is a decision your controller makes and we record on every report.

04

Back-load and reconcile

Twelve to twenty-four months loaded and reconciled to the summary journals already posted, so history becomes analysable rather than starting from now.

Questions

What people ask.

Which payroll systems do you support?
Gusto, Rippling, ADP, Paylocity, TriNet, Deel, and Justworks. Others are scoped case by case depending on what their API exposes.
Do you run payroll?
No. Your provider keeps running payroll, filing, and paying. We read the results and make them analysable.
Who can see individual salaries?
Only actors you grant field-level access to. Departmental cost and individual compensation are separate grants, enforced in the data layer rather than in the interface.
How is overhead allocated?
By a basis you choose — headcount, revenue, or direct labour — shown on every report that uses it. There is no objectively correct basis, so we show ours.
Does it help with project margin?
Substantially. Loaded cost per person joined to hours is the input most project margin calculations are missing.

Make your largest cost analysable.

Tell us your payroll provider and how it posts today. We will scope it in a week.