CostCtrl Demo: Shared Services Cost Dashboard

One shared services centre. Three answers about who really consumes it.

Everything below is computed live from the embedded transactional data of a fictional EUR 21M group business services organisation: invoices, tickets, onboardings, requisitions, minutes and capacity. Change the filters above and every engine recalculates.

Chargeback cascade: billings to net position (TDABC view)

Where 100 cents of chargeback go. Unused capacity is shown as its own block, not buried in transaction costs.

Practical capacity set at 85% of theoretical, following Kaplan and Anderson's TDABC convention, so transaction costs are not inflated by the cost of idle resources.

The method-divergence problem

Same centre, same ledger, different truths.

Monthly trend: chargeback, TDABC net position and Finance Operations utilisation

The chargeback is flat every month; the work is not. Year-end close peaks in January and December, August is the trough.

Flat FTE chargeback vs ABC vs TDABC: who is lying to you?

All three engines allocate the same operating cost pool. They just disagree, sometimes by 30 to 46%, about who caused it. Research on costing-system distortion reports errors of this magnitude on the extremes (IJISR).

How each engine thinks

The plain-language version.

Flat FTE chargeback (today): every business unit pays the same rate per supported employee, adjusted only by legacy negotiated indexes. Consumption is invisible: a BU that files 30 invoices per head with 25% mismatches pays the same rate as one that files 12 clean e-invoices.

ABC: cost pools with transaction drivers (per invoice, per ticket, per requisition). Better, but every driver is an average, and all capacity cost is pushed onto transactions, so idle time silently inflates every rate and behaviour differences vanish.

TDABC: each resource group gets a capacity cost rate in EUR per minute; time equations estimate the minutes each invoice, ticket, onboarding and data change actually consumes. Unused capacity is isolated instead of allocated.

Two real time equations from this model

This is what replaces allocation keys. Every coefficient is minutes.

invoice_processing_min = 3.2 (base)
+ 1.1 × invoice_lines
+ 14 if exception_or_mismatch
+ 6 if manual_approval

it_ticket_min = 7 (base)
+ 22 if high_priority
+ 22 if escalated_to_L2

Rates used here: Finance Operations 0.97 EUR/min, HR Services 1.12, IT Service Desk 1.04, Procurement 1.13, Master Data 1.19 (resource group cost / practical capacity minutes).

Cost per transaction and net position by engine

Cost per transaction under each engine. Arrows show rank moves from flat chargeback to TDABC.

Flat-chargeback margin vs TDABC margin

On the diagonal, the methods agree. Off it, your chargeback letters were negotiated on fiction. Points below the line are subsidised by the ones above it.

Why they differ: the flat rate sees only headcount, so behaviour-heavy units (paper invoices, mismatches, escalations, churn) are subsidised by clean high-volume ones. ABC catches transaction counts but averages away exceptions and priorities. TDABC prices the minutes.

Under the hood: the full model card

Every parameter behind the three engines. In an engagement these come from your GL, payroll, ticket dumps, invoice registers and workflow logs.

Capacity cost rate = resource group cost / practical minutes (85% of theoretical, Kaplan & Anderson). Each tower also carries a per-run fixed load (queue triage, batch jobs, close checklists) and a pass-through fee per transaction (OCR, e-invoicing network, background checks, licenses).

The whale curve: a few business units carry everyone else.

Sort business units from most to least over-recovered (TDABC) and accumulate. The curve climbs far above 100% of the net position, then the subsidised tail gives it back. Kaplan's Kanthal case found 225% at the peak; the same physics applies to internal customers of a shared services centre.

Cumulative net position curve (TDABC)

Hover a point for its P&L. Click to pin the card and see the three levers for the subsidised tail.

Detail: select a point on the curve

Full TDABC P&L of the hovered or pinned item.

Hover or click any point on the whale curve.

Fixing the tail: the three levers

The playbook for subsidised business units. Nobody gets fired: the point is to make consumption visible and priced.

1. Fix at source

Reduce the minutes: e-invoicing adoption, PO compliance, clean vendor master data, self-service password resets, structured onboarding packs. Most exceptions are manufactured upstream, in the business unit.

2. Behaviour and policy

Change the rules that create work: approval thresholds and delegation matrices, catalogue buying instead of free-text requisitions, ticket triage discipline, master data request quality gates.

3. Consumption-based chargeback

Publish a service catalog priced from time equations: per invoice, per ticket by priority, per onboarding, with exception and manual-approval surcharges. Business units self-correct when behaviour has a price.

Unused capacity: the cost line almost nobody measures.

An IMA field study found only 3 of 63 companies measured the cost of unused capacity. TDABC makes it a standing line item per resource group, following the CAM-I capacity model: theoretical, practical, used, idle.

CAM-I capacity bridge by resource group

Theoretical minutes, the 15% practical haircut (breaks, training, team meetings), productive use, and paid-for idle time.

Only 3 of 63 companies in the IMA capacity study measured the cost of unused capacity. Everyone else pays for it and then hides it inside the chargeback.

Utilisation of practical capacity by month

Finance Operations peaks at year-end close; the August trough idles every team while the chargeback keeps billing.

EUR of idle capacity, by resource group and month

Resource cost is fixed monthly; idle EUR = (practical minutes - used minutes) x capacity cost rate. Darker cells burn more cash.

Decision menu for idle capacity: take on more group entities or services, shift projects and training into the trough, resize a team, or hold it as growth headroom, but price the catalog as if it did not exist.

What a consultant would circle in red.

Nine findings, ranked by annual EUR impact, each computed live from this model. Expand any card for the calculation trace, the lever and the first action.

Identified in this model

EUR 0

Sum of the nine sized findings below. Not all are additive and not all are fully capturable; even the conservative half typically funds the costing programme many times over.

Ask the model. It computes, it does not improvise.

Type a question about chargeback, methods, capacity or what-ifs. Answers are calculated live from the demo dataset on this page.

Try these

How this works: this assistant computes answers from the demo model embedded in this page. It is deterministic, not a live AI. In an engagement, the same questions run against your data.

Overview

One shared services centre. Three answers about who really consumes it.

Chargeback cascade: billings to net position (TDABC view)

The method-divergence problem

Monthly trend: chargeback, TDABC net position and Finance Operations utilisation

Flat FTE chargeback vs ABC vs TDABC: who is lying to you?

How each engine thinks

Two real time equations from this model

Cost per transaction and net position by engine

Flat-chargeback margin vs TDABC margin

Under the hood: the full model card

The whale curve: a few business units carry everyone else.

Cumulative net position curve (TDABC)

Detail: select a point on the curve

Fixing the tail: the three levers

1. Fix at source

2. Behaviour and policy

3. Consumption-based chargeback

Unused capacity: the cost line almost nobody measures.

CAM-I capacity bridge by resource group

Utilisation of practical capacity by month

EUR of idle capacity, by resource group and month

What a consultant would circle in red.

Ask the model. It computes, it does not improvise.

Try these

See this with your data in 3 to 6 weeks