Off-Invoice Control Assessment · methodology

How we score

The exact method behind the Off-Invoice Control Index — laid open.

Read this first — what this is and isn’t. This is a directional self-assessment. It scores your answers to eight questions; it does not see your data, your systems, or your books. The 0–100 score is computed from Aurgus’s control-maturity rubric — our judgment, from 20+ years of rebate and off-invoice work — not from external benchmarks. The dollar figure is a category estimate modeled from industry ranges, not a measurement of your financials. The only way to know your real numbers is to trace your own data — what the 30-minute working session does. Everything below is the whole method.

The score: five dimensions, equal weight

Your composite is the simple average of five dimension scores (0–100 each), then a small deduction for realized losses.

composite = round( average(5 dimensions) − realized-loss penalty ), bounded 0–100

Why equal weight? We have no basis to claim one control matters more than another, so we don’t pretend to. As real assessments accumulate, we’ll revisit weighting if the evidence supports it.

Each answer maps to a fixed point value. Here is every one.

1 · Calculation Integrity — where the accrual is computed

Your answer	Points
A dedicated calculation platform	100
A native ERP rebate module	70
An ERP add-on requiring specialist config	52
A mix of ERP and spreadsheets	28
Primarily spreadsheets	8

2 · Estimation Discipline — how you accrue before a tier is reached

Your answer	Points
Expected final tier, documented method	100
Rate earned so far, true-up later	45
Varies / not standardized	30
Not sure	22
No accrual until the threshold trips (cliff)	15

3 · Audit Defensibility — time to trace one accrual to source

Your answer	Points
Minutes — drill to source	100
Hours of spreadsheet assembly	48
Days	24
Reconstruct from memory	6

4 · Number Coherence — do Sales, Finance, and the ERP agree

Your answer	Points
One governed number, all read it	100
Reconciled to the agreement at close	55
Three numbers, reconciled by hand	20
Not systematically compared	10

5 · Close Efficiency — how much of close rebate reconciliation consumes

Your answer	Points
Negligible — automated	100
A day or two	62
Most of a week	28
The bottleneck of our close	8

Realized-loss deduction — a recent finding sharpens the read

In the last two years	Deduction
None	0
A minor adjustment	−3
A significant true-up	−8
A restatement / material weakness	−15

What your tier means

Score	Tier	What it says
85–100	Defensible	Numbers are computed once and trace to source. The target state.
65–84	Governed	A systematic operation with real controls, but gaps still force reconstruction when someone asks the second question.
40–64	Managed	Programs are handled, but the number is defended by reconciliation, not lineage — where close time and audit exposure concentrate.
0–39	Reactive	The accrual is assembled on demand. A common source of audit findings and month-end fire drills — and the largest opportunity.

A note on the two tools. The Defensibility Check is a separate, lighter tool. It returns a three-tier verdict — Exposed → Reconstructable → Defensible — on the single dimension of traceability, not the five-dimension composite above. Both tools share “Defensible” as the top state; the Control Index tiers here are the broader five-dimension maturity picture.

The exposure estimate — a category benchmark, not your books

We model a range, not a single number, from the revenue band you selected — using the band’s floor and ceiling — so it never overstates a company inside a band.

low = band floor × off-invoice % × 1%
high = band ceiling × off-invoice % × 2%

Off-invoice % starts at 15% and rises toward 25% with the number of program types you run. Scaling by program count is an Aurgus assumption, not a published figure; the 15–25% band itself is from industry sources.
1–2% is an industry estimate of the inaccurate-or-invalid share of that spend (Infosys BPM).
Optional refinement: enter your actual revenue (no email) and the range collapses to a point estimate on your figure.

Worked examples (3 program types, ~19%)

Company	Modeled as	Estimate
$200M (Under-$500M band)	$100M–$500M band	$0.3M – $1.9M / yr
$2B ($1B–$5B band)	$1B–$5B band	$1.9M – $19M / yr
$2B, refined with exact revenue	$2B	$3.8M – $7.6M / yr

The band is wide because a band is wide — the exact-revenue refinement is what makes it precise. Mature controls compress the realized fraction; weak controls sit near the top of the range.

Sources

Trade / off-invoice spend ~15–25% of revenue, among the largest P&L lines after COGS — industry sources: Infosys BPM; NielsenIQ.
1–2% of trade spend inaccurate or invalid — Infosys BPM.
Recovery audits recover ~$1M per $1B of spend — PRGX.
Manufacturing close: ~6.4-day median, top quartile ~4.8, bottom quartile 10+ — APQC Open Standards Benchmarking (Cycle Time to Monthly Close).
Revenue recognition is consistently a top cause of financial restatements — Audit Analytics, 20-year review; variable-consideration is its hardest ASC 606 judgment.

The honest limit

This scores intent and process, from your answers. It cannot see whether your spreadsheet is actually right or your dedicated platform is actually trusted. It’s a mirror, not an audit. The working session is where we trace one real program from agreement to source and replace the estimate with your number.

Take the assessment → Book the working session →