The Findings|$95M Industrial Manufacturer · 16-Person Finance Org Audit Complete
Chapter 02

The Findings

This is how the finance function runs day to day, stripped of the org chart. Everything the diagnostic surfaced, grouped into a handful of views. Leadership can size the operating picture in minutes, not weeks.

How 16 People Actually Spend Their Time
Manual coordination35%
Data gathering & assembly25%
Review & validation20%
Exception handling (judgment)12%
Strategic analysis & planning8%
Roughly 60% of the team's hours go to chasing, assembling, and cross-checking: routine work agents are built to absorb. Only 12% calls for human judgment. 8% is the forward-looking analysis these hires were brought in to deliver.
Process Maps: SOP vs. Reality
Every documented workflow balloons once you follow it through the systems it touches.
Accounts Payable SOP: 4 steps→Reality: 13 steps6 handoffs
The SOP
  • Invoice received
  • PO matched
  • Approved
  • Paid
4 steps, linear
The Reality
13
steps
6
handoffs
4.2 days
cycle time
200–260 exceptions/mo
volume
Four SOP steps unfold into a 13-step path with six handoffs, approvals bounced over email, roughly 4.2 days to clear a single invoice, and 200 to 260 exceptions surfacing each month.
The Gap
  • 12 recurring exception categories drive about 70% of the workload, and each one resolves along a predictable set of rules.
  • Invoices come in over four different channels, none of which feed a single intake queue.
  • When PO prices don't line up, resolution moves by email and typically stalls about two days.
  • Absent goods receipts force someone to chase the warehouse by hand, costing roughly a day each time.
  • The team keeps letting 2/10 early-payment discounts lapse because invoices aren't cleared quickly enough to qualify.
Accounts Receivable SOP: 3 steps→Reality: 9 steps4 handoffs
The SOP
  • Invoice sent
  • Payment received
  • Applied
3 steps, linear
The Reality
9
steps
4
handoffs
11 days above benchmark
cycle time
38% manual match rate
volume
The three-step ideal runs nine steps: aging reports built by hand, phone-and-email collection runs sequenced on nothing more than balance age, and cash applied by manually pairing payments to invoices.
The Gap
  • Collection effort is ranked purely by how overdue a balance is, disregarding the payment habits the historical data already reveals.
  • Automated cash matching lands only 62% of the time, leaving the other 38% to be reconciled by hand.
  • Some accounts reliably run late on invoices over $50K yet pay everything else on time, but they get lumped in with habitual non-payers.
  • Days sales outstanding sits 11 days worse than the industry norm.
Close Management SOP: 5 steps→Reality: 14 steps8 handoffs
The SOP
  • Close checklist
  • Journal entries
  • Reconciliation
  • Consolidation
  • Reporting
5 steps, linear
The Reality
14
steps
8
handoffs
18–22 days
cycle time
4 entities
volume
All 4 entities keep their own close checklist in a standalone Excel file, and the Controller has to walk through each one by hand to gauge status. The two acquired entities routinely land past deadline.
The Gap
  • There is no single dashboard for close progress, so the Controller opens all 4 spreadsheets one by one.
  • The 18–22 day close is driven mostly by the backlog of AP exceptions upstream rather than the close work itself.
  • Intercompany reconciliation leans entirely on one employee's private cheat sheet mapping vendor names between systems.
  • A 4% intercompany error rate goes unnoticed until the external auditors flag it.
Expense Management SOP: 3 steps→Reality: 8 steps3 handoffs
The SOP
  • Employee submits
  • Manager approves
  • Finance processes
3 steps, linear
The Reality
8
steps
3
handoffs
2 days of close
cycle time
150–220 recategorizations/mo
volume
Ramp runs corporate spend, but its categories never line up with NetSuite's GL codes, so 150 to 220 transactions get re-coded by hand every close. The two acquired entities are also still booking travel through Concur.
The Gap
  • Two whole days of the close disappear into hand-mapping Ramp entries onto NetSuite GL codes.
  • Policy breaches only get flagged once the money has already been reimbursed, not beforehand.
  • Running Concur and Ramp side by side duplicates the exact same workflow twice.
Treasury SOP: 3 steps→Reality: 7 steps3 handoffs
The SOP
  • Check balances
  • Report cash position
  • Forecast
3 steps, linear
The Reality
7
steps
3
handoffs
2 hrs/morning
cycle time
4 accounts, 2 currencies
volume
Each morning the treasury analyst signs into three separate bank portals, gathers balances across 4 accounts in 2 currencies, stitches them together in Excel, checks them against NetSuite, and formats the result for the CFO. Two hours of work, already outdated by the time it lands.
The Gap
  • By the time the cash position reaches the CFO at 10:30am, overnight settlements have already moved the numbers.
  • The 13-week rolling forecast gets built again from scratch in Excel every single week.
  • FX exposure lives in its own spreadsheet that only gets refreshed whenever someone happens to think of it.
FP&A SOP: 3 steps→Reality: 11 steps5 handoffs
The SOP
  • Pull data
  • Analyze variances
  • Deliver package
3 steps, linear
The Reality
11
steps
5
handoffs
8–10 day lag
cycle time
3 data sources
volume
Nothing can begin until the close wraps around day 18–22. Figures are exported separately from NetSuite, Ramp, and Power BI, reconciled by hand, and the variance deck is built in PowerPoint. It is not delivered until day 28–30.
The Gap
  • There's an 8–10 day gap between when the close finishes and when the variance package goes out.
  • Every report requires manually reconciling three distinct data sources.
  • Leadership only sees the analysis once the month it covers has already ended.
Top 12 AP Exception Types
Type Freq/mo Resolution Repeatable Owner
PO Price Mismatch 28 2.1 days Yes AP Analyst
Missing Goods Receipt 24 1.4 days Yes AP Team Lead
Duplicate Invoice 23 0.5 days Yes AP Analyst
Quantity Variance 19 1.8 days Yes AP Analyst
Missing PO Reference 16 3.2 days Yes AP Specialist
Tax Calculation Error 14 1.1 days Yes AP Analyst
Vendor Master Mismatch 13 2.4 days Yes Controller
Currency Conversion 12 0.8 days Yes Treasury
Freight Charge Dispute 9 4.5 days No AP Team Lead
Payment Terms Mismatch 8 1.6 days Yes AP Analyst
GL Code Misclassification 7 0.9 days Yes AP Analyst
Approval Threshold Exceeded 6 2.8 days No Controller
About 70% of these exceptions resolve along a fixed decision path, so agents can clear them from day one. The remaining 30% still need a human call. Agents gather the full context up front so those calls take minutes, not hours.
Risk Register
criticalIntercompany reconciliation
The entire reconciliation hinges on one analyst's private mental map of vendor names spanning three systems. If she takes a day off, the process stops: no trained backup, nothing written down.
critical4% intercompany error rate
A 4% error rate slips through unnoticed until the external auditors catch it. There is no internal check or automated match in place, which leaves the door open to a restatement.
criticalMissed early payment discounts
Because invoices move too slowly, the team forfeits 2/10 net terms it qualifies for. We estimate that pace costs roughly $240K/yr in discounts left on the table.
highNo master vendor governance
One supplier shows up under four spellings depending on the system you check. Nothing dedupes the records, there is no golden master, and no shared ID ties them together across platforms.
highNo close status visibility
Nobody can see where the close stands across all 4 entities until the controller opens each spreadsheet by hand. The two acquired entities are late almost every period.
highDual expense systems
Ramp and Concur both run, split across different entities, and neither talks to NetSuite. The result is two parallel workflows and a standing reconciliation burden.
mediumStale daily cash position
Someone rebuilds the cash position by hand each morning and hands it over around 10:30am. By then overnight settlements have already moved the number. The CFO is steering on figures that are stale on arrival.
medium30+ day data lag
By the time FP&A ships the variance package, the underlying numbers are already more than 30 days old. Leadership is deciding against a picture of last month.