Sample reportThe company, people, systems and figures below are fictional. This shows what a real audit report contains.
Manufacturing Data & Systems Audit
Sample Engineering Co., a fictional 45-person precision machining business.
Summary
The business can see quoted against actual margin per job from its existing systems, without replacing its ERP. Jobs are numbered consistently everywhere, and the ERP's reporting views can be read safely. Two gaps limit accuracy, and both can be closed during the first build: labour is not always booked to the operation, and material is costed at standard rather than actual price.
- Feasible: a Job Profitability build on the ERP and the timesheet export.
- Not yet feasible: downtime reporting, because downtime isn't recorded in any system.
- Recommended: a fixed-scope first build, set out under “Recommended first build”.
Scope and method
The question set by the owner: which jobs make money, and why do some run over quote? Over five working days we:
- interviewed the owner, operations manager, office manager and workshop lead (45 minutes each);
- mapped every system that holds job, labour, material or delivery data;
- read a 12-month extract from each source, using read-only access;
- tested whether figures in each source agree with each other and with invoices;
- drafted KPI definitions and reviewed them with the operations manager.
Systems and access
| Source | What it holds | Proposed access | Data quality |
|---|---|---|---|
| ERP (SQL Server back end) | Jobs, quotes, purchase orders, stock, invoices | Read-only database account on reporting views | Good |
| Timesheet export (CSV) | Labour hours by job and operation | Nightly file to a shared folder | Fair |
| Production spreadsheet | Machine schedule, notes, downtime | Not proposed for the first build | Poor |
| Supplier price lists (PDF/Excel) | Material prices | Out of scope | n/a |
Nothing is written back to any system. Credentials stay with Sample Engineering Co. and can be revoked at any time.
Data quality findings
Labour is booked to the job but not to the operation on about a fifth of closed jobs.
Impact: Quoted vs actual labour can be compared per job, but not per operation.
Recommendation: Book to operation from a set date. Report per job until then.
Material is costed at standard cost, and actual purchase prices are not posted back to the job.
Impact: Margin on material-heavy jobs is overstated when prices rise.
Recommendation: Use PO line prices for actual material cost. Mapping covered in the build.
Downtime is recorded on a whiteboard and in spreadsheet notes only.
Impact: Downtime cannot be reported reliably from any system.
Recommendation: Excluded from the first build. A simple floor log is a candidate for phase 2.
Customer names are entered differently in the ERP and the spreadsheet.
Impact: Customer-level reporting double-counts some customers.
Recommendation: One mapping table, owned by the office manager.
Job numbers are consistent across all sources.
Impact: Jobs can be joined across systems without manual matching.
Recommendation: None. This is what makes the first build feasible.
KPI definitions
Agreed in draft with the operations manager. Every figure in the build is calculated exactly this way, and reconciled before sign-off.
| KPI | Definition | Sources |
|---|---|---|
| Quoted vs actual margin | (Invoiced value − actual material − actual labour − subcontract) ÷ invoiced value, per closed job | ERP, timesheets |
| Labour variance | Actual labour hours − quoted labour hours, per job | ERP quote lines, timesheets |
| Material variance | Actual material cost (PO line prices) − quoted material cost | ERP quote and PO lines |
| On-time in full | Order lines delivered complete on or before the promised date ÷ all order lines due | ERP sales orders and dispatch |
Risks and constraints
- ERP licence terms. Reading the reporting views is permitted under the current support agreement. This needs confirming in writing with the ERP reseller before the build starts.
- No test environment. All reads run against production reporting views, outside working hours.
- Key-person dependency. The production spreadsheet is maintained by one person. It is out of scope for the first build for this reason.
Recommended first build
Job Profitability Intelligence, on the ERP (read-only) and the nightly timesheet export, hosted on the server Sample Engineering Co. already runs.
In scope
- Quoted vs actual margin, labour and material variance per closed job, with drill-down to quote and PO lines
- Jobs ranked by variance, filterable by customer and month
- Nightly refresh, with a reconciliation report against invoiced totals
- Role-based access for owner, operations and office
- Documented deployment, data mappings and handover session
Out of scope
- Downtime and machine utilisation (no source data yet)
- Any write-back to the ERP
- Changes to how the workshop books time, beyond a written recommendation
Acceptance criteria
- Margin for 20 closed jobs, chosen by the operations manager, matches a manual calculation from the ERP to within rounding
- Nightly refresh runs for 10 consecutive working days without manual intervention
- The owner and operations manager can answer “which jobs ran over quote last month, and why” without exporting anything
Timeline and price: typically 2–4 weeks once credentials and approvals are in place. The fixed price is set out in the proposal that accompanies a real report.
Roadmap
| Phase | What | Depends on |
|---|---|---|
| 1 | Job Profitability (this proposal) | Read-only ERP access, timesheet export |
| 2 | Simple floor log for downtime, then Operations Intelligence | Two months of logged downtime |
| 3 | Customer order-status portal | Phase 1 data model in use |
Each phase is scoped and priced on its own. Nothing after phase 1 is committed by accepting this report.
What this audit is not
This is a diagnostic of data, systems and feasibility. It is not a cybersecurity assessment, a penetration test, a financial audit or an implementation.
A real audit is fixed-fee, takes 3–7 working days once access is in place, and ends with a report like this one and a fixed-scope proposal.