12
recurring client accounts
Same report structure, different client data.
Demonstration example
See the workflow behind the polished report.
A realistic, invented example showing how two recurring exports become one reconciled, client-ready monthly report.
Scenario
A boutique marketing consultancy rebuilds the same report for 12 clients every month.
The report is useful, but the production method depends on one person remembering hidden cleanup and formatting steps.
2
input exports
Campaign results and sales outcomes.
6h
illustrative monthly effort
Copy, clean, reconcile, format and export.
1
key-person dependency
The operating knowledge lives in memory.
Demonstration assumption: the current report already delivers value. The project improves production reliability; it does not invent the consultancy’s analysis or replace its client relationship.
01 · Before
A useful report produced through a fragile monthly ritual.
The risk is not only the six hours. It is that the logic, checks and exceptions are invisible until something goes wrong.
Eight recurring production steps
- Download campaign export.
- Download sales-outcome export.
- Rename channels and client identifiers.
- Remove duplicate or test rows.
- Paste figures into a master workbook.
- Repair formulas and chart ranges.
- Copy commentary into a slide template.
- Export, recheck totals and send PDF.
What the owner cannot safely delegate
- Labels differ between sources and months.
- Several cleanup rules are not documented.
- Chart ranges have to be repaired manually.
- Exceptions are noticed only through memory.
- Source totals and report totals are not recorded together.
- The handover instruction is “copy last month”.
02 · Inputs
Two stable exports, but inconsistent field names and definitions.
The diagnostic makes every source, owner, reporting period and transformation explicit before implementation.
Illustrative source evidence
Invented dataInput 01 · Campaign export
Results by source
| src_name | leads | spend_nzd |
|---|---|---|
| Paid Search | 48 | 3,240 |
| Referral | 31 | 0 |
| Direct / none | 22 | 0 |
| Partner-Net | 18 | 840 |
Reporting month appears in the filename, not in a field.
Input 02 · Outcome export
Sales outcomes by channel
| channel | qualified | won |
|---|---|---|
| Search | 23 | 8 |
| Referrals | 24 | 12 |
| Direct | 14 | 5 |
| Partners | 10 | 4 |
Channel names do not join until they are normalised.
03 · System
A controlled source-to-output architecture.
The method can be implemented with formulas, queries, scripts or a simple dashboard. The important part is that each stage has a defined purpose and review point.
Receive
Store the two approved exports in a controlled cycle folder with reporting dates.
Normalise
Map channel labels, client identifiers, dates, blanks and test records through defined rules.
Model
Join the sources, calculate approved metrics and preserve unmatched exceptions for review.
Reconcile
Compare source totals, row counts, periods and exception counts before release.
Publish
Refresh the client-branded output, add human commentary, approve and export.
04 · Defined logic
The hidden decisions become visible rules.
This is the difference between a clever spreadsheet and an operating workflow that can be checked and handed over.
Channel map
Paid Search and Search resolve to Search; unmatched values are not silently guessed.
Reporting period
Both exports must cover the same approved month. A mismatch blocks release and appears as an issue.
Duplicate treatment
Test records and exact duplicate identifiers are excluded according to an approved rule and counted in the QA note.
Metric ownership
Calculated facts are separated from human commentary so narrative never overwrites the underlying result.
05 · QA gate
A report is ready only when the output can be traced back to its source.
Checks are recorded with the reporting period, source refresh time, reviewer and exception status.
Source totals
Lead, qualified and won totals reconcile to the approved exports.
✓ ReconciledReporting period
Both files use the same monthly date range and source refresh timestamp.
✓ Same periodUnmapped label
One new source label is held outside the report until the owner confirms its category.
! Decision requiredOutput package
Charts, titles, units, source date, page count and PDF export are checked.
✓ Ready for approval06 · After
A repeatable operating cycle, not a black box.
The report still requires professional review and commentary. The routine work no longer begins from an undocumented blank page.
Illustrative monthly performance report
Demonstration onlyClient-facing summary
June performance summary
119Leads
71Qualified
29Won
40.8%Illustrative close rate
✓ Reconciled to source · 30 June 2026
Operating cycle
Six explicit actions
- Place approved exports in the cycle folder.
- Refresh the normalised model.
- Resolve any visible exceptions.
- Run and record the QA checklist.
- Add professional commentary and approve.
- Export, deliver and record the release version.
07 · Handover
The client receives the system, not only the output.
Every item is designed to reduce key-person dependence and make future changes visible.
Workflow map
Inputs, transformations, decision points, exceptions, reviewers and output path.
Data dictionary
Field names, source, definition, transformation, owner and sensitivity.
QA checklist
Source totals, row counts, period, filters, formulas, layout and approval.
Operating guide
Cycle preparation, refresh, exception handling, commentary, export and recovery.
Change log
Date, requested change, reason, risk, test, reviewer, approval and release version.
Support boundary
Fourteen days of bounded handover questions, plus optional ongoing care.
Your report will be different
Start with the workflow you already run.
A high-level description is enough to determine whether one report and up to two inputs fit the launch scope.
Book a 15-minute report fit check
No report upload required