Selected work / VM Bill Management

Transformation 01 / Allocation governance

Who owns the bill when the infrastructure is shared?

I translated inconsistent ownership rules, shared infrastructure, and messy master data into a governed attribution model designed for enterprise rollout.

designed for60+organizational units in intended rollout scope

The supplied case-study source identifies 60+ units as intended enterprise scope. Deployment and adoption are not claimed.

scaleThousandsof billing records in a typical month

Recurring operational volume recorded in the supplied case-study source; the public site does not expose exact client totals.

scale4connected attribution layers

Project, department, application, and virtual-machine relationships form the explainable chain.

estimated50–70%estimated reduction in manual reconciliation effort

Retrospective operational estimate from the supplied source. It is not a measured before-and-after result.

The actual project

What the business needed, and what I delivered.

The business
A multi-unit enterprise had to distribute recurring shared-cloud infrastructure costs. Finance and operations relied on inconsistent ownership files, embedded percentages, and knowledge held by individuals.
The product mandate
Create one bill-management product that could ingest monthly records, govern project, department, application, and VM relationships, validate unsupported structures, and explain every allocation in reporting.
My work
I mapped the existing reconciliation process, separated ownership from allocation, defined the four-layer data model and validation rules, and converted them into workflows, user stories, acceptance criteria, backlog coverage, and UAT scenarios.
Business effect
The model was designed for 60+ organizational units and thousands of monthly records. The supplied source estimates 50–70% less manual reconciliation effort; the reduction is not presented as measured.

Act 1 / The pressure

A final number is useful only when its path can be explained.

Change one condition and watch the consequence travel through the system. The model uses representative data and preserves the real decision.

Representative simulation · no client data

One bill. Two claims. Choose what the system should believe.

Record
SYN-1048
Service
Shared reporting cluster
Amount
$12,400
Choose an allocation approach

Incomplete allocation

The total exists, but its explanation does not.

Use one of the three approaches to see how a local assignment changes the reporting chain.

Before: a shared bill with one free-form ownership field can produce duplication or leave the department relationship unexplained.

Governed: ownership remains explicit while a validated 65/35 rule carries the cost through project, department, application, and VM.

Act 2 / The model

The complexity lived between four layers.

A virtual machine could support one application while that application served several departments. Shared services could benefit multiple programs. Historical files represented those relationships through combined fields, embedded percentages, inconsistent labels, and knowledge held by individuals.

A mistake at any layer could travel into reconciliation, ERP-aligned reporting, and the final interpretation of cost ownership. The product problem was therefore governance: accept legitimate complexity while refusing structures that could not be interpreted reliably.

  1. 01Project
  2. 02Department
  3. 03Application
  4. 04Virtual machine

Underlying structure: I translated inconsistent ownership rules, shared infrastructure, and messy master data into a governed attribution model designed for enterprise rollout.

Act 3 / The decisions

The product changed when the decisions became explicit.

01

Separate ownership from allocation

The pressure. Historical files placed multiple departments or percentages inside a single ownership field.

The move. Keep ownership explicit and handle shared cost through deliberate allocation behavior.

Inspect the tradeoff

Rejected: Preserve free-text multi-owner fields because they were familiar to spreadsheet authors.

Consequence: A shared service became a governed scenario that could be validated and reported consistently.

02

Make invalid structures visible early

The pressure. Unsupported combinations, incomplete mappings, and duplicate allocations were otherwise discovered at reporting time.

The move. Define validation expectations at data entry and master-data ingestion.

Inspect the tradeoff

Rejected: Accept every historical structure and resolve ambiguity during month-end reconciliation.

Consequence: The system could stop uninterpretable data before it contaminated downstream reports.

03

Preserve the explanation behind every number

The pressure. A total without its relationship path could not answer why a cost was assigned to a unit.

The move. Make reports traceable through project, department, application, and VM relationships.

Inspect the tradeoff

Consequence: Dashboard behavior, stories, and acceptance criteria all carried the same interpretability requirement.

04

Treat rollout readiness as a product requirement

The pressure. A configuration built around one team’s habits would not survive enterprise-wide use.

The move. Surface access, master-data, validation, reporting, and edge-case requirements before UAT.

Inspect the tradeoff

Consequence: Cross-functional validation focused on operational readiness instead of only screen completion.

Contribution boundary

My work turned interpretation into buildable behavior.

What I contributed

  • Mapped fragmented attribution workflows and their reporting dependencies.
  • Turned allocation and validation rules into implementation-ready requirements.
  • Defined expected visibility across configuration, billing, and reporting stages.
  • Identified unsupported master-data structures and operational edge cases.
  • Prepared user stories, acceptance criteria, and backlog coverage.
  • Clarified implementation behavior with engineering and QA and supported UAT readiness.

What remained outside my ownership

  • I did not own infrastructure policy, financial accounting decisions, engineering implementation, or enterprise adoption.
  • The estimated reconciliation improvement belongs to the governed operating model described in the supplied source - not to an independently measured personal result.

Act 4 / The change

The report gained an explanation, not just a total.

Evidence-qualified outcome50–70% estimated reduction in manual reconciliation effort

The work produced a centralized attribution model designed to support more than 60 organizational units and thousands of monthly billing records. It created a structured path from infrastructure ownership to reporting interpretation, with explicit validation and traceability across all four layers.

The supplied source estimates that the governed model could reduce manual reconciliation effort by 50–70%. Until a measured baseline exists, the public claim remains explicitly estimated.

Read the evidence boundary

The interaction and diagrams use representative synthetic data. No client records, internal identifiers, or live financial values are shown.

Act 5 / The lesson

Enterprise products often fail at the boundaries between systems, not inside individual screens. The central design challenge was deciding where flexibility should end and governance should begin.
What I would explore next

I would add anomaly signals for unexpected attribution changes, historical ownership comparison, and a safe simulation that shows the reporting effect of an allocation before submission.

Start a conversation

What are you trying to untangle?

Bring the unclear process. We can make the next decision obvious.