The Ops Accountant

Analyze · Data analysis for operations accounting

Reviewing the flow with data

The AP, JIB and Revenue sections explain how money moves. This section shows how to check it at scale: the queries, models, dashboards and templates a JIB supervisor uses to find the handful of wrong rows among thousands of right ones. Every case study runs on the same made-up dataset, and most of them run live in your browser.

Start here

The sample dataset

Wells, ownership decks, AFEs, invoices, JIB lines, non-op JIBs, production, revenue and accruals, all keyed on well name + cost center, with errors planted for the case studies to find. →

14wells
9tables
10months

Case studies

1 · SQL

SQL exception queries

A live SQL workbench with seven saved review queries you can read, change and run.

  • Billed, not paid, and revenue on a lookalike cost center
  • NRI that doesn't tie to WI, and AFE spend vs. authorized
  • Invoice coding exceptions, LOE per BOE, accrual accuracy
Open the workbench →
2 · Power BI

Operations accounting dashboard

LOE per BOE by well, AFE spend vs. authorized, and accruals vs. actuals, with a field filter.

  • A star schema keyed on well name + cost center
  • The DAX measures behind every number
  • Hover detail and a table view for each chart
Open the dashboard →
3 · Power Query

Combining non-op JIBs

Two operators, two file layouts, one clean table mapped to your wells and cost centers.

  • One M function per operator layout
  • A cross-reference table, not a name match
  • Unmapped wells and decimal checks kept, not dropped
Read the case study →
4 · Excel

Accrual review template

Last month's accruals against what actually came in, flagged and rolled up.

  • Structured tables with XLOOKUP, SUMIFS and IFS
  • A LET and FILTER follow-up list
  • A pivot of variance by month and group
Read the case study →
5 · AI

AI in the review process

Where assistants like Copilot and Claude help, the guardrails that keep the numbers right, and example prompts.

  • AI drafts, a person decides
  • A worked example where the first draft was wrong
Read the case study →

How the work is done

One key everywhere

Well name + cost center, in every join, lookup and relationship. Names repeat across wells; the key doesn't.

Find the exceptions

Each tool returns a short list of things to do, sorted by dollars or severity, not a wall of rows to scan.

Made-up data only

No employer, partner or owner data appears anywhere. Every well, vendor and figure is invented.

The material on this site is general educational information, not legal, accounting or tax advice. All examples use made-up data.