Not right, not wrong — different.

A finance team and an accounting team look at the same company and see two different companies.

The finance team is looking at projections. They see a forward curve — revenue by segment, by region, by product line, sometimes by individual customer. They are looking at what will be, sliced by the variables they need to make decisions about.

The accounting team is looking at the ledger. They see what was, posted to the chart of accounts, balanced, periodic, defensible to an auditor. The ledger has different variables — entry date, account code, debit, credit — and a tighter set of them.

The operations team is looking at contracts. They see commitments, terms, dates, parties. The variables there are different again — counterparty, effective date, governing law, payment schedule.

The legal team is looking at obligations. The same contracts, sliced differently — what’s enforceable, what’s expiring, what’s renewing automatically, what’s been waived.

None of these views is wrong.

Each one is correct for the job that view does.

---

The trouble starts when one view needs to consume another.

The finance team wants to project revenue. To do that they need to know what’s been booked. So they ask accounting for an extract — and accounting sends them the ledger, which is correct, but the ledger doesn’t know about region or product line or customer segment in the shape finance needs. Finance back-fills those dimensions from a customer-master extract in a different system. Sometimes it lines up. Sometimes it doesn’t, and somebody spends a Tuesday afternoon reconciling a number that should have been the same number in both places.

Operations wants to know what revenue is at risk. To do that they need to cross-reference contracts (theirs) against the ledger (accounting’s). They pull from both, build a workbook, and report up. Three weeks later a contract renews on different terms and the workbook is wrong. Nobody notices for a month.

Legal wants to know which contracts have hit certain thresholds. To do that they need usage data (operations) joined to contract terms (legal’s own) joined to billings (accounting’s). They send an email to operations. Operations sends a spreadsheet. Legal interprets it. Three lenses, three extracts, one analysis. The next time the question is asked, the extracts are rebuilt from scratch.

The practical question is never abstract. It sounds like: what cash is expected from contracts renewing in Q3, in regions where delivery is late, for customers whose invoices are already in dispute? Today that question crosses four systems before anyone can trust the answer.

The accountant isn’t wrong for needing the ledger view. The financial analyst isn’t wrong for needing the projection view. What’s wrong is asking each of them to rebuild the underlying record every time they need their view of it.

This is the failure pattern. It’s not that any single tool is broken. The accounting system is doing its job. The CRM is doing its job. The contract repository is doing its job. The FP&A tool is doing its job. It’s that the underlying facts of the company are scattered across all of them, and any view that needs more than one tool’s facts is, in practice, a fresh reconstruction.

The reconstructions are heroic. They’re also expensive, slow, and stale by the time they land.

---

There’s a clean way to think about this. Every view a function holds — finance’s projection, accounting’s ledger, ops’ contracts, legal’s obligations — is a projection of the same underlying record. The underlying record is what actually happened: a deal got signed, a service got delivered, a payment moved, a credit got applied. Each function reads that record through its own lens, organized for its own job.

A ledger is a projection of the underlying record onto the accounting axes — date, account, debit, credit.

A revenue projection is a projection onto the planning axes — segment, region, product, period, scenario.

A contract view is a projection onto the relational axes — counterparty, term, obligation, governing law.

An obligations view is a projection onto the temporal axes — what’s owed, what’s earned, what’s expiring, what’s been waived.

Same record. Different projections. None of them is the real record. The real record is what they all share, refracted into the shape that view needs.

When projections share an underlying record, two things become true that aren’t true today. One: any view that consumes another view’s facts can read them directly, not via an extract. Two: when the underlying record changes — a deal gets re-negotiated, a contract gets amended, a credit gets applied — every projection updates with it. There is no Tuesday-afternoon reconciliation, because there is nothing to reconcile. The number is the same number because it’s the same number.

---

There is an old joke in accounting that goes: “ask three accountants what the answer is, get four numbers.” The joke is funny because it’s true, and it’s true because the underlying record is not held in common. Each accountant reconstructs their answer from their own extract of their own slice. Each answer is internally correct. None of them quite agree.

The joke survives because the record that would let them agree didn’t exist. The pre-spreadsheet era had paper ledgers — physical, single-copy, hard to project from. The spreadsheet era had files that traveled by email and forked instantly. The SaaS era promised consolidation and delivered a different problem: forty tools, each holding its slice of the truth, each refusing to acknowledge that any of the other tools own anything.

What changes the joke isn’t a better extract. It’s an underlying record that every view reads from, that every change writes to, and that no view has the authority to call the version of.

The finance team still gets their projection.

The accounting team still gets their ledger.

Operations still gets their contract view.

Legal still gets their obligations view.

Each lens is built and tuned by the function that uses it, in the language that function speaks. None of them changes shape.

What changes is the record underneath. The lenses are reading the same record now. And when one of them is asked a question that crosses another’s territory — what’s the projected revenue from contracts up for renewal in Q3 that are currently in dispute? — the answer doesn’t require three extracts and a workbook. The lenses already speak to the same source.

This is the second category of thing the operating record changes. The first was the daily close: a continuously-true position the ledger eventually confirms. The second is this: a single underlying record that every function in the company reads through their own correct lens.

Not right, not wrong, different — and finally, sharing.