TL;DR: R2R, P2P and O2C are the three end-to-end process towers a finance shared services centre is built around. Record to Report produces the numbers. Procure to Pay spends the money. Order to Cash collects it. Most transformation programmes attack them with technology and get modest returns, because the binding constraint is usually ownership, not automation: finance organisations with formal, end-to-end order-to-cash process ownership run at 55% lower process cost than their peers. Fix who owns the process before you buy anything to run it.
What Are R2R, P2P and O2C?
They are the three transactional process towers of a finance function, named for their start and end points rather than the departments that perform them. That naming is the whole idea. Each cuts across functions deliberately, because the failures happen at the handoffs.
-
Record to Report (R2R) runs from transaction capture through to the financial statements and management reporting.
-
Procure to Pay (P2P) runs from identifying a need through to paying the supplier.
-
Order to Cash (O2C) runs from a customer order through to collected cash.
A note on nomenclature. The Hackett Group’s benchmark portfolio names the process account-to-report (R2R), while most of the market says record to report. Same taxonomy, different label. Do not let a vendor tell you their process model is proprietary because they renamed the boxes. The authoritative public taxonomy is APQC’s Process Classification Framework, which underpins its Open Standards Benchmarking database, launched in 2004.
Record to Report (R2R) Explained
R2R produces trustworthy numbers. It covers general ledger accounting, intercompany accounting and eliminations, fixed assets, reconciliations, accruals and provisions, the period-end close, consolidation, statutory reporting and management reporting.
It is the most underestimated tower, because its output looks like an obligation rather than an asset. Yet every other decision depends on it, which is why a broken close is never only a close problem.
The tell that R2R is failing is not a late close. It is a close that finishes on time and nobody trusts. When business heads maintain shadow spreadsheets alongside the official MIS, R2R has already failed, whatever the calendar says.
Procure to Pay (P2P) Explained
P2P runs from demand identification through sourcing, requisition, purchase order, goods receipt, invoice receipt, three-way match, approval, payment and supplier master maintenance.
APQC places accounts payable at 9.6.1, defining it as processing payments of operating expenses and other supplier charges, with steps including accruals and reversals, payables taxes, researching and resolving payable exceptions, and processing payments.
That third step, exceptions, is where the money goes. A P2P function’s cost is driven almost entirely by the proportion of invoices that fail to match automatically and need a human to investigate. Every published cost-per-invoice benchmark is really an exception-rate benchmark wearing a disguise.
And the exceptions are created upstream, in procurement, by a purchase order raised after the goods arrived or with terms nobody agreed. Which is precisely why the tower is defined as procure to pay, and why putting it under the AP manager guarantees it will not improve.
Order to Cash (O2C) Explained
O2C runs from customer order to cash in the bank. The Hackett Group breaks it into six aspects: credit management, order management, customer billing, cash application, collections management and dispute or deduction management. APQC’s cost measure spans the same ground.
O2C is where the tower model proves itself. Collections gets blamed for DSO. But a customer who has not paid because the invoice was wrong is a billing failure. A customer who cannot pay is a credit failure that happened before the order was accepted. Hackett makes the point directly: improving decisions at the front end delivers benefits across billing, cash application, collections and dispute resolution.
Chase collections harder and you get a more tired collections team.
The Lever Is Ownership, Not Automation
Here is the most useful number in this article, and it has nothing to do with technology. Finance organisations with formal, end-to-end order-to-cash process ownership have a 55% lower process cost than peer groups.
Not 5%. Fifty-five. From appointing someone to own the process end to end.
Read that against how most companies are organised. Credit sits under finance, order management under sales operations, billing under finance, disputes under customer service, collections under a manager three levels down. Six aspects, four reporting lines, nobody accountable for the whole. Every one of those managers can hit target while DSO climbs, because each is optimising a fragment.
This is why so many shared services business cases disappoint. The programme is sold as labour arbitrage and delivered as a lift-and-shift: the same fragmented process, performed by different people in a cheaper city. Cost per transaction falls by the wage differential and then stops falling, because the design was never the problem you fixed.
Hackett’s AI World Class research finds performance advantages of up to 75% for organisations pursuing process-led AI transformation, and is explicit that the gains come not from a single automation but from reimagining how work flows across the entire process. Their conclusion is worth quoting plainly: AI is not creating a technology gap, it is creating a performance gap.
Automation applied to a broken process produces a fast broken process.
What Good Looks Like in 2026
Hackett’s Digital World Class research finds top-tier finance organisations operate at 45% lower cost as a percentage of revenue, deliver 74% faster executive insights and produce forecasts 57% faster than peers, while running 48% lower days sales outstanding and 83% lower average days delinquent.
The frontier is moving further. The AI World Class Finance benchmarks published in July 2026 model order-to-cash process costs declining 52% to 59%, staffing requirements down 56% to 64%, dispute resolution 43% faster and delinquent days down 85%.
The pressure driving this is arithmetic. Hackett’s 2026 Finance Key Issues Study projects finance workloads rising 3.2% while headcount falls 2.1% and budgets fall 1.7%, opening a 5.3% productivity gap. AI implementation has climbed to the fourth-ranked finance priority, up from sixteenth in 2025.
Treat these as directional. They come from a firm that sells transformation consulting and are modelled scenarios, not audited outcomes. They remain the best public reference points available.
The India Layer That Global Process Maps Miss
Most R2R, P2P and O2C guides are written for a US or UK reader and then read by someone building a centre in Bengaluru. Three India-specific facts break the generic process map.
E-invoicing puts a hard clock on O2C. GST e-invoicing applies above ₹5 crore aggregate annual turnover, and since 1 April 2025, businesses with an AATO of ₹10 crore or more must report e-invoices to the Invoice Registration Portal within 30 days of the invoice date. Miss it and the IRP rejects the invoice, no IRN is issued, and the document is invalid for GST. Month-end batch uploading is not a process choice any more. It is non-compliance.
Wage redefinition reprices P2P’s payroll interface. The Code on Wages treats excluded allowances above 50% of total pay as wages, so if your centre runs payroll, the provident fund and gratuity arithmetic changed under it.
R2R now carries a transfer pricing dependency. If the centre serves a foreign parent, its cost base determines the inter company charge, which feeds the safe harbour election under the Income-tax Rules, 2026. R2R stops being an accounting process and becomes an input to a tax position.
The Lead Measure for Each Process
Every tower has a lag measure everyone reports and a lead measure almost nobody sets. Lag measures like DSO, cost per invoice and days to close are outcomes, known too late to act on, and the team reviewing them has no direct leverage over them. A lead measure is something the team controls and can watch while executing.
The question is not what our DSO is. It is whether we have identified measures our teams control that predict the outcome, rather than the outcome itself.
| Tower | The lag measure everyone reports | A lead measure worth setting |
|---|---|---|
| R2R | Days to close | Reconciliations completed and cleared before day one of close |
| P2P | Cost per invoice | First-time three-way match rate, and PO-before-invoice compliance |
| O2C | DSO | Invoices issued error-free and reported to the IRP on the day of issue |
I learned this the hard way on collections. The insight came from analysing the one month we hit target: the plan went out by the second, with customer and invoice-level data given to the team, and the focus was the top five customers by mid-month. So the dashboard changed. We stopped reporting month-end cash and started reporting collection performance against named customers by a named date. The lag measure looked after itself.
Narrowing to one or two lead measures takes time to tune. But it moves responsibility back to the operating team and out of the CEO’s office, which is where it belongs.
Frequently asked questions
- What is the difference between R2R, P2P and O2C?
- Three end-to-end finance process towers named for their start and end points. Record to Report runs from transaction capture to financial statements. Procure to Pay runs from identifying a need to paying the supplier. Order to Cash runs from a customer order to collected cash. Each spans multiple departments, because the failures cluster at the handoffs. Some firms, including The Hackett Group, call R2R account-to-report; it is the same process.
- What are the steps in the order to cash process?
- Six: credit management, order management, customer billing, cash application, collections management, and dispute or deduction management. Most O2C problems attributed to collections originate in credit or billing, which is why fixing collections alone rarely moves DSO.
- Does moving R2R, P2P and O2C to a shared services centre reduce cost?
- It reduces cost by roughly the wage differential and then stops, if you lift and shift the existing design. The larger prize is structural: formal end-to-end process ownership is associated with 55% lower order-to-cash process cost. If the centre inherits four reporting lines and nobody owning the whole, you have relocated the handoffs rather than removed them.
- Should we automate R2R, P2P and O2C with AI?
- Sequence matters. Hackett finds up to 75% performance advantage for process-led AI transformation, attributing the gains to redesigning how work flows rather than to any single automation. Establish ownership and fix the design first. Automation applied to a broken process produces a fast broken one.
- What do R2R, P2P and O2C mean for a GCC in India?
- These towers are usually the first scope a new capability centre takes on, and Indian rules change the design. E-invoicing requires businesses above ₹10 crore AATO to report to the IRP within 30 days, ruling out month-end batch processing in O2C. The Code on Wages affects payroll costs inside P2P. R2R’s cost base feeds the transfer pricing position. Design these in during setup.
What This Comes Down To
Governance is not a side hustle. The 55% number is not a technology finding. It is a governance finding wearing a process costume, and it says the cheapest intervention available to most finance functions is deciding who owns the thing end to end.
Great execution is thinking at the macro level and executing at the micro level. The three towers are the macro. The reconciliation cleared before day one, the purchase order raised before the goods arrive, the invoice reported to the IRP on the day it is issued: that is the micro. Both, or neither.
Astravise Services runs Agile Shared Services as part of an integrated operating model alongside Strategic CFO advisory and GCC Advisory, because a process tower reporting into a fragmented organisation will underperform whoever staffs it. Talk to us about a diagnostic.