Order to Cash Process: Steps and Owners
Most companies treat order to cash as an accounting problem. It is not. It is a sales problem, a legal problem and an operations problem that happens to end at a bank account, and the money it wastes is not small. The Hackett Group's 2025 Working Capital Survey, which analyzed the top 1,000 US publicly traded nonfinancial companies on 2024 data, found $1.7 trillion trapped in excess working capital, equal to 35% of gross working capital and 11% of aggregate revenue. Roughly $600 billion of that sits in receivables, driven by days sales outstanding.
That number describes large public companies. The pattern in a mid-sized business is identical and usually worse, because there is nobody whose actual job is the whole cycle. Sales owns the customer, finance owns the invoice, operations owns the delivery, and the handoffs between them belong to nobody. Cash falls through those gaps.
This guide is not a software comparison. It is the method I use when a company wants to fix the cycle rather than buy a tool: what the stages actually are, who owns each one by name, where the process breaks, which metrics tell the truth, what to automate first, and how to get a measurable result in ninety days.
What order to cash actually covers
Order to cash, often shortened to O2C, is everything between a customer deciding to buy and the money arriving and being applied to the right invoice. Most people describe it in three steps. In practice there are seven, and the ones people forget are where the losses live.
Stage 1: customer setup and credit. The customer record is created, credit is assessed, terms are set, tax status and billing entity are captured. This stage takes minutes and determines months of pain if done badly.
Stage 2: order capture and validation. The order enters the system, with pricing, discounts, quantities, delivery dates and any contractual specifics. Errors here surface much later, as disputed invoices.
Stage 3: fulfillment and proof of delivery. Goods ship or services are delivered, and evidence of that is captured in a form the customer will accept. No proof, no payment, however clear the contract.
Stage 4: invoicing. The invoice is produced, it matches what was ordered and delivered, it carries the references the customer's accounts payable system requires, and it reaches the right recipient through the right channel.
Stage 5: collections. Systematic follow-up before and after due date, segmented by customer value and risk, with escalation that actually escalates.
Stage 6: dispute and deduction management. Someone owns the difference between what was invoiced and what the customer intends to pay, resolves it against evidence, and closes it within a deadline.
Stage 7: cash application and reconciliation. Payments are matched to invoices, remittance is interpreted, short payments are routed, and the ledger reflects reality.
The cycle only works if every stage has a name attached to it. Not a department, a person. Departments do not return calls.
The handoff map, and why the gaps cost more than the stages
Each stage in isolation usually runs fine. The money disappears at the boundaries, and there are five of them.
Sales to finance, at credit. The salesperson wants the order, credit wants protection. When credit is a rubber stamp, bad receivables enter the book. When credit is a black box, deals stall and sales learns to route around it. Both failures come from the same cause: no published rule that says what gets approved automatically, what needs review and how long review takes.
Sales to operations, at order accuracy. An order written with a shorthand product code, a verbally promised discount or a delivery date nobody confirmed becomes a dispute sixty days later, and by then the person who made the promise has moved on to next quarter.
Operations to finance, at proof of delivery. Finance invoices what the system says was shipped. If the signed delivery note, the timesheet or the acceptance email is not attached to the transaction, any customer who wants an extra thirty days now has a free option.
Finance to customer, at invoice delivery. An invoice that arrives without the purchase order number, to a general inbox, or outside the customer's portal, is not late because the customer is slow. It is late because it never entered their approval queue. This single failure explains more overdue balances than any credit problem, and it is entirely self-inflicted. The mirror image of this dynamic, seen from the buying side, is described in the guide on procure to pay process design and controls, which is worth reading precisely because your customer runs something like it.
Finance to finance, at cash application. A payment arrives covering fourteen invoices with three deductions and no remittance detail. Until someone unpicks it, those invoices still look open, collections chases a customer who has paid, and the relationship takes the damage.
Write these five boundaries on a page with a name next to each. In most companies, two of the five have no owner at all. That is the diagnosis, and it takes an hour.
Stage by stage: the owner, the rule and the failure mode
Customer setup and credit
Owner: finance, with a published rule that sales can read.
The rule should fit on one page: order value thresholds that pass automatically, thresholds that need a credit check, what evidence is required, the maximum turnaround time, and who decides in a disagreement. Terms should be a decision, not a default inherited from the first deal a salesperson closed in 2019.
Failure mode: credit limits set once and never revisited. A customer who was good three years ago and is now slow keeps their old limit, because nobody owns periodic review. Set a review cadence by segment and hold it.
The customer master record matters more than it looks: legal entity, billing address, tax identifiers, purchase order requirements, portal details, invoice delivery channel, contact for disputes. Every field missing here becomes a phone call later. The connection between how you bring a customer on board and how fast they eventually pay is direct, and I covered the wider onboarding sequence in the guide on the customer onboarding process, stages and owners.
Order capture and validation
Owner: sales operations, or the person who runs the quoting system.
Every order should pass three checks before it is accepted: the price matches an approved price list or a signed contract, the terms match the customer record, and the deliverable is described in terms the invoice can repeat. If your invoice cannot quote the order back to the customer in their own language, expect a dispute.
Failure mode: the side agreement. A discount, a rebate, an extended payment term or a free service promised by email and never entered into the system. It surfaces as a deduction, months later, and finance has no evidence to refuse it. The structural answer is that commercial terms live in the contract record and flow into billing automatically, which is the argument made in the guide on contract lifecycle management.
Fulfillment and proof of delivery
Owner: operations, with a defined evidence standard.
Decide what counts as proof for each product or service line: signed delivery note, carrier confirmation, acceptance email, approved timesheet, system log. Then attach it to the transaction at the moment it happens, not when a dispute starts. Evidence gathered retroactively is evidence you will not find.
Failure mode: services delivered without an acceptance step. In project work, the gap between "we did the work" and "the client agrees we did the work" is where the largest write-offs happen.
Invoicing
Owner: finance, measured on first-time acceptance rather than on invoices issued.
The invoice must carry whatever the customer's accounts payable system needs to process it without human intervention: purchase order number, line references, cost center, correct legal entity, agreed currency, required attachments. Send it through the channel they actually read, which for larger customers means their portal, not an inbox.
Failure mode: invoicing in batches at month end. A ten-day average delay between delivery and invoice adds ten days to every single receivable in the book, and no collections effort can win that time back. Invoice on the event, not on the calendar.
Collections
Owner: a named person, even part time, with authority to escalate.
Segment the ledger rather than working it top to bottom. A workable split: high value and strategic accounts get a relationship-based approach with contact before due date; mid-market gets a fixed reminder sequence; small balances get automation and a hard stop rule. Contact before the due date is the highest return activity in the entire cycle, because it converts a payment failure into a process question while there is still time to fix it.
Failure mode: escalation that never escalates. If the third reminder says the same thing as the first, customers learn your deadlines are decorative. Define what happens at each step and do it, including the uncomfortable step of pausing new shipments.
Disputes and deductions
Owner: whoever caused the category, coordinated by finance.
Every dispute has a root cause, and the cause is almost never finance: wrong price, wrong quantity, damaged goods, missing documentation, service not accepted, promotion not applied. Code them, report them monthly to the function responsible, and give each dispute a resolution deadline.
Failure mode: treating disputes as a collections problem. If finance negotiates its way through them one by one, the underlying causes never get fixed and the same deductions come back every quarter.
Cash application and reconciliation
Owner: finance, measured on match rate.
The aim is that most payments match automatically and the exceptions reach a person quickly. Ask customers for structured remittance, use the payment reference fields your banking channel supports, and never let unapplied cash sit for more than a few days. The parallel discipline on the payables side is described in the guide on how to automate accounts payable, and the two functions should compare notes, because they are mirror images of each other.
Failure mode: unapplied cash used as a cushion. When the reconciliation account carries a permanent balance, every aging report is fiction and collections chases customers who have already paid.
The metrics, in causal order
Most companies track days sales outstanding and nothing else. DSO is an outcome, and outcomes move last. Track the causes.
| Metric | What it measures | Warning threshold |
|---|---|---|
| Time from delivery to invoice | Speed of the billing engine | Above 2 working days |
| First-time invoice acceptance rate | Invoice quality at source | Below 95% |
| Percentage of invoices disputed | Upstream order and delivery accuracy | Above 3% |
| Average dispute resolution time | Ownership and evidence quality | Above 15 days |
| Cash application auto-match rate | Remittance and reference discipline | Below 85% |
| Current ratio of the ledger | Share of receivables not yet overdue | Below 80% |
| Days sales outstanding | The overall outcome | Rising for two consecutive quarters |
| Bad debt as share of revenue | Credit quality and escalation discipline | Above sector norm, rising |
The order matters. If invoice acceptance is 80%, working on collections is treating the symptom: one invoice in five was never going to be paid on time regardless of how politely you asked. Fix the source, then measure the outcome.
Two cautions. First, DSO moves with sales mix and seasonality, so compare like with like and use a rolling measure. Second, best possible DSO, calculated as if every customer paid exactly on terms, tells you how much of your DSO is actually collectible improvement versus terms you granted yourself. Many companies discover their real problem is the terms they agreed to, not the collections they run.
Seven failure modes I see repeatedly
One: terms granted as a sales tool. Payment terms are pricing. Ninety day terms on a thin margin can erase the profit on the deal, and the salesperson who granted them is measured on the booking, not on the cash. This is a compensation design problem before it is a finance problem, and it connects directly to how you set prices in the first place, a topic covered in the guide on B2B pricing strategy.
Two: no contact before the due date. Companies that call at day 45 on 30 day terms are negotiating. Companies that confirm at day 20 that the invoice is approved and scheduled are managing a process. The second costs less and works better.
Three: the ledger worked top to bottom. Starting at the largest balance feels rational and wastes the day, because large customers are usually late for structural reasons that a phone call cannot solve. Segment first.
Four: disputes with no deadline. An open dispute with no resolution date is a permanent hole. Give every dispute an owner and a date, and report the ones that pass it.
Five: month end billing. Everything queues, the team burns three days, and every customer receives their invoice at the same time as everyone else's, competing for the same approval attention. Continuous billing flattens the peak and speeds the cycle.
Six: credit review that happens once. Customer risk changes. Reviewing limits only at onboarding means the deteriorating accounts are exactly the ones carrying the largest exposure.
Seven: no single owner for the cycle. When sales, operations and finance each own a piece, the end-to-end result belongs to nobody. Someone has to be accountable for cash conversion as a number, which is the organizational argument behind revenue operations.
What to automate first, and what not to
Automation follows process, never the other way around. Automating a broken invoice flow produces broken invoices faster and adds a subscription. That said, four areas repay quickly when the rules are clear.
Cash application matching. The highest return, lowest risk automation in the cycle. Matching payments to invoices from remittance data is a pattern recognition task with an objective right answer, and a strong match rate frees a person from a job nobody enjoys.
Reminder sequences for the tail. The small and mid balances that consume most of the contacts and least of the value. A defined sequence, sent automatically, with escalation to a person only on exception.
Invoice delivery and portal submission. If a meaningful share of your customers require submission through their own portals, every manual submission is a delay and a failure point.
Dispute routing and coding. Getting each dispute to the right owner automatically, with a code and a clock, is what makes root cause analysis possible at all.
What not to automate early: credit decisions on material exposures, collection conversations with strategic accounts, and dispute resolution where the evidence is ambiguous. Those are judgment tasks, and a system that guesses will cost you a customer.
Two rules before any tool decision. First, write the rule you intend to automate in plain language on one page. If you cannot, the tool will not fix it. Second, measure the process for a month first, because you cannot evaluate an improvement against a baseline you never took.
Four maturity stages
Stage 1: reactive. Invoicing happens in batches, collections happen when cash gets tight, disputes are resolved by whoever picks up the phone. DSO is known but not explained. Most companies under twenty million in revenue live here, and many above it.
Stage 2: defined. Stages have owners, the credit rule is published, a reminder sequence exists, disputes are coded. Cash improves without new software, usually by ten to twenty days of DSO in the first year, because most of the gap was process rather than effort.
Stage 3: measured. The causal metrics are tracked monthly, dispute root causes go back to the function responsible, credit limits are reviewed on a cadence, the aging report is trusted because cash application is current.
Stage 4: optimized. Automation carries the routine volume, people handle exceptions and relationships, terms are a commercial decision made with margin visibility, and the cycle is reported as one number owned by one person.
Skipping from stage one to stage four by buying software is the most expensive mistake in this area. The tool inherits the confusion.
What it costs to fix, and where the money comes back
The visible costs are software licenses and perhaps a collections resource. The real costs are elsewhere.
Cleaning the customer master. Duplicate records, wrong legal entities, missing purchase order requirements, obsolete contacts. This is unglamorous work that determines whether anything downstream functions, and it always takes longer than planned.
Rebuilding the invoice template and data. Getting every required reference onto the invoice, in the format each large customer needs, is a project of weeks, not an afternoon.
Writing the rules. Credit policy, escalation ladder, dispute codes and deadlines, terms approval matrix. A few days of senior time, and the single highest return activity in the list.
The sales conversation. Making terms and credit part of the commercial discussion changes behavior that has been rewarded for years. Expect friction, and expect it to be the real bottleneck.
| Source of return | When it shows | How to measure it |
|---|---|---|
| Faster invoicing | Month 1 | Average days from delivery to invoice |
| Higher first-time acceptance | Month 2 | Share of invoices paid without dispute |
| Contact before due date | Month 2 | Share of current ledger contacted pre-due |
| Cash application automation | Month 2 | Auto-match rate, unapplied cash balance |
| Dispute root cause removal | Month 4 | Disputed value by cause, month on month |
| Credit discipline | Month 6 | Bad debt as a share of revenue |
The economics are simple enough to calculate on the back of an envelope. Take annual revenue, divide by 365, multiply by the number of DSO days you expect to recover. That is cash released once, permanently, at zero cost of capital. For a company doing twenty million with a realistic ten day improvement, it is roughly five hundred and fifty thousand of cash that was already yours.
Deloitte's Working Capital Roundup 2025, covering more than 2,300 companies, found that recent cash conversion gains have largely come from tactical moves in payables and inventory, while DSO rose under collection pressure, and concluded that durable improvement is more likely to come from stronger forecasting, workflow automation and closer collaboration. That is the distinction that matters: stretching your own payables is a one-time trick your suppliers will notice, while fixing the order to cash cycle compounds.
Eighteen questions to ask your own team
Run this as a ninety minute session with sales, operations and finance in the same room. The disagreements are the finding.
- Who owns the order to cash cycle end to end, by name?
- What is our average time from delivery to invoice, measured not estimated?
- What share of invoices are paid without a single query?
- What are the top three dispute causes by value this year?
- Who resolves disputes, and what is their deadline?
- When did we last review credit limits on our twenty largest accounts?
- Which customers have terms that do not match their contract?
- How many customers require portal submission, and do we always use it?
- What is our cash application auto-match rate?
- How much unapplied cash is sitting in reconciliation right now?
- What percentage of the ledger is contacted before the due date?
- What happens at reminder three, and does it actually happen?
- Who has authority to put an account on hold, and when was it last used?
- How do side agreements reach the billing system?
- What evidence of delivery do we hold for our five largest open invoices?
- Is anyone in sales measured on cash rather than bookings?
- What is our best possible DSO, and how far are we from it?
- If our largest customer went to ninety days tomorrow, what would we do?
Question fifteen is the one that goes quiet. If nobody can produce delivery evidence for the five largest open invoices inside ten minutes, the collections problem is actually a documentation problem, and no amount of chasing will fix it.
A case from another sector, the same principle
The pattern behind all of this shows up outside finance. In a medical center I worked with, delivered capacity grew by twenty percent without hiring anyone and without buying equipment. The hours were already there, but they were badly distributed: schedules built by habit, sequences that created dead time, information someone had to hunt for while a patient waited.
An order to cash cycle behaves the same way. Before hiring a collections person because receivables are out of control, it is worth measuring how many invoices go out wrong, how many days pass between delivery and billing, and how much cash is sitting unapplied. In most companies the answer is that the capacity was there and the sequence was wrong. In one distribution business I advised, tightening the commercial process rather than adding headcount was also what produced a thirty percent sales increase, because the same discipline that fixes billing accuracy fixes quoting accuracy.
If you want to know whether your own bottleneck is the credit rule, the invoice, the evidence or the follow-up, describe how your cycle runs today and I will tell you where I would look first. In most cases the starting point is not the one the company had in mind when it decided it had a collections problem.
Roadmap for 30, 60 and 90 days
Days 1 to 30: measure and assign
Week 1. Take the baseline. Days from delivery to invoice, first-time acceptance rate, current ratio of the ledger, unapplied cash, DSO and best possible DSO. Pull the last ninety days of disputes and code them by cause. Without these numbers you will never prove the project worked.
Week 2. Draw the handoff map and put a name on each of the five boundaries. Where there is no name, appoint one. This step costs nothing and resolves a surprising share of the problem on its own.
Week 3. Write the credit rule and the escalation ladder, each on a single page. Get sales leadership to sign both, in a meeting, not by email.
Week 4. Segment the ledger into three tiers and define the contact approach for each, including the pre-due contact for tier one.
Days 31 to 60: fix the source
Move invoicing from batch to continuous for the largest customer segment. Fix the invoice template so it carries every reference the top twenty customers require, verified with them rather than assumed.
Attach delivery evidence to transactions at the point of delivery, with a defined standard per product line. Start the pre-due contact routine on tier one accounts and record what you learn, because the reasons customers give in those calls are your dispute root causes arriving early.
Clean the customer master for the top one hundred accounts: legal entity, purchase order requirement, delivery channel, dispute contact.
Days 61 to 90: automate and hold
Automate cash application matching and the reminder sequence for the tail. Introduce a fifteen minute weekly review of disputes over the deadline, attended by the functions that cause them, not only by finance.
Report the causal metrics monthly alongside DSO. At day ninety, compare against the baseline. If DSO has not moved but first-time acceptance and time to invoice have, the improvement is coming and the lag is normal. If none of them moved, the rules were written and never enforced, which is a management problem with a short and uncomfortable fix.
When the process is not the real problem
There is a case where all of this is wasted effort, and it should be said plainly. If the company sells on terms it cannot afford in order to hit quarterly numbers, and nobody in sales carries any consequence for cash, then the finance team is being asked to collect its way out of a commercial decision. It cannot be done. The receivables will grow in line with revenue, and every process improvement will be consumed by the next quarter's concessions.
The signal is easy to spot: ask who can approve ninety day terms and how often it happens. If the answer is that anyone can and it happens routinely, the order to cash problem is a governance problem wearing a finance costume.
The same applies to evidence. Without someone accountable for proof of delivery at the moment of delivery, every disputed invoice becomes a negotiation you enter without documents, and the outcome of that negotiation is decided before it begins.
Fix the rule, then the sequence, then the tool. Companies that reverse that order pay twice, and the second payment is the credibility of whoever proposed the project.
Where order to cash meets the rest of the business
Three boundaries get confused often enough to be worth drawing explicitly, because each one hides a different kind of ownership gap.
Quote to cash versus order to cash. Quote to cash starts earlier, at the configuration of the offer and the quote itself, and it includes pricing approval and contract signature. Order to cash starts once the order exists. The distinction matters because a large share of billing disputes are created before the order is placed, in a quote that promised something the billing system cannot express. If your dispute codes keep pointing upstream, your problem is quote to cash and you are fixing the wrong half.
Order to cash versus record to report. Record to report is the closing cycle: accruals, reconciliations, financial statements. It consumes what order to cash produces. A month end close that is regularly delayed by unapplied cash or unresolved disputes is not a closing problem, it is a receivables problem presenting itself late. When the controller complains about the close, look at the aging report first.
Order to cash versus customer service. Customers do not distinguish between a service ticket and a billing query, and routing a billing question through a service desk that cannot see the invoice adds days to every resolution. Decide who answers billing questions, give them visibility into the ledger and the delivery evidence, and publish a single contact. Splitting this responsibility across two teams is one of the quietest causes of slow payment, because the customer waits while your organization decides internally who owns the answer.
There is a fourth boundary that only appears in companies selling subscriptions or recurring services: the link between the contract's billing schedule and the revenue recognition rules. When those two live in different systems maintained by different people, the invoice and the reported revenue drift apart, and reconciling them becomes a permanent monthly tax. Deciding early that the contract record is the single source for both is cheaper than any correction later.
FAQ
What are the order to cash process steps and owners?
There are seven steps, and each needs a named owner. Customer setup and credit belongs to finance, working from a published rule sales can read. Order capture and validation belongs to sales operations. Fulfillment and proof of delivery belongs to operations, with a defined evidence standard. Invoicing belongs to finance, measured on first-time acceptance rather than volume issued. Collections belongs to a named person with authority to escalate. Dispute management is coordinated by finance but owned by the function that caused the dispute category. Cash application belongs to finance, measured on auto-match rate. The gaps between these owners cost more than the steps themselves.
How long is a realistic order to cash improvement project?
Ninety days to a measurable result is realistic for a mid-sized company: thirty days to measure and assign ownership, thirty to fix invoicing speed and accuracy at the source, thirty to automate the routine volume and establish the review cadence. Ten to twenty days of DSO improvement in the first year is a common outcome, and most of it comes from process rather than software. Full maturity, where the cycle is automated and terms are a deliberate commercial decision, usually takes twelve to eighteen months.
Which metric should we fix first, DSO or something else?
Not DSO. It is an outcome and it moves last. Start with time from delivery to invoice and first-time invoice acceptance rate, because those two determine the ceiling on everything downstream. If one invoice in five is queried, no collections effort can produce a good DSO. Once invoices go out fast and correct, move to cash application auto-match rate and the share of the current ledger contacted before due date. DSO then improves as a consequence, and you will be able to explain why.
Should we automate order to cash or fix the process first?
Fix the process first, then automate the parts whose rules you can write in plain language on one page. Cash application matching, reminder sequences for small balances, invoice delivery to customer portals and dispute routing are the four areas that repay quickly. Credit decisions on large exposures, conversations with strategic accounts and ambiguous disputes should stay with people. Automating a broken invoicing flow produces wrong invoices faster and adds a subscription cost to the original problem.
Why are so many invoices paid late when the customer is solvent?
Usually because the invoice never entered the customer's approval queue correctly. Missing purchase order number, wrong legal entity, sent to a general inbox instead of the required portal, or missing the delivery evidence their process demands. From the customer's side this is not a decision to pay late, it is an invoice that cannot be processed. The fix is to confirm with each large customer exactly what their accounts payable system requires and then always send it that way, which typically removes a larger share of overdue balance than any collections change.
Who should own collections in a company without a credit department?
One named person with part of their time formally allocated and with authority to escalate, including the authority to place an account on hold with a defined approval path. The role does not need to be full time below a certain ledger size, but it cannot be a shared responsibility, because shared responsibility produces inconsistent contact and customers calibrate to the weakest follow-up they experience. What matters more than the title is that the person can trigger consequences, not just send reminders.
How do payment terms affect profitability?
Directly, and more than most sales teams realize. Terms are pricing. On a thin margin, extending from thirty to ninety days can consume a significant part of the profit on the deal once the cost of financing that gap is counted, and the effect is invisible in a booking report. The practical control is an approval matrix that ties term length to deal margin and customer risk, plus at least one sales incentive linked to cash rather than to bookings alone. Without that, the collections function spends its life managing the consequences of decisions it never saw.