Collections Metrics: What to Measure When You Want to Get Paid Faster
Collections Metrics: What to Measure When You Want to Get Paid Faster
Ask a business how its receivables are performing and you'll usually get one number: days sales outstanding. It is the most reported metric in the discipline and among the least useful for the purpose it's used for, because it moves with sales volume and timing independently of how well you actually collect. A strong sales month raises DSO even if you collected everything perfectly; a weak one lowers it while collections deteriorate. This guide covers the measures that actually drive decisions — what share of collectible money you collected, which balances are moving in the wrong direction, where recovery stops being economic, and how to build a dashboard that changes behavior rather than describing it after the fact.
What you'll learn
Why DSO misleads
Days sales outstanding expresses your receivables balance in terms of days of sales. It's a legitimate long-run indicator and a poor month-to-month performance measure, for four reasons:
- Sales timing distorts it. A large invoice issued on the last day of the period sits in receivables in full and inflates DSO, regardless of collection performance.
- Sales volume distorts it. Growing revenue mechanically raises DSO; declining revenue lowers it. A business celebrating falling DSO during a sales slump is celebrating the slump.
- It blends different terms. Customers on net 15 and net 60 average into one figure that describes neither.
- It can be improved by things that aren't improvements — selling to fewer customers on shorter terms, or factoring receivables, both reduce DSO without collections getting better.
The useful version is best possible DSO — what your DSO would be if every customer paid exactly on terms — compared against actual. The gap between the two isolates the collection problem from the terms problem, which are different things requiring different fixes.
Keep DSO. Report it as a trend over quarters. Just stop treating it as the measure of whether collections is working.
Collection effectiveness index
The metric that answers the question DSO is asked to answer: of the money that was available to collect this period, what share did you actually collect?
Because the calculation compares collections against what was collectible — beginning receivables plus period sales, less ending receivables, measured against beginning receivables plus sales less ending current receivables — it isolates performance from sales fluctuation. A value approaching one hundred percent means you collected nearly everything that was available.
Why it's the better summary number:
- It's not distorted by growth, so a growing business can tell whether collections is keeping pace.
- It's comparable period to period in a way DSO isn't.
- It responds to actual effort, which makes it a legitimate performance target rather than one that rewards the wrong behavior.
Two implementation notes. Track it monthly and look at the trend rather than any single value, since month-end timing still introduces noise. And segment it — by customer type, product line, and channel — because a stable overall figure can conceal one deteriorating segment offset by another improving.
Reading an aging report properly
Nearly every business produces an aging report and most read it wrong, in two specific ways.
Read percentages, not dollars. An aging report in absolute dollars grows with your business, so a larger over-90 balance may simply reflect more revenue. The share of total receivables in each bucket, tracked over time, is what shows direction — and direction is what you act on.
Age from the due date, not the invoice date. Aging from invoice date blends customers on different terms and makes a net-60 customer look delinquent at day 45 when they're current. Days past due is the meaningful measure.
What a properly read aging tells you:
- The current bucket's share is your headline health indicator. A falling share means invoices are aging past terms faster than they're being collected.
- Bucket concentration — whether the over-90 balance is a hundred small invoices or three large ones, which calls for entirely different responses.
- Customer concentration within buckets, since one large slow-paying customer can dominate the picture and is a relationship conversation rather than a collections process problem.
- Deduction and dispute balances separated out, because these are not collection problems and mixing them in produces both bad measurement and wasted contact — the routing point in our deductions guide.
Roll rates: the early warning
Roll rates measure the percentage of balances moving from one bucket to the next in a period — current to 30, 30 to 60, and onward. They are the earliest reliable indicator available, because they detect a change in flow before it accumulates into a visible change in balances.
The value is timing. An aging report shows a problem after it has formed; a rising current-to-30 roll rate shows the problem forming while the aging still looks acceptable. For a business with a quarterly reporting rhythm, that difference is a full quarter of response time.
How to use them:
- Track each roll rate monthly, and watch the trend rather than the level.
- The current-to-30 rate is the most valuable, because it's furthest upstream and because balances at 30 days are the most recoverable — the decay curve in our collections framework means intervention here is worth far more than the same effort later.
- Segment by customer cohort — new customers rolling faster than established ones is a credit policy signal, not a collections signal, and points back to the terms decisions in our credit application guide.
- Watch for rate changes without balance changes, which is exactly the early signal the metric exists to provide.
Recovery by age and segment
Every business working receivables has a decay curve — the percentage recovered as a function of how old the balance was when work began — and almost none of them have measured their own.
Building it requires tracking, for each cohort of accounts: age at first contact, the balance, and whether and when it was recovered. The output is a table showing recovery rate by age band, and it's the foundation for nearly every decision in the function.
What it enables:
- Placement timing. Knowing where your own curve falls sharply tells you when internal effort stops working and placement should occur — the decision framework in our agency guide.
- Write-off thresholds. Expected recovery equals recovery rate times balance; where that falls below cost of pursuit, continuing destroys value, per our write-off analysis.
- Prioritization. Working accounts in expected-value order rather than balance order or age order, which is the single largest efficiency gain available to a capacity-constrained team.
- Segment differences. Recovery rates by customer type, size, industry, and reason for non-payment frequently differ enough to justify entirely different treatment.
The related measure worth tracking alongside: promise-to-pay kept rate. What share of customers who commit to a date actually pay by it, broken down by how the commitment was obtained. This is a direct measure of whether your contact process produces real commitments or polite ones, and it's actionable in a way few metrics are.
Cost to collect
Recovery that costs more than it returns is activity rather than value, and only this metric surfaces it.
Cost to collect is total collections expense — staff time loaded, systems, agency commissions, legal costs — as a percentage of amounts collected. It should be calculated in aggregate and, more usefully, by segment and balance band.
What the segmented version reveals, reliably:
- Small-balance accounts are frequently uneconomic to work manually, which is the argument for automation rather than for abandoning them — automation moves the boundary at which pursuit stops making sense.
- The largest hidden cost is opportunity cost. Staff time on a six-month-old account is time not spent on a thirty-day one where recovery odds are far higher, which quietly manufactures the next cohort of write-offs.
- Legal and agency costs need to be measured against incremental recovery, not gross recovery — the relevant question is what those channels recovered beyond what would have come in anyway.
Two related figures worth having: accounts worked per person per day, which indicates capacity and whether the queue is realistic, and contact rate — the share of attempts reaching the right person — which is frequently the binding constraint and is fixable through data quality rather than effort.
Quality and relationship metrics
Collections operates on customers you generally want to keep, which means recovery numbers alone are an incomplete picture.
- Dispute rate — the share of contacted accounts that raise a dispute. A high rate points upstream at invoicing accuracy or fulfillment rather than at collections.
- Complaint volume, tracked and reviewed. This is both a compliance control and a signal about tone, and it should be reviewed by someone who can change the scripts.
- Retention among contacted customers versus comparable customers who weren't. This is the metric that tells you whether your collections process is costing you future revenue, and almost nobody measures it.
- Payment plan completion rate by plan structure, which tells you which arrangements actually get paid — per our plan design guide.
- Self-service share — the proportion of payments arriving without human contact, which is the cheapest recovery available and a direct measure of whether your payment paths are working.
- Compliance exceptions, including contact frequency violations and disputed accounts still receiving collection contact, which is the control our compliance guide treats as non-negotiable.
Building the dashboard
A dashboard that describes performance is a report. One that changes behavior has three properties: it's short, it's segmented, and every number has an owner and a threshold.
Weekly, operational:
- Roll rate, current to 30
- Accounts worked and contact rate
- Promises made and promises kept
- New disputes raised
- Watchlist accounts with no action in seven days
Monthly, management:
- Collection effectiveness index, total and by segment
- Aging by percentage, with the trend
- All roll rates
- Recovery rate by age band
- Cost to collect by segment
- Complaint volume and compliance exceptions
Quarterly, strategic:
- DSO trend and best possible DSO gap
- Write-offs against forecast
- Recovery decay curve, refreshed
- Retention among contacted customers
- Credit policy performance by cohort — whether the customers you approved are the ones paying
Two design principles that determine whether any of it matters. Every metric needs a threshold and an owner, because a number with no assigned response is decoration. And segment everything — the recurring lesson across this desk's measurement coverage is that blended figures conceal the very variation you need to act on, and a stable total is frequently two offsetting movements you'd want to know about.
Measurement that comes with the process
HL Hunt AI Debt Collection works every account on a defined cadence under your own brand and reports what actually happened — roll rates, recovery by age band, promise-kept rates, plan completion by structure, and self-service share — so the dashboard is a byproduct of the work rather than a separate project.
Frequently asked questions
It moves with sales volume and timing independently of collection performance, blends customers on different terms, and can improve for reasons that aren't improvements. Useful as a long-run trend, misleading month to month.
The share of receivables available for collection in a period that you actually collected. Because it measures against what was collectible rather than against sales, it isolates performance from sales fluctuation.
The percentage of balances moving from one delinquency bucket to the next. It detects a change in flow before it accumulates into visible balances, which makes it the earliest reliable warning available.
When expected recovery — your own recovery rate at that age, times the balance — falls below the loaded cost of pursuit. That requires measuring your own decay curve rather than assuming one.
Key takeaways
- DSO is distorted by sales volume and timing; use collection effectiveness index as the summary performance measure instead.
- Read aging by percentage and age from the due date, not the invoice date.
- Roll rates are the earliest warning — the current-to-30 rate is the most valuable number in the function.
- Measure your own recovery decay curve; it drives placement timing, write-off thresholds, and prioritization.
- Cost to collect by segment reveals where pursuit destroys value and where automation moves the boundary.
- Track quality alongside recovery — dispute rate, complaints, plan completion, and retention among contacted customers.
Fix the current-to-30 roll rate first
Most recovery value is decided in the first thirty days. HL Hunt AI Debt Collection works every invoice from the due date with segmented messaging, self-service payment and plans, and compliance enforced automatically — which is where the roll rate actually moves.
This guide is educational and does not constitute financial or accounting advice. Metric definitions vary across organizations; consistency of definition over time matters more than matching any particular convention.