Deduction KPIs: Definitions and Formulas
Deduction performance is usually invisible until someone puts a number on it. This guide defines the KPIs finance and credit teams use to measure a deduction workload: how large it is relative to sales, how old it is, how much gets challenged, how much comes back, and how much it costs to run the function. Every formula here is a definition, not a target. No metric on this page carries a benchmark, an industry average, or a statement of what counts as a good result. What counts as good depends on a company's retailer mix, its reason-code mix, and its own resourcing, and any number presented as typical without a named, checked source doesn't belong here. Looking for the underlying distinction between a deduction and a chargeback? Start at the glossary's chargeback-vs-deduction entry.
Balance and velocity metrics
Days Deductions Outstanding (DDO)
Days Deductions Outstanding measures how long a supplier's open deductions have been sitting relative to sales, expressed in days. It is the deduction-side counterpart of Days Sales Outstanding (DSO), a standard accounts-receivable metric: instead of measuring how slowly customers pay, it measures how large the open-deduction balance is relative to the sales volume generating it.
Formula:
DDO = (Open Deductions Balance divided by Total Sales in the Period) multiplied by Number of Days in the Period
Where:
- Open Deductions Balance = the dollar total of deductions taken but not yet resolved (not reversed, not written off, not otherwise closed), as of the measurement date.
- Total Sales in the Period = gross invoiced sales dollars over the same period the balance is being measured against.
- Number of Days in the Period = the length of the period used for the sales figure, stated explicitly by whoever runs the calculation.
What it's for: DDO gives you a single, sales-normalized figure for how much deduction balance is outstanding, so you can track it over time without it getting distorted by sales growing or shrinking on its own.
Common pitfalls:
- Using a sales figure from a different period than the balance date (a trailing sales figure mixed with a point-in-time balance, or the reverse) without stating which convention you're using.
- Counting a deduction that's been partially reversed as still "open" at its original full amount, instead of at its remaining unresolved amount.
- Comparing a DDO computed on gross sales to one computed on net sales (after returns) as if they were the same figure. They're not, unless both parties state and match the sales basis.
Deduction rate (share of sales)
The deduction rate expresses total deductions taken in a period as a percentage of gross sales in the same period. It is one common way to calculate deduction volume and velocity, not a validity metric: it says nothing about whether the deductions taken were correct, only how large they were relative to sales.
Formula:
Deduction Rate = (Total Deductions Taken in the Period divided by Gross Sales in the Period) multiplied by One Hundred Percent
Where:
- Total Deductions Taken = the dollar sum of all deductions posted against remittances in the period, across all reason codes, before any dispute outcome is known.
- Gross Sales = total invoiced sales dollars in the same period, before returns or allowances.
What it's for: it's a normalized way to compare deduction exposure across periods, retailers, or business units of different sizes, with no benchmark implied.
Common pitfalls:
- Mixing a deduction figure measured by count with a sales figure measured by dollars, or the reverse. The rate has to use the same unit on both sides.
- Using an invoice-date period for sales and a remittance-date period for deductions. A deduction against an invoice from an earlier period will otherwise inflate or deflate the rate, depending on which side lags.
- Treating this rate, on its own, as a measure of how much was wrongly deducted. It measures scale, not validity. Pair it with the valid-vs-invalid share metric below for that question.
Open deduction aging buckets
Deduction aging groups the open-deduction balance into buckets by how long each deduction has been open, the same technique used for accounts-receivable aging. It surfaces how much of the balance is old enough to be at risk of a retailer's dispute deadline passing, rather than treating the whole balance as one undifferentiated figure.
Formula:
For each open deduction, Age equals Current Date minus Deduction Posted Date. Assign the deduction's dollar amount to a bucket based on that age, then sum the dollar amounts within each bucket to get the bucket total.
Where:
- Deduction Posted Date = the date the deduction was taken on the remittance or ledger, stated explicitly (not the date it was discovered, and not the date a dispute was filed).
- Current Date = the date the aging report is run.
- Bucket boundaries = zero to thirty days, thirty-one to sixty days, sixty-one to ninety days, and ninety-plus days is a widely used convention, the same one used for standard receivables aging. It is presented here as a convention, not a rule. Retailer dispute windows differ from retailer to retailer and from this convention, so these buckets are a starting point, not a deadline calendar; see the guide to shortage-deduction disputes for how individual retailer windows actually work.
What it's for: it prioritizes review and dispute effort toward the oldest open dollars, especially against retailers whose dispute windows carry a fixed deadline after which a deduction can no longer be contested.
Common pitfalls:
- Anchoring age to the wrong date. "Posted date," "discovered date," and "dispute filed date" are three different clocks. Mix them across an aging report and you'll understate or overstate the true age.
- Not stating whether the clock restarts or continues when a deduction is partially resolved. A partial reversal that leaves a remainder open needs a stated, consistent rule.
- Aging a deduction that's already been disputed and is only awaiting a retailer response the same way as one that hasn't been worked at all. These are operationally different states, even though both are technically "open."
Dispute and outcome metrics
Dispute rate
The dispute rate is the share of deductions received in a period that are actually disputed, meaning formally challenged with the retailer, rather than accepted or left unworked. It is one common way to measure effort and coverage, distinguishing deductions that get reviewed and contested from ones that pass through unchallenged.
Formula:
Dispute Rate = (Number of Deductions Disputed in the Period divided by Number of Deductions Received in the Period) multiplied by One Hundred Percent
Where:
- Number of Deductions Disputed = count of deductions for which a dispute, meaning a formal challenge to the retailer, was submitted in the period.
- Number of Deductions Received = count of deductions posted against remittances in the period, regardless of what happens to them afterward.
What it's for: it shows how much of the incoming deduction volume is actually being worked, as opposed to accepted by default or left sitting in a queue.
Common pitfalls:
- Cohort mismatch: a deduction received near the end of a period is often disputed in the following period, so a strict same-period count can understate the true dispute rate for recent volume. State whether the metric is same-period or cohort-tracked.
- Counting a deduction as "disputed" the moment it's queued for review, before a dispute is actually submitted to the retailer. That overstates the rate relative to a definition based on submission.
- Using a dollar-based numerator against a count-based denominator, or the reverse. State which unit the rate is measured in and keep both sides consistent.
Recovery (or reversal) rate
The recovery rate, also called the reversal rate, measures, of the deductions actually disputed, what share came back as a credit or reversal from the retailer. It is one common way to measure outcomes for disputed dollars specifically, not for the whole deduction population, and it carries no fixed or typical value on this page.
Formula:
Recovery Rate = (Dollar Amount of Deductions Reversed divided by Dollar Amount of Deductions Disputed) multiplied by One Hundred Percent
Where:
- Dollar Amount of Deductions Reversed = the net dollar amount credited back or reversed by the retailer against disputed deductions in the period, netted to the original deduction amount.
- Dollar Amount of Deductions Disputed = the dollar total of deductions that were formally disputed in the same cohort.
What it's for: it shows the effectiveness of dispute effort on the subset of deductions actually challenged, separate from the dispute rate above, which measures how much got challenged in the first place.
Common pitfalls:
- Counting reversals twice: if a single deduction is partially reversed more than once, sum the reversals to the deduction's net recovered amount rather than adding every ledger reversal transaction as if each were a separate deduction.
- Measuring recovery against total deductions received instead of total deductions disputed. These are two different denominators, and they produce two different figures that aren't interchangeable.
- Mixing time windows: a deduction disputed in one period may not reverse until a later one. State whether recovery rate is measured by disposition-date cohort or strict same-period, and don't compare a same-period figure to a cohort figure.
Valid vs invalid deduction share
This metric splits reviewed deductions into those determined valid, meaning the retailer's deduction was correctly taken or the supplier concedes it, and those determined invalid, meaning the deduction should not have been taken, each as a share of total deductions reviewed. It is one common way to answer a validity question, distinct from the volume question the deduction rate answers.
Formula:
Valid Share = (Dollar Amount Determined Valid divided by Total Dollar Amount of Deductions Reviewed) multiplied by One Hundred Percent
Invalid Share = (Dollar Amount Determined Invalid divided by Total Dollar Amount of Deductions Reviewed) multiplied by One Hundred Percent
Where:
- Total Dollar Amount of Deductions Reviewed = deductions that have actually been assessed, not the full open balance, which may include unreviewed items.
- Determined Valid or Invalid = the outcome of that assessment, per whatever determination process the business uses, whether internal review, the retailer's final ruling, or both stated separately.
What it's for: it separates the question of how large the deduction workload is from the question of how much of it was actually correct, which is the input that should drive process or compliance fixes rather than dispute volume alone.
Common pitfalls:
- Treating an internal validity assessment and a retailer's final dispute ruling as the same category. They can disagree, and reporting one as if it were the other misstates the metric. State which determination source you're using.
- Comparing a valid/invalid share computed by dollar amount to one computed by count without noting the difference. A small number of large-dollar deductions can dominate a dollar-based share while being a minority by count.
- Reporting a share only for reviewed deductions as if it applied to the entire open balance, when a large share of the balance may be unreviewed and therefore undetermined.
Time to resolution
Time to resolution measures how long it takes, on average or at the midpoint, for a deduction to reach a final state, meaning reversed, written off, or otherwise closed, after it is received. It is one common way to measure cycle time for the whole deduction-handling process, not specific to disputes alone.
Formula:
Time to Resolution, for one deduction, equals Resolution Date minus Deduction Received Date. Report either the Mean Time to Resolution, the average across a cohort of deductions closed in a period, or the Median Time to Resolution, the midpoint value across that same cohort.
Where:
- Resolution Date = the date the deduction reached a final, closed state in the business's own records.
- Deduction Received Date = the date the deduction was first posted.
- Cohort = the set of deductions closed within the reporting period, chosen deliberately, with the choice stated.
What it's for: it tracks how quickly the deduction-handling process closes items out, which affects both cash timing and how much of the open balance is aging into risk of a missed dispute deadline.
Common pitfalls:
- Reporting a mean without stating it: a small number of very old, slow-to-resolve deductions can pull a mean well above the typical case, while a median wouldn't. State which statistic you're using, since the two answer different questions.
- Defining "resolved" inconsistently, for example counting a deduction as resolved when a dispute is submitted rather than when a final determination or reversal actually posts.
- Comparing a time-to-resolution figure across two different resolution definitions, such as closed in an internal system versus cash actually settled, as if they measure the same thing.
Write-off rate
The write-off rate is the share of closed deductions in a period that were absorbed as a loss, meaning written off, rather than reversed or otherwise recovered. It is one common way to measure how much of the resolved workload ends in an accepted loss, distinct from the recovery rate, which measures success among disputed dollars specifically.
Formula:
Write-Off Rate = (Dollar Amount Written Off in the Period divided by Total Dollar Amount of Deductions Closed in the Period) multiplied by One Hundred Percent
Where:
- Dollar Amount Written Off = deductions closed in the period by being accepted as a final loss, rather than reversed.
- Total Dollar Amount of Deductions Closed = all deductions that reached any final state in the period, whether written off, reversed, or otherwise settled.
What it's for: it distinguishes how much of the deduction workload ends in an absorbed loss, which is useful alongside the recovery rate for a full picture of where closed dollars land.
Common pitfalls:
- Conflating a deliberate write-off, meaning a decision not to pursue or a lost dispute, with a time-barred deduction, meaning one that missed a retailer's dispute deadline before it could be worked. These have different causes and, where the underlying data supports it, should be tracked as separate reasons rather than pooled into one write-off figure.
- Measuring the rate against total deductions received instead of total deductions closed in the period. The two denominators aren't interchangeable and will produce different-looking rates.
- Omitting partial write-offs, meaning a deduction partially reversed and partially written off, from either side of the ratio. This understates both the write-off and the recovery pictures if the partial amounts get dropped rather than allocated to both metrics.
Composition and cost metrics
Deductions by reason category
This metric shows what share of total deduction dollars falls into each reason category, for example shortage, pricing, or compliance, so effort can be targeted at the categories driving the most volume. It is one common way to measure composition rather than a rate against sales, and it depends on a consistent reason-code taxonomy across sources.
Formula:
For each Reason Category, Category Share equals (Dollar Amount in that Category divided by Total Deduction Dollars in the Period) multiplied by One Hundred Percent, summed across all categories to the full total of the period.
Where:
- Reason Category = a grouping of individual retailer reason codes into a smaller, consistent set of categories; see the category list at /codes.
- Total Deduction Dollars = the sum across every category in the same period.
What it's for: it identifies where deduction dollars concentrate, so review, dispute, and process-fix effort can be prioritized by category rather than treated as one undifferentiated pile.
Common pitfalls:
- Summing category totals across retailers whose reason-code taxonomies haven't been mapped to a single internal category set first. A shortage code at one retailer and a differently named but equivalent code at another have to be mapped before they can be added together, or the category mix won't be comparable retailer to retailer.
- Letting an unmapped or "miscellaneous" code bucket grow large without flagging it. A large residual category usually means the mapping is incomplete, not that the business genuinely has a large unclassified category.
- Comparing a category's percentage share across two periods when the underlying reason-code taxonomy changed between them. That changes the category definition and makes a period-over-period comparison misleading unless you remap to the same taxonomy.
Cost to process
Cost to process expresses the total cost of running the deduction-handling function, divided by how many deductions were processed, as a cost-per-deduction figure. It is one common way to measure unit economics for the function itself, separate from any measure of how much money was recovered or lost.
Formula:
Cost to Process, per deduction, equals Total Cost of the Deduction-Handling Function in the Period divided by Number of Deductions Processed in the Period.
Where:
- Total Cost of the Deduction-Handling Function = the labor, systems, and any allocated overhead cost attributable to working deductions in the period, per whatever cost-allocation method the business uses.
- Number of Deductions Processed = the count of deductions actually worked in the period, with the business stating whether that means touched at all or fully resolved.
What it's for: it gives you a per-unit cost figure for the deduction-handling function, useful for staffing and process decisions, and for evaluating whether a process change actually reduces unit cost.
Common pitfalls:
- Choosing a "processed" definition, touched versus fully resolved, without stating it. The two produce very different denominators, and therefore very different cost-per-deduction figures, from the same cost base.
- Allocating shared overhead into the numerator using an implicit or inconsistent method between periods. That makes period-over-period comparisons of this metric unreliable unless the allocation method is stated and held constant.
- Comparing a cost-to-process figure that includes only direct labor cost to one that also includes systems or overhead cost, as if the two were computed the same way.
Frequently confused metrics
What is the difference between the deduction rate and the dispute rate?
- The deduction rate expresses total deductions taken as a percentage of sales; it says nothing about what happens to those deductions afterward. The dispute rate expresses how much of the deductions received actually get formally challenged. A business can carry a low deduction rate and a low dispute rate at the same time, or a high deduction rate paired with a high dispute rate. The two just measure different things.
What is the difference between the recovery rate and the write-off rate?
- The recovery rate measures, of disputed dollars, how much came back as a reversal. The write-off rate measures, of all closed dollars whether disputed or not, how much was absorbed as a loss. A deduction that's never disputed at all can still be written off, which is why the write-off rate's denominator, deductions closed, is broader than the recovery rate's denominator, deductions disputed.
What is the difference between Days Deductions Outstanding and time to resolution?
- Days Deductions Outstanding measures the size of the current open balance relative to sales, at a point in time. Time to resolution measures how long it takes an individual deduction, once received, to reach a final closed state. A business could post a fast average time to resolution but still carry a large DDO if deduction volume is high relative to sales, or the reverse.
What is the difference between valid vs invalid share and the write-off rate?
- Valid vs invalid share is a determination about whether a deduction was correctly taken in the first place. Write-off rate is about the final disposition of a closed deduction, which can happen regardless of validity, for example writing off a deduction the business believes is invalid because the dollar amount doesn't justify further dispute effort. A deduction can be determined invalid and still be written off rather than recovered.
See how your own deductions break down
Every KPI on this page depends on having deduction data in one place, mapped to a consistent reason-code taxonomy and time basis. The definitions and formulas on this page hold for any deduction-management setup.