Metric Definition
Correct coding rate
Track from
Expense categorisation accuracy
Expense categorisation accuracy is the percentage of expense transactions that are coded to the correct general-ledger category. It measures how trustworthy your spend data is before anyone reports on it. When categorisation is wrong, every downstream budget, forecast, and tax figure inherits the error.
7 min read
What is expense categorisation accuracy?
Expense categorisation accuracy is the percentage of expense transactions coded to the correct general-ledger category. If 1,000 transactions are reviewed and 940 are confirmed to sit in the right category, accuracy is 94%. It is a measure of data quality in the books, not of how much was spent.
Accuracy matters because categorisation is the foundation every financial report is built on. Budget-versus-actual comparisons, departmental spend analysis, tax deductibility, and forecasting all assume transactions are in the right buckets. A travel charge miscoded as software, or a one-off capital purchase coded as recurring opex, quietly distorts the picture and sends teams chasing variances that are really just coding errors.
Miscategorisation is rarely random. It clusters around ambiguous merchants, new vendors with no coding history, employee-submitted expenses with thin descriptions, and rules that auto-assign a default category when they cannot decide. Because the totals still reconcile, the errors hide in plain sight until someone audits a category and finds it full of things that do not belong.
Accuracy is not the same as completeness
A transaction can be fully recorded and reconciled yet still sit in the wrong category. Expense categorisation accuracy measures whether the code is right, not whether the transaction exists. Judge it by sampling and re-coding against the actual ledger policy, since auto-applied categories that were never reviewed are exactly where errors hide.
How to calculate expense categorisation accuracy
Categorisation accuracy is the number of correctly coded transactions divided by the total reviewed, expressed as a percentage. The accuracy of the figure depends entirely on how rigorously you confirm what correct means, since a transaction is only right relative to your coding policy.
For example, if a finance reviewer samples 500 transactions for the month and finds 465 coded correctly, accuracy is 93%. Sampling is usually more practical than checking every line. Stratify the sample so ambiguous merchants and employee-submitted expenses are over-represented, because that is where the errors concentrate and a flat random sample understates them.
- 1
Define the correct category per policy
Document the coding rules so reviewers judge transactions against a consistent standard rather than personal interpretation.
- 2
Select a representative sample
Stratify by source and merchant so ambiguous and employee-submitted transactions are well covered, not drowned out by easy recurring charges.
- 3
Re-code each sampled transaction
Independently assign the correct category and compare it to the one already on the books, recording each mismatch and its reason.
- 4
Divide correct by reviewed
Divide confirmed-correct transactions by total reviewed, multiply by 100, then break the rate down by source, merchant, and category to find the weak spots.
Expense categorisation accuracy in a metric tree
Accuracy is one number, but the errors behind it come from distinct sources that need different fixes. A metric tree decomposes the headline rate into the points where a category gets assigned, so a dip points to a specific stage rather than to the books in general.
The first branch is data quality at entry: whether the transaction arrives with enough detail, a clean merchant name, a description, a receipt, to be coded confidently. The second is rule coverage: whether automated coding rules and merchant mappings exist for the spend you actually have. The third is the human coding step: whether employees and approvers select the right category when rules cannot. The fourth is review and correction: whether a check catches errors before they reach the reports.
Metric tree insight
When accuracy slips, the tree separates a data problem from a rules problem from a people problem. KPI Tree puts RACI ownership on each branch, so merchant enrichment sits with the systems team and the coding-policy branch sits with finance. When the rate moves, the accountable owner for the branch behind it is notified, and the verified impact loop checks whether a new mapping rule or a policy change actually lifted accuracy rather than just shifting the errors elsewhere.
Expense categorisation accuracy benchmarks
Accuracy depends heavily on how transactions are coded. Fully manual coding from thin expense reports drifts low, while well-mapped automated coding on enriched card data can run very high. The ranges below give a practical bar by coding method.
| Coding method | Typical accuracy | Strong accuracy | Main error driver |
|---|---|---|---|
| Manual employee coding | 80 to 90% | 92% or above | Thin descriptions and guesswork |
| Rule-based auto-coding | 88 to 95% | 96% or above | Default-category fallthrough |
| Enriched card data with rules | 93 to 98% | 98% or above | New or ambiguous merchants |
| Reviewed and corrected | 96 to 99% | 99% or above | Late or skipped review |
How to improve expense categorisation accuracy
Improving accuracy means reducing the errors at their source, starting with the branch that produces the most reclassifications. Fix the upstream cause, a missing mapping or a noisy merchant name, rather than correcting the same transactions by hand every month.
Enrich transactions at entry
Clean merchant names and attach receipt and metadata before coding. A transaction with a clear merchant and description is far easier to code right, by rule or by person.
Expand merchant mappings
Turn every recurring reclassification into a mapping rule so the same merchant is never miscoded twice, and shrink the share of spend that falls to a default category.
Audit by sample, not by total
Review a stratified sample each period, weighted toward ambiguous and employee-submitted spend, so errors are caught before they reach the reports.
Guide coding at the point of submission
Prompt employees with the likely category and a short reason field at submission, so the right code is chosen first time rather than corrected later.
Common mistakes when tracking expense categorisation accuracy
- 1
Trusting that reconciled means correct
Reconciliation proves a transaction is recorded, not that it is in the right category. Accuracy needs its own check.
- 2
Sampling randomly across all spend
A flat sample is dominated by easy recurring charges and understates the error rate. Stratify toward the ambiguous transactions.
- 3
Ignoring default-category fallthrough
Auto-rules that assign a catch-all category when unsure look like coverage but are a hidden pool of errors. Track the fallthrough rate.
- 4
Correcting symptoms each month
Hand-reclassifying the same merchants repeatedly wastes effort. Convert recurring corrections into mapping rules so the fix sticks.
Related metrics
Budget utilisation rate
Spend vs allocation accuracy
Financial Metrics
Metric Definition
Budget Utilisation Rate = (Actual Spend / Allocated Budget) x 100
Budget utilisation rate measures the percentage of allocated budget that is actually spent during a given period. It is a core financial planning and analysis (FP&A) metric that reveals whether the organisation is executing its financial plan effectively, whether budgets are set at appropriate levels, and whether spending is aligned with strategic priorities.
Category spend analysis
Financial MetricsMetric Definition
Category spend analysis is the process of grouping organisational expenditure into logical categories such as software, travel, marketing, and professional services, then examining patterns within each group. It transforms raw transaction data into actionable intelligence about where money goes and where savings can be found.
Spend forecast accuracy
Financial MetricsMetric Definition
Spend Forecast Accuracy = (1 - |Actual Spend - Forecasted Spend| / Forecasted Spend) x 100
Spend forecast accuracy measures how closely actual expenditure matches the predicted spend for a given period. It evaluates the quality of the organisation's financial forecasting process and directly affects cash flow planning, budget allocation, and investor confidence in financial guidance.
Receipt compliance rate
Financial MetricsMetric Definition
Receipt Compliance Rate = (Transactions With Valid Receipts / Total Transactions Requiring Receipts) x 100
Receipt compliance rate measures the percentage of expense transactions that have a valid receipt or supporting document attached. It is a fundamental control metric for finance teams, affecting audit readiness, tax recoverability, and the accuracy of expense categorisation.
Metric trees for finance teams
Metric Definition
See where expense categorisation accuracy sits within the wider set of metrics a finance team owns and decomposes.
How to debug a broken metric
Metric Definition
Work through a structured diagnosis when expense categorisation accuracy drifts so you can find the coding step that is going wrong.
Turn coding accuracy into an owned number
Build expense categorisation accuracy as a metric tree in KPI Tree, with data quality, rule coverage, human coding, and review as branches that each have an accountable owner. When accuracy slips, the owner of the responsible branch is notified, and the verified impact loop confirms whether a new rule or policy actually raised the rate.
