Classification & Matching
Analyze is_order/category classification accuracy and customer matching performance.
Time range
Customer
Showing data from Oct 30 – Nov 29, 2025
Avg. category precision
Average precision across all category classifications.
91.8%
across categories +1.8 pp
vs previous 30 daysAvg. category recall
Average recall across all category classifications.
90.8%
across categories +2.1 pp
vs previous 30 daysIs-order accuracy
Accuracy of distinguishing orders from non-orders.
97.2%
order vs non-order +3.0 pp
vs period startAutomated match rate
Percentage of customer matches done automatically.
94.6%
without manual override +2.5 pp
vs previous 30 daysClassification metrics by category
Which categories are classified accurately? Precision and recall per category.
Is-order accuracy over time
Is the system better at telling order vs non-order?
Accuracy
FP Rate
FN Rate
Customer match methods
How are customers being matched – automatically or manually?
45%
24%
15%
11%
5,422 total matches
Domain-based
2,450 (45.2%)
VAT-based
1,280 (23.6%)
Email-based
820 (15.1%)
Name/LLM suggestion
580 (10.7%)
Manual override
292 (5.4%)
Customer match accuracy by customer
For which customers is matching strong or weak? Top customers by volume.