Returns Glossary
The essential reference for understanding product returns, AI classification, and e-commerce metrics that matter to Shopify merchants.
A
- AI Classifier
- A machine learning model that automatically categorizes inputs—like return reasons or customer feedback—into predefined groups.
C
- Chargeback
- A dispute filed by a customer with their bank to reverse a payment, often resulting from a failed return or refund experience.
- Confidence Score
- A number between 0 and 1 indicating how certain the AI model is about its classification decision.
- Customer Satisfaction Score
- A metric measuring how satisfied customers are with their purchase and return experience.
F
- False Positive
- When the AI incorrectly assigns a return reason that doesn't match the actual reason.
- False Negative
- When the AI fails to identify the correct return reason and assigns an incorrect one.
- Feature Engineering
- The process of selecting and transforming raw data into the most useful inputs for an AI model.
- False Return Claim
- A return request where the stated reason doesn't match the actual condition or circumstances of the return.
- Fulfillment Error
- Any mistake in the order fulfillment process that results in the customer receiving an incorrect or unsatisfactory order.
I
- Inference
- The process of using a trained AI model to make predictions on new, unlabeled data.
L
- Labeled Data
- Training examples where humans have already assigned the correct category or answer.
M
- Machine Learning Model
- A mathematical system trained on historical data to make predictions or classifications on new data.
- Model Accuracy
- The percentage of classifications the AI model gets correct across all categories.
- Multi-class Classification
- A classification task where each input can be assigned one of three or more possible categories.
N
- Natural Language Processing
- A branch of AI that helps computers understand, interpret, and generate human language.
- Net Promoter Score
- A customer loyalty metric based on likelihood to recommend, correlated with return behavior and churn.
O
- Order Accuracy
- The percentage of orders shipped correctly with the right items, quantities, and condition.
P
- Product Defect
- A manufacturing flaw, material failure, or quality issue that prevents a product from functioning as intended.
- Predictive Analytics
- Statistical techniques using historical data to forecast future outcomes like return likelihood.
R
- Return Reason Classification
- The process of categorizing why customers return products using AI to analyze descriptions, images, or structured data.
- Return Rate
- The percentage of orders that customers return within a given period, a key metric for product and operational health.
- Return Reason Code
- A standardized label assigned to each return describing why the customer sent the item back.
- Return Policy
- The documented rules governing whether and how customers can return products for refunds or exchanges.
- Refund Rate
- The percentage of revenue returned to customers through refunds within a given period.
- Restocking Fee
- A charge deducted from customer refunds to cover the cost of processing and inspecting returned items.
- Return Fraud
- Deceptive return practices where customers exploit return policies for financial gain without legitimate cause.
S
- Sentiment Analysis
- AI that determines whether customer text expresses positive, negative, or neutral emotions.
- SKU Mismatch
- A fulfillment error where the product shipped doesn't match the SKU ordered, resulting in customer returns.
- Size/Fit Issue
- Returns where customers cite incorrect sizing, poor fit, or items not matching size expectations.
T
- Training Data
- The labeled examples used to teach an AI model how to categorize returns correctly.
- Threshold
- The minimum confidence score required for the AI to auto-assign a classification versus flagging for human review.
W
- Wrong Item Shipped
- An order fulfillment error where customers receive a different product than what they ordered.