How to audit an e-commerce business

1. Define the decision and baseline

“Increase sales” is too broad to guide an audit. Specify the customer segment, channel and time period. Choose a primary business outcome and record the factors that could distort the comparison, such as a promotion, assortment change or stock shortage.

• Separate placed, paid and fulfilled orders.
• Compare similar periods and customer groups.
• Record definitions and data owners for each metric.

2. Reconcile order economics

Reconcile recognized sales with product cost, discounts, payment charges, fulfilment, return costs and acquisition. State which costs are included so that teams use the same definition of contribution.

• Calculate contribution by channel and meaningful customer segment.
• Check whether discounts or delivery subsidies explain a conversion increase.
• Distinguish variable order costs from fixed overhead.

3. Find friction in the buying journey

Review product discovery, search, availability, product information, basket, delivery selection and payment. Segment the funnel by device, new or returning customers and acquisition source before interpreting an average conversion rate. Use support contacts and observed customer behavior to explain the numbers.

• Test representative journeys on a phone.
• Check when the customer first sees the full cost and delivery promise.
• Validate event tracking against actual orders.

4. Follow the order through fulfilment

A stock discrepancy, a late delivery or a difficult return can erase the benefit of a smoother website. Examine cancellations and returns by cause and connect operational issues to the customer promise that preceded them.

• Compare promised and actual service levels.
• Assign owners to the handoffs between digital sales, stores and support.
• Include the cost of failed and returned orders.

5. Examine repeat purchases by cohort

A single repeat-purchase rate can mix customers who have had very different opportunities to return. Compare cohorts with equal observation windows and consider the natural replenishment cycle of the category. In an omnichannel business, check whether a customer who disappears online is still buying in stores.

• Use a consistent customer identity where available and permitted.
• Separate organic repeat behavior from promotion-driven reactivation.
• Track repeat contribution as well as repeat revenue.

6. Turn findings into a testable backlog

Each recommendation needs an observed problem, an explanation, an expected mechanism and a way to check the result. Record confidence and dependencies. A missing measurement may deserve priority over a visually impressive redesign when it prevents the team from judging any change.

• Name an owner, a metric and a review point for each initiative.
• Define the margin, service or quality limits a test must respect.
• Keep untested ideas labeled as hypotheses.

Related experience and research

Apply this to your business

A consultation can help turn these questions into a scope, a data request and a sequence of decisions for your team.