E-commerce Analytics Dashboard Planning Guide

An e-commerce dashboard should explain where money is earned and where operational friction reduces it. Store revenue, advertising revenue and accounting revenue are different views of the business; combining them requires explicit definitions.

Begin With Orders And Adjustments

Choose an order identifier that survives exports and integrations. Define gross product sales, discounts, shipping, tax, refunds and net sales separately. Decide whether refunds are reported on their processing date, against the original order cohort, or in both views with clear labels.

For example, a store may ship 100 orders this month while receiving refunds for last month's promotion. A current cash view includes those refunds now. A promotion profitability view assigns them back to that campaign cohort. Neither should quietly replace the other.

Join Product And Channel Detail Carefully

Use stable variant identifiers for product analysis. Titles change, and one product can have sizes with different availability or return rates. Record acquisition source separately from the platform's attributed conversion totals. Adding attributed revenue across ad platforms can count the same order more than once.

Include stockouts, cancellation reasons and return rates beside demand. Falling sales for a bestseller may reflect unavailable inventory rather than weak marketing. Show data freshness for fulfillment and inventory feeds.

Reconcile Before Automating Decisions

Select ordinary orders, discounted baskets, partial refunds and multiple-currency purchases. Confirm how each flows into totals. Agree the exchange-rate policy with the reporting owner rather than mixing converted figures from different systems.

Build The Dashboard Around Merchandising Decisions

Start with decisions that repeat in the store: which variants to replenish, which promotions to repeat, where checkout needs investigation and which products generate disproportionate returns. Give each decision a supporting view and a responsible user. A merchandising manager and a customer-support lead may need different detail from the same orders.

Use a small summary with drill-down into products, channels and exceptions. Show the reporting period, currency and included order states. Keep cancelled test orders out of commercial totals through a documented rule, not an invisible manual adjustment that only one analyst remembers.

Write a metric dictionary before designing charts. For average order value, specify the revenue basis and which orders form the denominator. For conversion rate, identify whether the denominator is sessions, users or a defined checkout cohort. Two accurate numbers can disagree when their definitions differ.

Work Through A Discounted Order And Partial Refund

Consider an illustrative order containing two items priced at 40 and 60. A ten-unit basket discount reduces product sales from 100 to 90. The customer also pays ten for shipping, and tax is recorded separately according to the store's configuration. Keep those components distinct rather than labeling the entire amount product revenue.

The second item is later returned. If the discount was allocated proportionally, its discounted product value was 54 and the first item's was 36. The refund workflow must preserve that allocation or the reporting must use the authoritative refund amounts. Do not assume the refund equals the original undiscounted price of 60.

Suppose the refund occurs in the following month. The cash-adjustment view shows the refund when processed; the original-order cohort view can show that the sale eventually retained only 36 in product revenue. Both views are useful, but they should have different labels and should not be compared as though they describe the same period.

For GA4 implementation, Google's e-commerce measurement documentation describes purchase and refund events, including transaction and item identifiers. Map analytics events to the actual order and refund records, then verify the values using examples like this one. An event appearing in a debugger does not prove that its financial meaning is correct.

Keep Analytics And Order Records Reconciled

Use the store or order-management system to establish the operational order population. Analytics may observe only part of that population because collection depends on the implementation and visitor environment. Diagnose differences instead of expecting perfect agreement by default.

Build a reconciliation report with order identifiers, timestamps, amounts and event status. Investigate duplicate purchase events, missing transaction identifiers and refunds that never reached analytics. Avoid sending personal information simply to make reconciliation easier; stable non-personal identifiers usually provide the necessary connection.

The GA4 revenue-validation guide is a useful companion for event-level testing. Keep the dashboard planning document focused on which source owns each measure, what differences are tolerable and who investigates exceptions.

Add Product Economics And Availability

Revenue alone can reward products that are costly to fulfill or frequently returned. Include contribution measures only when the relevant costs are available and consistently defined. State whether the figure includes product cost, shipping subsidy, payment fees, advertising or return handling. Do not label a partial calculation net profit.

Analyze variants where size, color or material affects demand and returns. A strong parent product can hide one variant that is repeatedly unavailable and another that rarely sells. Use stable identifiers so a title cleanup does not break the historical series.

Show availability alongside sales. A product with falling units may have spent half the period out of stock. A promotion could look unsuccessful because the advertised size was unavailable, not because the audience disliked the offer. Record stockout periods if replenishment decisions depend on them.

For bundles, decide whether reporting follows the sold bundle, its components or both. Avoid adding bundle revenue and component allocations together as separate sales. The operations and merchandising views may need different groupings, but they must reconcile to the same order ledger.

Compare Channels With Explicit Attribution Limits

Store the order's observed acquisition context where it can be collected appropriately, but keep it separate from each advertising platform's attributed result. Several platforms may claim influence over one sale. Adding their reported revenue can exceed the store's actual revenue without any individual report being arithmetically broken.

Show spend, observed orders and platform-attributed conversions as distinct measures. Label windows and reporting settings. Use longer-term customer or cohort analysis where repeat purchases matter, and avoid treating a first-click or last-click view as a complete description of influence.

When comparing channels, inspect product mix, discounts and returns as well as cost per order. A channel producing inexpensive orders may be acquiring low-value baskets with high return costs. The decision to scale should reflect the store's operating constraints and margin assumptions, not just the lowest headline acquisition cost.

Make Freshness And Permissions Visible

Different sources update at different intervals. Order status may be current while advertising cost or refunds arrive later. Show the last successful refresh for material sources and label recent periods as provisional when appropriate. A green refresh icon for one connector should not imply that every metric is equally current.

Restrict customer-level detail to users who need it. A product manager may need aggregate return reasons without names and addresses. Exports and scheduled reports require the same review as the dashboard itself, because downloaded files can outlive the access permissions in the application.

Set an owner for broken feeds and unmapped products. Create an exception view that shows stale data, unknown identifiers and reconciliation gaps. The dashboard should help staff fix these problems rather than quietly hiding the affected records to keep the charts tidy.

Pilot A Small Trading Period

Choose a representative period containing ordinary orders, discounts, a partial refund, a cancellation and a stockout. Reconcile totals manually with the store's records. Ask merchandising and support users to answer real questions using the dashboard, then note where they need to leave it to investigate.

Use the pilot to remove unnecessary charts and add missing context. A table showing products with rising return rates and their reasons may be more useful than another sales trend. Prioritize what changes a decision before widening the reporting scope.

Conversion Rate Optimization (CRO) is relevant when analytics and shopping behavior point to funnel problems that need diagnosis. It should not be described as an included custom dashboard build without an agreed scope. The dashboard governance guide can help establish the longer-term reporting responsibilities.

Frequently Asked Questions

Which system should provide the revenue total?

Choose the source that matches the question. The order system can establish sales and refunds; accounting can establish financial reporting; analytics can describe observed marketing behavior. Define the differences rather than forcing one number to serve every purpose.

Should refunds reduce the original month or the current month?

Both views can be useful. A processed-refund view shows current activity, while a cohort view explains the eventual outcome of earlier sales. Label them clearly and follow the business's reporting policy for official financial statements.

How much detail belongs on the first screen?

Show enough to identify the main decision and unusual changes, then provide drill-down. Keep definitions, filters and freshness visible. A summary is useful when users can inspect the records behind it without losing the context of the question.