Start with one order
A channel report starts with claimed revenue. A customer journey starts with the order and works backwards through observable visits. The same order can contain a Meta visit, a Google click and an email click, but its revenue still exists only once.
Check the order, lookback window and session links first. Keep orders without a recognized session separate and do not silently relabel an unknown source as direct.
Read each touchpoint literally
A UTM tag, referrer or click ID describes what was recorded for a visit. It does not independently prove causation. Keep clicks, impressions, email opens and direct visits distinct, and preserve missing source data as unknown.
- Check the time, source and available campaign data for each visit.
- Do not treat clicks, impressions, email opens and direct visits as the same event.
- Keep missing source information labeled as unknown.
- Use the same session and lookback definition for every order.
Describe channel roles
The sequence can show channels that appear early, bring a buyer back or sit directly before the order. Describe these as observed roles. A prior introduction through word of mouth, a podcast or an unrecognized session may be absent.
Phrase the conclusion carefully. Say that a channel often provides the first measured contact, not that it created all demand. Split new and returning customers where the available order history supports that distinction.
Compare attribution models
First touch credits the first eligible contact. Last touch credits the last. A linear model divides credit across eligible touchpoints. Compare them on the same orders, period and rules, then add cost, new-customer status and margin before changing budgets.
- Identify which channel often provides the first measured session.
- Identify which channel often appears directly before the order.
- Compare how revenue credit shifts when the model changes.
- Review cost, new customers and margin before moving budget.
Turn the timeline into action
Use the journey to choose a focused question. If paid traffic appears early and branded search closes, investigate whether acquisition creates demand that search later captures. Stronger claims about incremental sales require experimental evidence.