One customer discovers a product on TikTok, returns through Google, sees a Meta ad, and buys. Several platforms can claim that purchase. Adding their totals does not tell you how many orders the business received.
Start with the order record and a written purchase definition. Then compare each platform’s receipt and attribution rules with that shared commercial reference.
Define the order before comparing channels
Before comparing channels, decide what a purchase means. Use one order identity, one currency rule, one treatment of discounts, and one policy for refunds and cancellations. Apply the same discipline to the steps that precede the order, such as product views, add-to-cart events, and checkout starts.
Names alone are not enough. Two systems can both report purchase while disagreeing about value, time, or status. A shared event language includes the properties and rules that make an event comparable.
Preserve the journey without forcing certainty
First-party identifiers can help connect interactions when the customer has provided them and the use is permitted. Source, campaign, timestamp, and order context should remain attached as the journey develops. When identity is uncertain, the record should say so.
A stitched timeline is evidence, not permission to claim that every touchpoint caused the sale. The goal is to preserve what happened well enough to compare explanations, including the possibility that more than one channel contributed.
Use attribution as a model, not a verdict
In the TikTok, Google, and Meta journey, last-click reporting may favor search while a view-through model favors an earlier ad. An independent customer timeline does not make those models disappear. It makes their assumptions easier to inspect.
Teams can then ask better questions. Did discovery precede branded search? Did retargeting appear before checkout? Does the order reconcile with the platform event? These questions are more useful than asking which dashboard is the single winner.
Bily organizes the customer journey around consistent events so channel reports can be compared with the same commercial outcome. The result is not perfect certainty. It is a clearer account of what is known, what is modeled, and where the data disagrees.
Let evidence guide the budget
For the next channel review, record one order rule, reporting period, time zone, value rule, and treatment of refunds. Classify differences as scope, timing, delivery, or attribution before using them to change spend.
An unexplained difference is a question to investigate. Keep it visible until the evidence supports a specific action.
