When your channel reports add up to more than you sold, nothing is broken. Nothing is deduplicating either.
Klaviyo claims $412,000. Meta claims $530,000. Postscript claims $96,000. Google claims $284,000. Add them up and your channels drove $1,322,000 of a $1,850,000 month, which sounds plausible until you add organic and direct, and the total passes what you actually sold.
Nothing is malfunctioning. Every platform is measuring its own involvement in the orders it can see, and there is no layer anywhere in your stack whose job is to stop the same order being claimed four times.
Each platform runs its own attribution independently, on its own data, using its own window:
| Platform | Default window | Counts an order when |
|---|---|---|
| Meta Ads | 7-day click, 1-day view | An ad was clicked, or merely seen, before the purchase |
| Klaviyo | 5-day click or open | An email was clicked or opened |
| Postscript | Typically 1 to 7 days | An SMS was delivered or clicked |
| Google Ads | 30-day click | An ad was clicked |
| GA4 | Session-based | Assigns to the last non-direct click only |
Note the last row. GA4 is the only one that enforces exclusivity, and it does so by picking a single winner, which is why GA4's channel numbers are always lower than what the platforms self-report.
Meta cannot see that Klaviyo also claimed the order. Klaviyo cannot see Meta. Neither can see Postscript. They are not in communication and were never designed to be. Each answers "did my channel touch this buyer inside my window," and for a customer who saw an ad, got an email and received a text in the same week, all three answers are yes.
Crestline's month, self-reported figures against GA4's exclusive last-click view of the identical orders.
| Channel | Self-reported | GA4 last-click | Difference |
|---|---|---|---|
| Meta Ads | $530,000 | $486,000 | $44,000 |
| Klaviyo email | $412,000 | $214,000 | $198,000 |
| Postscript SMS | $96,000 | $61,000 | $35,000 |
| Google Ads | $284,000 | $243,000 | $41,000 |
| Total claimed | $1,322,000 | $1,004,000 | $318,000 |
Crestline Co. is a simulated DTC brand used for demonstration. These figures are illustrative and are not a client result.
The platforms collectively claim $318,000 more than a single-winner model assigns them. That $318,000 is the overlap: orders touched by two or more channels and counted by each.
Klaviyo accounts for $198,000 of it, well over half. That is the open-attribution problem doing most of the work, and it is why email is almost always the largest source of double counting in a DTC stack.
Now the part that breaks budget meetings. Add GA4's organic search and direct, $498,000 combined, to the $1,322,000 of self-reported channel revenue and you get $1,820,000 against $1,850,000 in net sales. It looks like it reconciles. It does not. Roughly $318,000 of that total is the same orders counted twice, which means something close to $318,000 of genuinely organic and direct revenue is being credited to paid and owned channels.
The rule that prevents most of the damage. Never sum attributed revenue across platforms. The total is not a real quantity. Compare each channel against its own history instead, and when you need a number for the whole business, use net sales.
Pick one source of truth and accept it is imperfect. Use GA4's last-click for all cross-channel comparison. It undercounts genuine multi-touch influence, but it is internally consistent, and consistency is what you need to compare channels. Cheap and immediate.
Tighten every window toward click-based attribution. Switch Klaviyo to click-only. Drop Meta's 1-day view. This does not eliminate overlap but it removes the most spurious portion, the credit claimed on a view or an open where the customer did nothing. This is the highest return for the effort involved.
Measure incrementality on the channels that matter. Holdout tests are the only method that gives you a defensible answer, because they measure what happens when a channel is absent rather than who was standing nearby. Slower, and the right tool for the two or three channels carrying most of your spend.
The overlap itself is unavoidable. What is avoidable is budget decisions made as though it does not exist. Postlytix reads every channel's attribution against actual order data at the same time, sizes where the same revenue is being claimed more than once, and shows what each channel contributes once the duplication is removed.
This is one of eight places the numbers come apart. Why your e-commerce revenue numbers never match explains why fragmentation happens across the whole stack, and links to the rest.