Analysis for people buying traffic.

Add up what every platform claims and compare it against the orders you actually took. That single figure is the size of the double counting and is usually larger than people expect.

It is also the number that makes the rest of the analysis legible, because every channel-level figure is inflated by some share of it.

Match platform conversions against order records and credit each channel with what it can be tied to, counting the unmatched claims separately rather than discarding them.

Without a channel field on the orders, matching falls back to time and value, which is weaker. Say so rather than presenting the same confidence on worse evidence.

Revenue per channel flatters whichever one sells discounted, high-return products, and that is frequently the channel with the best-looking return.

Where product costs exist, set spend against contribution instead. That is the number that decides whether a campaign is worth running, and it occasionally reverses the ranking entirely.

Your own order data covering the period, since that is the table the figures are computed from: one table, whichever is largest. Column names differ by platform and by year, so they are matched by content rather than by an expected header.

The platform reports and any product cost sheet can go up alongside. Each is read, converted and handed back as a download, and the report names any relationship it can see, but no figure is computed across two files, so setting spend against contribution is a comparison you make from the two outputs.