Multi-Touch Revenue Attribution Automation: Which Touch Actually Closed the Deal

A deal closes, and everyone has an opinion about what caused it — the ad that first caught the contact's attention, or the email that arrived right before they said yes. Without a system that actually tracks each touch and applies a consistent rule for splitting credit, "which channel worked" stays a guess dressed up as an observation, and budget ends up allocated by whoever argues most convincingly rather than by what the contact's own history actually shows.

A single source field captures one moment in a relationship that can span weeks or months of separate touches — it structurally cannot represent the fact that a contact was reached by more than one campaign before converting. Genuine multi-touch attribution requires actually recording every touch as its own event, tied to a timestamp and a channel, and then applying an explicit, consistent rule for how credit splits across however many touches actually happened — a rule that has to handle one touch, two touches, and three-or-more touches differently, because a straight even split stops making sense once there is a touch that is clearly closest to the moment of conversion.

The moment an opportunity is marked won with a linked contact, Flow's attribution engine looks back 30 days across that contact's recorded campaign touches. Under the first-click model, the earliest touch in that window gets 100% of the deal's value. Under last-click, the touch closest to the win gets 100%. Under the multi-touch model, the split depends on how many touches actually happened: a single touch gets 100%, two touches split 50/50, and three or more touches split 30% to the first, 40% to whichever touch is closest to the actual conversion event, and 30% to the last — with the weights automatically rebalanced if that middle touch happens to land at either end of the sequence, so the percentages always add up to 100%.

All three models are computed and stored automatically, every time — nobody has to pick a model in advance or re-run anything after the fact. On the attribution dashboard, a team switches between first-click, last-click and the multi-touch split with one control, and sees the resulting revenue broken down by channel, computed live from the same underlying records — never a separately-run report that can drift out of sync with what actually closed.

A contact clicks a social ad in January, opens a nurture email in February, and signs a $10,000 deal in March after a final follow-up call logged as a touch. Under first-click, the January ad gets full credit for the $10,000. Under last-click, the March follow-up gets it instead. Under the multi-touch model, the January ad gets $3,000, the February email — the touch closest to the deal actually converting — gets $4,000, and the March follow-up gets $3,000. A marketing lead reviewing the dashboard can toggle between all three views for the same closed deal without re-running anything, and immediately sees that no single model tells the whole story on its own.

Attribution becomes a live number every team member can check, not a periodic manual reconciliation between deals and campaign logs. Because all three models compute automatically and simultaneously, a team is never locked into a single view of "what worked" — they can check first-click for top-of-funnel credit and multi-touch for a more balanced picture of the whole journey, from the same underlying data.

A team that wants time-decay or position-based weighting models on this same dashboard will not find them here — the visible attribution view supports first-click, last-click and the 30/40/30 multi-touch split; other model types exist elsewhere in Flow's attribution infrastructure for a separate, per-conversion computation, but are not yet surfaced on any customer-facing report, so this page does not claim them as an available view. A team whose contacts rarely have more than one recorded touch before converting will see the multi-touch and last-click models converge to the same answer most of the time — the split only becomes meaningful once a contact genuinely has multiple recorded touches.

A contact record with one "source" field can only ever represent the most recently recorded touch, structurally erasing everything that happened before it. Centriu Flow instead keeps every recorded touch and computes three separate, internally consistent attribution models the instant a deal closes — so "what worked" has three honest, simultaneously available answers instead of one field that already lost the earlier part of the story.