Skip to main content
Multi-touch attribution (MTA) splits the credit for a conversion across every marketing touchpoint that contributed to it. Rockerbox calculates credit using all of the touchpoint data available at the time of calculation. Because of this, the credit assigned to a channel is not fixed. It changes whenever the set of data behind a conversion changes. This page walks through a single conversion to show how that happens.
The numbers on this page are for illustration only. Actual results depend on the model being used.

​
No matter how many touchpoints we add, it is still 1 conversion

A conversion only happens once. Adding more touchpoints never creates more conversions or more credit, it only changes how that one conversion is divided. When more channels are known to be involved, each one takes a share, and the credit going to the others has to come down to make room.

​
Example: one conversion, four channels

Imagine a single conversion that was influenced by Google, The Trade Desk (TTD), Facebook, and TikTok. The conversion is the same in every step. Only the data Rockerbox has received changes.
1

Google data arrives first

Google is the only known touchpoint, so it receives 100% of the credit.
2

TTD data is added

TTD now has a claim on the conversion. Credit is redistributed, for example 60% Google and 40% TTD.
3

Facebook data is added

A third touchpoint joins the path. Google and TTD each give up some credit to Facebook. Facebook takes the second largest share, ahead of TTD, which was second until now.
4

TikTok data is added

The fourth touchpoint takes its share, and the other three decrease again.
In this example, each channel’s credit goes down as other channels are added. Google’s 100% was never wrong, it was the correct answer for the data that existed at that moment.

​
Late data changes past results

Attribution is recalculated when new data arrives, even if the conversion happened a long time ago. Suppose a month later you add linear TV data from Tatari. Rockerbox backfills the Tatari touchpoints against the conversions they influenced during that period. Tatari’s credit has to come from somewhere, so it comes out of the credit that Google, TTD, Facebook, and TikTok previously held. The result is that:
  • Credit for those four channels decreases for conversions Tatari touched, even though a month has passed
  • Previously reported numbers for those channels can change after a backfill
  • It is still only 1 conversion, so the credit is divided differently but never increased

​
Any backfill can shift attribution

Adding a new channel like Tatari isn’t the only thing that can move your numbers. Any backfill has the chance to shift conversion data. The longer it has been since a conversion, the less likely it is to be altered. Examples:
  • Backfilling historical data for a channel that is already connected
  • Re-pulling a platform’s data to fill gaps
  • Adding touchpoints that were previously missing for part of a period

​
What affects MTA results

  • Which integrations are connected. A channel with no data cannot receive credit, so its share goes to the channels that do have data.
  • When data arrives. Late or backfilled data changes the credit assigned to every other channel on the same conversions.
  • Data completeness. Gaps in a channel’s data, such as missing days or missing campaigns, understate that channel and overstate the others.
  • Granularity of the data. User-level and event-level touchpoints give more precise credit than aggregated data.
  • Lookback window and attribution settings. These determine which touchpoints are eligible to receive credit for a conversion.
  • Conversion event definition. Which events count as conversions determines which paths are evaluated.