Convertro’s algorithmic attribution model accounts for past purchases by modeling them in a regression analysis and determining the appropriate conversion revenue to allocate to the marketing that drove a user’s prior conversions. This reflects the idea that it’s harder to earn a new customer than to keep an existing one. Marketing sources that initially acquired customers for you receive a greater share of credit compared to those that serve to bring back already acquired customers.
The model also analyzes historical conversion data to understand when marketing sources are (or are not) helping push users through stages of a purchase funnel. Credit for marketing sources that occur in between two stages of a conversion funnel (“cart” and “sale” for example) is determined based on how much the marketing contributes to the customer’s progress to the next funnel stage. For example, if “cart” almost always leads to “sale” even without the presence of additional marketing exposures in between the two events, the marketing between those two steps will get less credit.