Marketers often toss around the word “attribution,” but there are many misconceptions about this technology. Firstly there are inadequate attribution models floating around the industry like first-touch, last-touch and digital-media only attribution. These insufficient strategies deliver little to no value. To truly understand the complexity of the purchasing process, it is more beneficial for marketers to deploy multi-touch attribution models that analyze and measure marketing campaigns cross-channel cross-device, and are able to link on and offline activity. This approach takes into account complexities, such as media fragmentation, offline tracking, cookie deletion and the use of multiple devices, that trip up inferior models. For example, the average household has 5.7 Internet-connected devices, which means marketers must be sure they can track and distinguish each member of a household across their devices to measure effectiveness of campaigns.
Secondly, there is another ugly industry secret: not all algorithms are created equal. Advanced algorithms, if informed by accurate data, can better predict which visitor is most likely to convert, according to the path that led her to the website. For example, Convertro’s multi-touch attribution uses a machine learning algorithm that combines present marketing, domain intelligence, regressions and precise behavioral controls to scientifically attribute credit to each marketing touchpoint. Sample data is not enough, marketers need the entire data set because the algorithm for attribution and the results are both very sensitive. The larger the number, the more accurate the results will be, which is why access to historical data – including non-converting paths, is so important.
Additionally, effective attribution algorithms should not rely on third-party cookies, which can lose track of a user if she deletes her cookies before converting. The ability to collect accurate user-level data that recognizes when the same user interacts with a brand – no matter which device or personal login she uses to connect to the Web – means the data is more accurate, and therefore more insightful. Hence, better algorithms lead to better recommendations and more profitable marketing decisions.