Business Case: Attribution 2.0
Online marketing: Allocating online budgets optimally via Attribution 2.0
There are lots of possibilities when it comes to creating a budget – even in the world of online marketing. The question often arises whether the chosen online marketing channel (SEO, affiliate, SEM, etc.) produces the desired result, such as more clicks, more subscriptions, or more sales. No one wants to throw money down the drain. An attribution model is a statistical way to reliably assign an outcome to one of many possible causes. With Attribution 2.0 we have developed a model that lets companies allocate their marketing budgets optimally – that is, based on causality – to the channels that are proven to perform well. This way, marketing managers always know how to best spend their budget.
Creating a statistical attribution model requires so-called customer journey data. This is detailed data about customers’ different contact points when visiting a website, and includes customer activity starting from where they enter the site all the way to their completed purchases. Ideally, this data is enriched with off-site contacts (customer contact points outside of your own website, such as banner ads that didn’t get clicked on) and with shopping cart, CRM and other external data. The first step consists of selecting relevant attributes. The amount of data is often huge, and there is a lot of background noise in the bits and bytes. This means that really relevant data is few and far between. Considerable experience is therefore needed to pick out the most relevant data. It is not uncommon for tracking artifacts to distort the data. Such problems must be identified and resolved before beginning the actual analysis.
After the data has been properly prepared and validated, our Attribution 2.0 statistical model is now put into action. The model has a challenging task: It aims to understand the purchase process and identify the so-called conversion drivers. The model employs a data-driven approach. This involves comparing unsuccessful and successful customer journeys in order to identify which factors help increase conversion (leading, for example, to a completed purchase or a newsletter subscription) and which factors hinder conversion. Unlike well-known heuristics such as Single Source Attribution or Fractional Attribution, this procedure is statistically sound and fair. Both are important because the attribution model determines how large marketing budgets are regularly allocated. Such decisions should be based on facts and not gut feelings. The procedure also isn’t a black box. In fact, quite the opposite is true: Many of the rules identified by Attribution 2.0 are consistent with the experience of marketing experts. Our model has repeatedly earned the confidence of marketing managers, because its validity can be objectively measured and quantified.
The last step involves automating the attribution modeling and implementing it in the company’s or the tracking provider’s system. In addition to its use for online channels, Attribution 2.0 can also integrate offline contacts compiled from traditional media such as TV and catalog advertising. Overall, Attribution 2.0 enables companies to allocate their budget in the most effective way, to achieve cost savings by identifying unprofitable channels, and to assess new partners or campaigns faster. The attribution model is also flexible and can be further refined at any time, and thus grows as business requirements and needs increase.