The term attribution originated in psychology where it describes the connection between causes and the actions and behaviors that are triggered by them. It found its way into modern marketing at a time when advertisers were shifting their budgets from traditional offline ads to digital media. This made huge amounts of customer data available to them via channels such as paid and organic search, as well as display and email marketing.
Based on this development, marketing attribution is now understood to be a process of analysis through which solid results can be achieved using consistently reliable measures. For this purpose, every available channel and advertising media is considered to bring a potential customer into contact with a product or a service, increase their interest and then convert them into a buyer. These touchpoints can be, for example, social media campaigns, SEO efforts or supporting existing customers who share their product experiences on review portals.
Successful attribution for an optimal marketing mix
Various methods of marketing attribution can now be used to quantify how successfully the respective touchpoints drive conversion. When using this method, the customer’s interactions with advertising measures are not considered just on their own, but rather within their chronological sequence and interaction so as to identify possible correlations and determine the respective shares in the overall conversion rate. The aim is to use any insights gained to optimize the marketing budget for current and future advertising campaigns.
Which attribution model to choose?
There are a variety of different attribution models that can be used depending on the target group and the measures employed. Single-source attribution methods, for example, credit individual touchpoints such as initial contact with the product or the last click before purchase with the total conversion performance. This approach is easily implemented because it leaves out the complexity of a decision-making process.
In contrast to single-source attribution, fractional attribution methods assign a contribution to conversion to all existing touchpoints. Examples of this are:
- Position-Based Attribution that emphasizes the first and last touchpoints in the evaluation over all others;
- Time-Decay Attribution that values more recent interactions higher than those at the beginning of the decision-making process;
- Multi-Touch Attribution (MTA), the most data-centric and sophisticated fractional attribution model, that assigns variable values to all touchpoints based on intensive web analytics.
AI-supported marketing mix as an alternative
The problem with MTA is that its dependence on Big Data is one of its greatest weaknesses – for it to arrive at meaningful results, MTA needs both extensive and the most precise customer data possible. One of the most important sources of this is tracking cookies that record a website visitor’s interactions even beyond the domain visited to generate a detailed user profile. However, tracking has become less and less relevant in recent years.
Why does tracking cause such problems?
One of the main problems is that tracking cookies can only provide data for operations on the web – advertising efforts that continue to spread via print and TV or sales in bricks-and-mortar stores cannot be tracked.
The same is true for closed ecosystems that only allow proprietary tools to measure and evaluate marketing efforts and are blind spots for web analytics from outside of these “walled gardens”. But the strongest impact on data-centric attribution models has come from laws such as the European General Data Protection Regulation or the California Consumers Privacy Act that have placed tight limits on the intensive collection of customer behavior by the use of strict opt-in requirements and cookie banners.
The resulting inaccuracy in the evaluation of touchpoints leads to ineffective allocation of the marketing budget – which means that MTA cannot fulfill its own precise requirements. Alternative methods are available that rely instead on holistic AI-supported monitoring of advertising measures and their results and do not require attribution. In particular, decision intelligence-based methods such as the paretos Omnichannel Optimization offer privacy-compliant and easy-to-use tools for improving the marketing mix.