Data professionals everywhere are scrambling to find a sustainable and effective successor to the third-party cookie. Here, for our Deep Dive on Data, Bobby Gray and Aleksandra Semenenko of Drum Network member agency Artefact tell us to look no further than causation-based approaches.
Online advertising introduced many advantages, but also made the equation more complicated, a situation exacerbated by the development of so many new digital channels (social, VOD, digital radio, etc). The rise of influencers and online communities makes understanding the true impact of marketing more nuanced than ever.
It’s increasingly complicated to decipher customer behavior and the behavioral effects of marketing. Marketers still lack a ‘single source of truth’ to which everyone is aligned, a challenge heightened by the impending loss of the third-party cookie, which is integral to mapping the individual customer journey, to inform strategies and measure full marketing funnel performance.
While the impending death of the cookie is understandably a pain point for many organizations, it also has the potential to pave the way for next-generation marketing.
Understanding marketing return-on-investment
Data options abound for understanding marketing return-on-investment (MROI) in post-cookie world.
Correlation and causality-based approaches will be the most robust options. The former identifies a relationship between marketing variables and sales. Generating insight in this way is a lengthy process, making it best suited for monthly or quarterly budget planning. Causality-based approaches, by contrast, determine the individual marketing variables that have a direct impact on business outcomes (or the success of other marketing variables) and allow for weekly or even daily optimization.
Correlation analysis, such as Marketing Mix Modelling (MMM), has been used for a long time. It’s not reliant on customer data but provides a good view of what is and isn’t working. However, it infers that any fluctuations in marketing delivery are directly correlated (either positively or negatively) with changes to success factors such as sales. This makes it tricky to provide clear reasons for performance uplifts and can lead to reliance on statistically-relevant trends that can result in erroneous decisions influenced by personal biases.
For example, a retailer might increase paid social spend and run a 25% onsite discount over a 30-day period. Correlation-based analysis would draw the conclusion that the resulting surge in online sales was a result of both marketing activities. It is unlikely to accurately determine the precise impact of each, or the inter-relationship between them.
Causality, which is truly omni-channel, identifies an event as being the direct consequence of another; it provides the ‘why’ and ‘so what’ for outcomes of marketing activity/ Understanding this is at the root of optimizing future activity.
Causality-based approaches show that A (marketing budget change) caused B (sales increase). The proportion of the uplift can be directly attributed to each channel. This is critical for making budget decisions on the marketing mix. Causality-based approaches allow marketers to understand the influence that one marketing activity has on another, and how this feeds into outcomes.
So a causality-based approach might determine that 35% of the increase in sales was down to paid social spend, while 65% was attributable to the 25% discount promotion. This method could also show the relationship between the two separate activities, and how this affects sales, proving which optimizations are the most important.
MMM studies were never intended for near real-time insights. Causality-based approaches, however, use real-time demand signals to show the impact that every potential major business driver (media, competition, price, brand demand signals) has on key business outcomes (both digital and offline) at any given moment. They match the decisions that businesses need to make with the measurements required to support those decisions.
Cookie data shows an individual consumer’s journey across various platforms. They might start with Facebook, move on to YouTube and then use a search engine before making their eventual purchase. Correlation, while it shows where they have visited, does not show the sequence – a major disadvantage in view of the different roles played by each channel at each stage of the buying decision process. But causality enables the customer path to be re-built from a macro level – allowing a brand to see not just that advertising on Facebook works, but its interrelationships with other channels in the marketing mix (the impact on YouTube or PPC for example) and their influence in driving results.
The bottom line
MROI recognizes that marketing measurement must be more than mere justification of advertising spend after the event. Based on causation, MROI provides in-depth insight to enable each level of the decision-making process to be optimized so that it underpins everything, from analyzing weekly campaign performance, through quarterly planning, to annual budget allocation.
MROI modeling is individual to each organization, depending on their specific business objectives, budgets, data and business drivers. The bottom line is that it enables smarter decision-making.