The Role of Creative in Measuring Marketing Effectiveness

Multi-touch attribution relies on the ability to track each exposure, interaction and sale back to an individual. Today, a lot of that measurement happens using third-party cookies and app tracking. Therefore, the depreciation of cookies is likely to lead to a shift in investment from MTA to MMM. This shift is a good time to reevaluate the approach to both methodologies and consider an important piece of the puzzle that has long been ignored: creative.

Different But Similarly Blind to Creative

MTA seeks to understand the consumer path to conversion by tracking every exposure of an individual across various channels. Tracking exposure across web pages and apps relies on cookies and cross-app tracking (at least historically). Credit is then assigned to each of those touch points, the magnitude of which is determined by the attribution model used. Depending on the framework used, some touch points may receive more credit and others less, most often depending on the sequence of exposures.

MMM primarily leverages impression streams across publishers and channels to determine effectiveness against metrics like return on ad spend. This is done using multi-linear regressions to determine which variables have a relationship with a particular metric and to what degree. Unlike MTA, MMM does not rely on tracking audience exposures, and does not seek to understand the ideal consumer path but instead helps identify the appropriate media mix.

As a result of signal loss, advertisers are expected to invest more heavily in MMM measurement. While MMM has the benefit of not relying on cookie data, it comes with its own set of challenges, namely that the model is not complete. Historically, creative has been an unaccounted variable in the models, despite it being the most important input, accounting for 70% of performance.

Instead, MMM tends to rely most heavily on media variables such as spend, channel, target, etc. and at times, consider outside factors such as seasonality, macroeconomic conditions or weather. The lack of creative variables can lead to the misattribution of performance to other factors, possibly misleading budget allocation and media mix decisions.

Fit Adherence Scoring Variables Into Existing MMM

As Google looks to follow Apple’s lead and also restrict cross-app tracking on Android in the next two years, marketers are likely to be further pressured to adopt different solutions to mitigate signal loss. Cookie depreciation diminishes the effectiveness of hyper-targeting and further stresses the importance of measuring creative influence on performance. In response, introducing creative variables into sales/ROI modeling frameworks is likely to become more common. The challenge will be to understand which creative elements are most important to the marketing model, from key messages to emotional levers, sound, and imagery.

Third-party cookies and app tracking will make MTA harder, but there are ways to improve upon MMM to include the influence of the creative on performance using a designated creative variable.

However, including too many creative variables with too much granularity cannot be accommodated with multi-linear regressions. Instead, it’s best to score assets against particular adherence criteria (roughly four to five, max) and introduce them as variables to the model. These variables should be distinct, easy to measure, and informed by insights as to what works and what doesn’t work. In most cases, these criteria may align with platform best practices, however, as more analytics is carried out on an advertiser’s creative, there is an opportunity to tailor those criteria to brand-specific best practices. Introducing creative variables via scoring is expected to help provide better readouts of effectiveness, without crediting performance to the wrong variables.

Read about how data can drive your next digital marketing strategy, like going from Global to Local.