Marketing measurement has been perplexing advertisers since the dawn of digital. It’s been almost 30 years since the first online ad – a static banner for AT&T that appeared on HotWired.com back in 1994.
While ad creatives, formats and targeting have become far more sophisticated since the early days, measuring an ad’s effectiveness easily and accurately remains somewhat of a pipe dream. Despite a whole host of potential solutions, the industry can’t seem to get it right.
So where are we going wrong? Why does accurate marketing measurement remain so elusive? Let’s evaluate some of the most commonly used tactics.
The lacking last click
The most common way to measure ads before the introduction of privacy legislation was via user-level data and last-touch attribution. Last-click attribution is easy to implement and evaluate, which is why many marketers still use it despite being aware of its flaws.
But the problem is that last-click attribution doesn’t account for today’s complex customer journeys. And, on average, consumers see between 4,000 and 10,000 ads a day across a variety of digital and offline screens – so how can we say which ad led them to a decision?
Marketers know that the chance of an individual seeing an ad for new headphones once and then instantly buying them is unlikely. Yet they’re happy to assign all credit for a transaction to the last click, which completely negates the value of the other ads the consumer may have been served in the lead up to a purchase. It’s a lazy and inadequate means of marketing measurement.
Missing the mark with MTA
Multi-touch attribution (MTA) was initially touted as an improvement to last-click attribution that takes the entire customer journey into account and accurately assigns credit to each touch point. The problem is it doesn’t actually work.
First, MTA doesn’t take offline channels into account. Linear TV ads, billboards and other nondigital ads have no value where MTA is concerned. Moreover, privacy legislation, ad blockers, Apple’s IDFA, the walled gardens’ data stranglehold and the decline of third-party cookies all work to prevent marketers from consolidating data across channels, making MTA impossible. MTA was a great idea, but one that has sadly never lived up to its promise.
Deterministic deceptions
Deterministic attribution is just a fancy term for a form of multi-touch attribution. It assigns an identity to an individual through a personal or device ID and tracks their actions across digital channels to measure the value of various ads served to that consumer. Like MTA, it has no offline capabilities.
Given that it requires an ID to work, deterministic attribution has also been impacted by the deprecation of Apple’s IDFA. And even where that’s not the case, deterministic data is extremely hard to come by because it requires a user to be ‘identifiable’ (i.e., logged in on all devices) for identities to be matched.
Forget fingerprinting
Fingerprinting refers to creating a digital fingerprint for an individual user. While it has been around for a while, it has really only come to prominence with the decline of third-party cookies, as marketers search for alternative ways to identify and track users.
Like cookies, fingerprinting encroaches on consumer privacy. It does so by deriving data from different attributes of a device to give it a unique identity. But it doesn’t work as well as cookies, relying on probabilities and a mishmash of information to identify a user online. And fingerprinting’s flouting of user privacy hasn’t gone unnoticed, as it’s been outlawed by both Apple and Google.
The probabilistic guessing game
Probabilistic attribution uses machine learning to predict a user’s behavior based on the actions of an already-converted customer. The biggest issue with this type of marketing measurement is it’s based on probability rather than certainty.
Furthermore, like other methods, probabilistic attribution is not possible without user-level data, and this type of data will soon become obsolete.
It’s time for a paradigm shift
As an industry, we need to change our expectations of measurement and concede that it simply isn’t possible to track everything. Not only because of privacy restrictions, but also because even the most advanced AI cannot know what a user is really thinking when they visit a certain webpage.
We must also acknowledge that privacy legislation has been introduced for a reason. Any measurement platforms that rely on IDs must be shown the door. Once brands accept this, they can focus on implementing privacy-safe solutions that provide them with real answers.
All brands need to know is which marketing activities are performing and which are not, so they can uncover wasted spend and convert it into value.
“Data-Driven Thinking” is written by members of the media community and contains fresh ideas on the digital revolution in media.
Follow INCRMNTAL and AdExchanger on LinkedIn.
For more articles featuring Maor Sadra, click here.
