Beware the Pixel Pushers: 3 Reasons to Quit Click-Based Attribution

With ongoing economic uncertainty, ecommerce marketers are under more pressure than ever to maximize the value of each advertising dollar spent. To do so, they need to know how many sales ads are actually driving so they can move budget to the best performing channels and campaigns. Every marketer wants to maximize the efficiency of their ad spend, but the wrong measurement methodology will lead to risky investments.

Let’s explore why attribution solutions with nomenclature like “pixel-based,” “1st-party attribution,” or “click-path analysis” should be evaluated with a close eye.

Reason #1: Pixel-based attribution is based on the assumption thatcorrelation equals causation. It does not.

Platforms like Facebook and Google have a siloed view of marketing. Using “last-touch” attribution, they will take credit for any conversion that was preceded by their ads, even when other ads or other channels may have been involved.

“Multi-touch attribution” (MTA) is a popular evolution of this concept offered by many 3rd party measurement providers. MTA attempts to track all the clicks on a consumer’s path to purchase and assign variable credit for a conversion to each touchpoint. Connecting individuals to each ad they saw and the products they purchased is an attractive story that is easy for marketers to conceptualize, hence MTA’s continued staying power despite its many issues.

What’s important to understand about both MTA and last-touch attribution methods is that they are based on correlation, not causation, meaning they do not actually prove that an ad caused a conversion. MTA assumes that because a user clicked an ad, it contributed to the sale, but that is frequently not the case. Correlation-based methodologies do not account for purchases that would have happened regardless of the ad. This often leads to over-attributing conversions.

Additionally, pixel-based attribution cannot account for view-through conversions, as they cannot collect impression data, further worsening its ability to build a complete perspective of the mythical consumer click path.

Reason #2: Pixel-based Attribution Measures an Incomplete Picture, Despite Promising User Journey Analysis

Pixel-based attribution is trying to work around increasing restrictions on user-tracking and data access to keep multi-touch attribution (MTA) alive, which was a sub-optimal methodology even at the peak of third-party user data access in the early 2010s.

As new data-privacy policies such as Apple’s ATT, and Facebook’s reduced lookback windows following iOS 14.5 continue to mount, Last-touch and Multi-touch providers are struggling to get a complete picture of the consumer journey. In fact, soon after Apple’s rollout of ATT, Meta found that last-click models undervalued marketing on FB/IG platforms by 47% on average and recommended incrementality as the only viable measurement solution.

Pixels are not a new technology. Conversion tracking pixels are the fundamental component of Google Analytics, already in use for last-click measurement by the vast majority of ecommerce marketers. In response to the increasing restrictions and regulations, MTA providers are rebranding their methodology as “Pixel Based Attribution” or “1st Party based Attribution.” Savvy marketers will be well served by understanding that these rebranded methods are rife with the same old challenges, often leading to misguided performance insights and sub-optimal marketing investment decisions.

New pixel-based attribution providers simply apply a new counting model to the data their pixel collects. The most common issues are incomplete data and data duplication. Pixels rely on user cookies to connect the dots between the ad platform and the site visit. With cookies all but gone on the majority of user devices, providers are attempting to replace this data with “Synthetic Touchpoints,” “Machine-Learning” algorithms, or third-party identity resolution solutions to fill in the missing pieces.

Due to compressed attribution windows and limited data collection, with pixel-based applications, upper-funnel, prospecting, and brand awareness campaigns often look like bad performers, when in reality they are actually driving many incremental sales. In a recent incrementality test Measured conducted for a premium fashion brand, pixel-based measurement was only able to account for 1/4th of the net-new sales contributed by the ad campaign being tested. This is not uncommon, as pixel-based measurement often optimizes towards lower funnel media, such as retargeting, due to the higher volume of click-data visible near the purchase point.

Incrementality-based measurement does not have these same limitations, as it uses test-and-control based methodologies and actual sales transaction data to measure the impact of marketing channels and campaigns. Incrementality does not require user tracking, and takes into account all of a campaign’s influence, clicks or views, over a purchase. Most importantly, incrementality doesn’t try to paint the impossible “consumer journey mapping” as possible or calculate results in a mysterious black box of modeling and voodoo.

Reason #3: Pixels Wreak Havoc on Websites

The display advertising ecosystem recognized this issue nearly a decade ago during the rise of publisher website optimization and SSP exchange-based bidding. Ecommerce brands are now following suit and taking more care to ensure their websites are designed to reduce consumer friction and maximize sales.

According to ecommerce giant Shopify, conversions decrease an average of 7% for every additional second it takes a website to load . As tracking Pixels ping one or more external servers on every page load, over time these server calls can severely impact site performance. They additionally require frequent maintenance, and are prone to breakage, which leads to frustrated site teams and further reduced measurement. Marketers working to optimize business impact should take strong consideration before adopting a pixel-based approach.

Perhaps even more importantly, tracking pixels must hide their identity to appear as a component from the brand’s native website in order to access any available user identity information as a 1st party. In order to do so, they require CNAME or Canonical Name changes. For example, instead of appearing as a pixel from pixel.mtaprovider.com, they appear as mtaprovider.yourwebsite.com. This practice also reduces the likelihood of the provider’s URL triggering an adblocker.

The problem with this, is that CNAME capabilities, like cookies, are likely going to be deprecated in the near future , and this loophole will no longer be available to allow third party pixel providers to access user or ad platform information. This presents yet another challenge for pixel-based attribution providers as they to keep the outdated multi-touch attribution story alive.

Ultimately, pixel-based attribution has an expiration date that is rapidly approaching. And, for all the reasons above, it's a bad idea to risk any of your advertising dollars on the questionable insights click-based MTA vendors are providing. Measuring for incrementality is the only way to get reliable results, validated by sales data - no user-level tracking required - so you can confidently allocate ad spend for maximum return.

Check out how other brands are using incrementality to optimize their ad spend and increase ROAS, then contact us to learn more about the Measured Incrementality Platform!