Facebook Attribution deep dive: Democratizing attribution for digital marketers

Facebook has put aside any platform bias and now there is a way to compare performance across platforms.

Chat with MarTechBot

Facebook ads app icon

Picture this – for a given period, your Google Ads campaigns record 500 conversions and for that same period, Facebook Ads campaigns show 700 conversions but the total number of conversions in Google Analytics was closer to 1000 instead of 1200? I see this every day and it’s one of the most frustrating aspects of being a digital marketer – and it makes it challenging for stakeholders to understand the true impact of your marketing initiatives.

As investment in digital marketing continues to increase, we have seen significant fragmentation in the channels marketers are using to reach their target consumers. In step with the diversification of channels, buyer journeys have become more complicated to track as consumers move freely across the open web, and traditional approaches to attribution have fallen short on reporting the true value of different marketing initiatives.

The promise of attribution

The promise of attribution is an exciting one – it’s something that marketers have longed for since the dawn of the field. Many of you will be familiar with the challenge posed by John Wanamaker at the dawn of the 21st century – “Half the money I spend on advertising is wasted; the trouble is I don’t know which half.” And while this may be a gross oversimplification, the question remains – how effective is my marketing mix, and where should I invest to grow

Fundamentally, attribution requires two components:

Screen Shot 2018 11 08 At 2.30.49 PM

  • Customer Journey Map – The detection of marketing touchpoints (clicks & impressions) associated with a unique user across devices & time.
  • The Applied Model – This includes both rules-based and data-driven models. Traditionally, most marketers have focused on Last Touch as this has been the default model applied in most analytics platforms, but also includes First Touch, Linear, Position-Based, Time-Decay & Data-Driven.

Over the past few years, I’ve observed the way we, as marketers, talk about attribution. There has been a strong focus, almost entirely, on the applied model with limited thought given to the customer journey mapping. Perhaps this is because we have given up on finding a solution that will allow us to combine clicks and views across platforms within the same customer journey map, because it seems too hard or complicated. Whatever the reason, with the launch of Facebook Attribution we can finally look towards the future with greater clarity and certainty.

Before we dive into why Facebook Attribution presents the potential for a great leap forward in marketing technology, it’s important that we understand the output of attribution. Ultimately, the point of attribution is to understand the various interactions across channels that lead to conversions and the relative weight that should be applied to each interaction. Attribution will not prescribe the path forward, rather, it is a solution that informs and guides decisions around our marketing mix moving forward. Looking to dig deeper into attribution theory? Check out this recent presentation I shared on the topic of Data-Driven Attribution at SMX.

Why Facebook?

Three words: People-Based Marketing. As the largest database of humans on the internet, Facebook is uniquely positioned to create a unified view of marketing touchpoints across consumers path’s to purchase. People-Based Marketing in a concept that Facebook pioneered in 2014 with the vision to create an advertising world that allowed for streamlined and uniform audience marketing across platforms. While this is the promise of programmatic advertising in general, we start to dilute the waters as we use different platforms and data sources for audience targeting and limit our clarity around our audiences being targeted.

The cross-device and cross-platform nature of People-Based Marketing puts Facebook in a unique position to create a solution that brings all online marketing touchpoints together. In this author’s opinion, this is likely why Google Attribution has been teased for so long and has yet to be released. With that said, Google has been instrumental in driving the Data-Driven Attribution movement and will continue to be a powerful player here given the mass adoption of Google Analytics as the ‘source of truth’ for many organizations.

Getting started with Facebook Attribution

To get started with Facebook Attribution, you’ll need to have the Facebook Events Pixel implemented on your website. At a bare minimum, you’ll need to have the PageView and Conversion events passing data back. Given that you’re reading this article, it’s highly likely you already have the pixel installed correctly. As a general note, conversion data can also be imported from Offline Conversions as identified by user data upload to Facebook.

Facebook has put together this guide for configuring the platform, and while there are some technical aspects, the majority of 3rd party platform tracking can be handled by simply created automated reports within these platforms and applying click trackers (provided by Facebook).

Of note, Facebook currently supports the inclusion of data from the following platforms:

Preconfigu Manual Config Chart

There is also an option to generate tracking for a custom platform, but it should be noted that LinkedIn, Pinterest, Snapchat and Twitter have chosen not to support this level of tracking at this time. In a move that would be more likely to appear on a website protesting the actions of a politician, Facebook actually makes reference to this within the platform and even asks you to make this request on their behalf.

twitter screen

In addition to integration with various paid channels, Facebook Attribution also includes UTM tracking so we are able to bring in click data from organic search, email or other social platforms that have driven traffic to your site.

Leveraging Facebook Attribution

After you have everything set up, you’ll need to wait a few days for data to start recording and appearing within the platform. In general, you should look to wait at least 30 days before using the data within the account to inform decisions. This is based on the click attribution window you are planning on applying, which be either 1, 7, 14, 28, 30 or 90 days. Given the default window for most platforms will be 28 days for click data and 1 day of impression data, I highly recommend using this within the platform in order to maintain consistency and start using the platform to guide decisions. Over time, you may choose to experiment depending on the length of the path conversion, but to go any shorter, would be to occlude the impact of upper funnel initiatives that are particularly hard to measure the impact of.

You can change the applied model & attribution window within the Performance pane of the platform:

Screen Shot 2018 11 08 At 3.28.53 PM

Which will allow for the creation of a high-level overview of the top-channels performance below:


Wait, is that Google Ads taking the top spot? The output of this report obviously departs on your marketing mix – but it’s important to note that Facebook has put aside any platform bias and produced a platform that finally allows for the combination of conversions from Facebook and Google Ads (as well as many others!) reported as a single source of truth. This is something that we have been in search of for years – and we finally a way to compare performance across platforms.

Digging deeper into the platform, we can utilize the Custom reports pane to actually dive into the models and leverage Source for a macro-level view or Campaign for micro-level. This distinction is big as it allows for a channel-level overview of performance with the ability to get more granular once you have detected overarching trends by channel. There is also a custom breakdown component, although it appears to be a little clunky right now, the promise of campaign segmentation by audience targeted (new vs. returning) and stage within the funnel is incredibly exciting and will go a long way to aiding the discovery of potential efficiencies throughout the various stages of the buyer’s journey.

Screen Shot 2018 11 08 At 3.39.13 PM 1

When extracted and studied over time, we can use this report to compare against changes in campaign strategy and understand the incremental impact of new activities, as well as the relative cost-per-acquisition and efficiency across channels and campaigns.

As previously mentioned, Facebook has the industry-leading cross-device graph available, and provide a Cross-device pane for marketers to get anecdotal insights into the number of users that engage with ads on one device and convert on another. This data is certainly interesting and provides a different spin on the typical cross-device graphs provide by various analytics platforms.

The challenges

  • Limited Data Sources – This will likely continue to be a challenge for all attribution platforms until the end of time. There are privacy, technical and competitive issues associated with the aggregation of this data. Notably absent are many Social Platforms, Amazon, TV and OOH media – however, the fact we can combine so many platforms (including those owned by Google) is amazing and something we need to embrace so we can encourage Facebook to continue innovation here and break down barriers to form new relationships with these other platforms.
  • Model Application – Currently users can select between last touch, last click or visit, even credit, positional 30 percent, positional 40 percent, time-decay 1-day and time-decay 7-day models. These models are all rules-based which is inherently biased and cannot account for the wide variety of touchpoints that may influence a customer. Facebook has provided a data-driven model, but at this time, this is only available to be used with data from Facebook Ads. While we would love to see a data-driven model implemented for all channels, time-decay 7-day is a good one to leverage for decision making.
  • Education – Although this is the least technical challenge, it may, in fact, be the hardest. How do we use attribution models? How do we build trust with decision makers that this is the source of truth? We must take responsibility for telling this story within our organizations and encouraging the adoption of Facebook Attribution.

Looking toward the horizon

So, what does the future look like? It looks complicated – that much we know. Just wait until Amazon starts marketing via their talking microwave and you need to figure out how to attribute that touchpoint – of course, I’m kidding (maybe) but if the past 5 to 10 years are anything to go by, our channels will continue to become more fragmented, but our user journeys will continue to become more connected as tracking technology becomes omnipresent. (In a totally non-dystopian future kind of way of course!)

The more immediate future will likely see Facebook release API reporting access here, which will enable for this solution to scale. We’re already dreaming of the ways we can visualize this within Google Data Studio and suspect they will likely release some degree of reporting automation in the near future.

Facebook Attribution could be in a turning point in digital marketing measurement – but it will only succeed if marketers embrace this technology. This will democratize access to attribution and make it accessible to organizations of all sizes beyond those that have historically looked to invest in an enterprise attribution platform such as Visual IQ. Oh yeah, did I mention Facebook Attribution is free? This could be as big as Google Analytics for Facebook and the future is shaping up to be a very measurable one.

Opinions expressed in this article are those of the guest author and not necessarily MarTech. Staff authors are listed here.

About the author

Simon Poulton
Simon has had a passion for finding creative ways to measure real-world challenges from an early age. Combining an affinity for psychology, statistics, and digital marketing, he is currently the VP of Digital Intelligence at Wpromote. Simon regularly speaks at industry events, including SMX West, SMX East, Cleveland Research Group's eCommerce Catalyst for Change, and SMX Advanced on a variety of topics related to data-driven digital marketing.

Get the must-read newsletter for marketers.