A Quick Guide to Facebook Attribution

From an advertisers perspective, Facebook advertising has become an essential part of the marketing mix, with increasing amounts of spend being dedicated to Paid Social activity on the channel.

But ensuring you get value for your spend is increasingly important in a highly competitive advertising market.

So, how does Facebook attribution work?

In this blog, we dig into that question and discuss:

• How Facebook attribution works
• iOS 14.5’s impact on Facebook reporting
• The challenges with Facebook attribution analytics
• Facebook attribution models available
• Alternatives to Facebook attribution

A quick overview on Facebook

For a business that was launched just under 19 years ago out of a college campus in Harvard, Facebook has had a disproportionately big impact on the world – and the advertising world, in particular.

With a potentially reachable market of 2.11bn users it has become the ‘go-to’ part of the media mix with advertising spend on the platform in Q4 of 2021 alone of a staggering $27.1bn.

However, the platform hasn’t been without its challenges in recent years. With issues around data breaches and growing concerns from advertisers around the ability to track the effectiveness of the increasing amounts of spend being ploughed into the channel. And ensuring that you are able to effectively measure ROAS on your activity.

How Facebook attribution works

So how does Facebook attribution work?

Unlike other analytics tools like Google Analytics – which uses Last Click by default and uses cookie tracking for attribution purposes – Facebook attribution works on the basis of allocating conversion credit based on two types of attribution windows:

• View-through – where is someone views you are ad within 24 hours and makes a purchase, Facebook is credited with the sale conversion
• Click-through – any purchase connected to a click on a Facebook ad within the last 28 days is credited to Facebook

Up until relatively recently, Facebook was tracking activity across a customer journey that reached from view content, to add to cart, to initiate checkout and then on to purchase.

And applying conversion in a View-Through attribution window of 24 hours and Click-Through window of 28 days.

One of the attractions this approach had for advertisers was that it took account of purchases that involved longer customer journey times.

But for many advertisers, there was also a frustration around the type of conversion data that Facebook was putting in front of them.

Particularly when they compared it with what they were seeing in their Google Analytics data.

The simple example below illustrates this:

• Imagine a user is browsing Facebook and sees an ad for your coffee machine…
• They click through on the ad but don’t buy right away
• 10 days later, they see a YouTube ad for your brand
• Do a quick search on Google for coffee machines
• See a re-targeting ad for your product and click to buy

Who gets credit for the sale?

Facebook would give credit to the click through on the original Facebook ad – because it is focusing only on platform ad impact and it is within its 28 day attribution window for clicks.

Whereas, Google Analytics would tell you that it was the re-targeting ad that should get 100% of the credit (due to the fact that it is using a Last-Click model that gives full credit to the last touchpoint).

The simple truth is that all of the touchpoints on this journey have had an impact and that should be reflected in the attribution reporting.

It is a situation that has had marketers scratching their heads at the big discrepancies between Facebook and GA reporting – and the advent of iOS14.5 has only exacerbated things.

iOS14.5’s impact on Facebook

Back in 2021, Apple took the decision to require users to opt-in for tracking on apps on their iOS devices.

And the impact on Facebook has been significant.

First of all, Facebook lost access to a significant chunk of its analytics data – by effectively losing tracking capability across a huge proportion of iOS devices.

At the same time, it reduced the tracking attribution window in its attribution model to 7 days.

Which has had a marked effect for brands with customer journeys that are outside of what is a very short period.

And this only serves to skew the conversion attribution discrepancies that we outlined above.

Facebook have responded to the situation, and the data gap, by effectively modelling the performance of your ads.

So, not considering full data about every click and conversion but more creating an average attribution model, based only on the data it can see. And certainly not taking into account conversion related activity on other channels.

One of the key things to be clear about up front here is that Facebook attribution is firmly focused on reporting on the effectiveness of ads within the Meta Ecosystem – which comprises of:

• Facebook
• Messenger
• Instagram
• Whatsapp

It is not multi-channel in terms of its conversion attribution and will always attribute conversion to a Facebook campaign – and not take account of activity on other channels – within its 28 days click and 1 day view attribution modelling windows.

Challenges with Facebook attribution analytics

And all of this is causing marketers to increasingly question the type of reporting and analytics that Facebook is putting in front of them.

In fact, our own research indicates there is real concerns around bias in AdTech reporting with an incredible 80% of marketers indicating they are concerned about the issue.

And taking a closer look at the data would appear to bear this out.

Take the example on the right which is for one of our clients.

Their challenge – due to the issues around different attribution windows and models – was that, while whilst Facebook was reporting a figure of £450k against a specific campaign they were running, Google Analytics was reporting attributable revenue of only £20k for the same campaign!

The heart of the issue here was the way that both Facebook and Google Analytics were choosing to interpret the data available to them.

Fortunately, we are able to use our own attribution solution, Corvidae, to effectively rebuild the broken data views and get to the true campaign revenue figure which was closer to £250k.

What Facebook attribution models are available?

To date we have only focused at a high level on the way that Facebook attributes conversion credit based on attribution and conversion windows.

The 28 day click model is the default option but it is worth pointing out that there are a choice of attribution models including:

• Even credit – where each and every touchpoint on the journey gets credit
• Last-Click – where, similar to the default in GA, full credit is given to the last touchpoint before conversion
• First-Click – turns Last-Click on its head and full credit goes to the initial click on the journey
• Positional – divides the highest percentage of credit to First and Last touchpoints and divides the rest up between the remaining ones
• Time decay – gives most credit to the touchpoints closer to conversion and less to touchpoints closer to the front-end of the journey

There is much discussion about what type of Facebook attribution model you might use depending on your own marketing needs.

For example, if you are looking to drive growth you might choose First-Click because top of the funnel matters more than anything.

But for us this type of analysis is missing the point a little.

Regardless of your choice of attribution model, it is clear to us that there are some fundamental challenges around the way that Facebook delivers attribution analytics including:

• the fact that is it only reporting on activity within the Meta Ecosystem which ignores a huge chunk of the wider customer journey
• issues around the shortened length of conversion window it is analysing
• limitations in the data available to analyse conversion post iOS14.5

What is the alternative to Facebook attribution?

So, what is the alternative to Facebook attribution?

Many brands are continuing to rely on Google Analytics.

But as we saw in our example above, GA is also prone to issues around the accuracy of its reporting.

Part of the reason is that it relies on a cookie-based model and cookies are notoriously poor at tracking cross-browser, cross device user journeys.

And our own work with clients has consistently shown that 80% of marketing data is incorrectly attributed.

And as the death of third-party cookies looms, many marketers are taking the opportunity to review their options and consider cookieless alternatives.

Which is where AI and Machine Learning driven attribution comes in.

We have pioneered the use of this type of technology which provides the capability to effectively rebuild your broken marketing data and make the break from siloed reporting in platforms like Facebook and Google Analytics.

To get a more accurate view of the impact of your marketing touchpoints right across the customer journey – not just in the context of a single channel like Facebook.

If you’d like to learn more, download a copy of our eBook, Is Cookie-Free Attribution a Myth?, which covers:

• The main reason cookie based attribution and platform reporting is flawed
• How to fix your 80% broken analytics data
• How our unique patented technology effectively replaces the cookie