Agnostic attribution: the vital ingredient for a single source of marketing truth  

Marketing attribution is assessing which of your marketing activities and touchpoints are driving conversions and revenue for your brand.  

It can be a tall order to get to the true picture on this, but it can be nigh on impossible when biased attribution data obscures the real truth of the situation. Too often, attribution data provides only a partial view of things, that is more concerned with driving spend into a particular channel than giving an accurate assessment of overall performance.  

Here we take a close look at the issue as we consider:

• What is marketing attribution bias?
• 4 reasons you should be concerned about it
• Why does marketing attribution bias happen?
• Why agnostic attribution holds the key to removing it
• How to sidestep attribution bias with Corvidae

What is marketing attribution bias? 

Marketing attribution is the process of identifying the extent to which individual touchpoints along the customer journey influence a decision to purchase.  

The ability to accurately quantify the true impact of marketing channels, campaigns and media creative is foundational for digital marketers, as they look to develop their strategy and optimise spend.  

Delivering effective attribution in what is an increasingly multi-device and multi-channel world can be challenging – with research from Gartner suggesting that 60% of CMOs plan to cut marketing analytics teams, and our own research indicating that 62% of marketers believe data to support cross-channel decision-making is broken. It is a problem that has its roots in a range of issues from increasingly siloed data sets to the use of outdated technology to track and analyse customer marketing activity. 

One of the contributing factors that often flies under the radar, but is increasingly a concern for marketers, is the issue of attribution bias. The term attribution bias refers to the process of inaccurately assigning value to specific touchpoints in the customer journey when determining the impact they are having on conversion. So, it is effectively about telling you things that might not necessarily be true, or even hiding fragments of the truth from you. 

Attribution bias happens for a range of reasons that we will explore next but it is clear that marketers are concerned about the issue – with 80% of marketers indicating they are increasingly concerned about the potential for bias in their AdTech reporting in particular.   

4 reasons you should be concerned about attribution bias

So, attribution bias exists but why should marketers care? We think there are 4 key reasons as follows:

1. It’s bias and it is wrong…plain and simple – at the most fundamental level bias in any shape or form is wrong, and when it is being displayed by some of the digital ad and analytics solution providers you are placing your trust in, it something you are going to want to root out.
2. It leads to sub-optimal spend allocation decisions – partial and selective views of the customer journey and black box style approaches to measurement are skewing performance metrics and pushing higher levels of spend into audience targeting and campaigns that are unlikely to convert.
3. There is a knock-on impact on ROI and revenue – failure to properly connect spend to revenue accurately due to marketing data bias leads to a lack of control over ROI and can mean making the case for marketing investment becomes infinitely harder.
4. Encourages an unhealthy focus on lower-funnel media – something we will return to later, but much of the bias being shown by the big AdTech players is significantly overvaluing costly touchpoints closer to conversion and driving up CPAs in the process

Why does marketing attribution bias happen?

So why is marketing attribution bias even a thing anyway?

Marketing bias happens for a wide range of different reasons but here are some of the key ones we come across:

1. Subjective bias

Before we get into more of the specifics of bias in AdTech measurement it is worth pointing out that digital marketing – despite its heavy reliance on marketing technology – is still a people business on the front end.

Everyone from Performance Marketing Directors to Heads of Analytics are human after all, and the reality is that bias can creep into all aspects of your attribution and measurement approach without you even necessarily being cognisant of it. From your choice of AdTech solution, to which attribution model or models you apply and how you – or your technology solution – chooses to weight and rank the impact of specific channels, campaigns and individual creative executions.

So being mindful of your conscious, or even unconscious, bias is key as you set your measurement model up. 

2. The impact of outdated and obsolete technology

However, it is also the case that outdated approaches to marketing measurement – that are simply not built to be able to cope with the intricacies of multi-channel and multi-device customer journeys – are showing a heavily biased view of the impact of your marketing.

Consider the example of third-party cookies, the underlying technology that is used to deliver attribution in some of the most commonly used analytics solutions like Google Analytics and Adobe. Much of the discussion around the recent on-off soap opera that developed around Google’s plan to remove (and then not remove!) third-party cookies from its Chrome browser missed the main point – which is that cookies do an awful job of tracking complex customer journeys.

This is shown clearly in the example below. The top user path, which was constructed using our own AI-driven attribution platform, is able to follow the customer journey and accurately value customer touchpoints – even as the user switches devices.

However, because of limitations in the cookie/pixel technology for tracking purposes, the cookie-enabled solution (in this example GA) is not able to follow the user and splits the journey into 2 paths A and B.

This has huge implications for the attribution story it is telling its users, by over-emphasising the impact of lower funnel activity (the only part of the journey it can see and connect to conversion) like PPC whilst ignoring higher value touchpoints earlier in the journey.

The result? Attribution bias on a grand scale.

3. More cynical manipulation of marketing measurement

So technical bias in their chosen analytics solution is clearly a worry for marketers, but what is perhaps more of a concern is the more cynical use of data bias by some of the big AdTech providers.

Consider the example of Google recently. Not only does it control a large chunk of digital ad spend on its Google Ad platform but it also provides its Google Analytics measurement solution to assess the effectiveness of those ads. This creates a clear misalignment in terms of the potential for bias in assessing ad performance on its network.

And two recent developments have left Google wide open to accusations of bias in this respect.

• Firstly, and incredibly, Google changed the weighting on its data-driven attribution model to give more credit to Paid Search in Google Ads. A move with a clear conflict of interest if there was one!
• Secondly, they also removed all other attribution models from its platform – with the exception of Last-Click – which means that if you are using GA to assess the effectiveness of your marketing performance you are extremely unlikely to find any result other than that Google’s own adverts are performing best for you.

These recent issues with Google are only the tip of the iceberg for marketers. Consider the customer example here where they found that Facebook was reporting just under £450k of revenue for a campaign – whilst GA was indicating just £24k for the same level of activity!

The discrepancy comes down purely and simply to the respective platforms effectively marking their own homework and creating versions of the truth that best fit their need to drive revenue onto their platforms. This leaves marketers grasping for the real answers to campaign effectiveness as they continue the search for a more accurate picture of the truth.

Why agnostic attribution holds the key to removing attribution bias

For companies that sell attribution AND AdTech or ads, the motivations are clear. If they tell you that spending more money on their platforms is working well and driving revenue for you, the likelihood is you will spend more with them. Why would you go anywhere else?

With large independent attribution providers increasingly partnering with companies that are selling ad space the situation gets more and more messy. How much can you really trust the in-house attribution and reporting solutions from companies that simply want you to spend more on their platform?

Which is why our Corvidae attribution platform is proudly agnostic. We aren’t under any pressure to try and funnel your ad spend into certain channels – and our only ‘skin in the game’ is to provide you with a single, independent source of marketing truth.

Here are 3 key ways that Corvidae helps you strip away the biased and siloed view you are getting from many AdTech providers.

1. Rebuilding your broken marketing data to give one single source of the truth

Your marketing data is foundational in terms of effective attribution.

Unfortunately, the siloed nature of this data across the various advertising and analytics platforms you use is holding your attribution back. Our AI-enabled attribution can collect data from your various platforms, it then uses probabilistic modelling, deep Machine Learning and our patented stitching technology to effectively rebuild and unify this broken data to deliver one single source of the truth as a base for all of your attribution activity.

This helps to strip away the bias inherent in platform-specific reporting.

2. Struggling to reconcile paid spend on social – no problem!

We mentioned earlier the very real concerns that marketers have regarding bias in social.

The good news? Our approach enables you to hold your siloed AdTech data firmly to account.

In the customer example below, we were able to show that Meta was effectively over-reporting on campaigns by an incredible 300%. By effectively pulling Meta’s data into Corvidae using their API and showing it alongside with the genuinely attributed data – which is available as part of our rebuilding and stitching process.

3. Attribute, optimise and grow – free of bias

And that is the power of our AI-driven attribution solution.

It enables marketers to side-step the issues of bias and partial reporting on customer journeys that are inherent in cookie-enabled AdTech platforms. This allows you to drive revenue and ROAS with a clear analysis that shows where to commit spend – and, just as importantly, where not to!

Sidestep attribution bias with Corvidae now

Find out how Corvidae can help provide the single view of the truth you need on your marketing efforts – and consign attribution bias to history. 

Request a demo with one of our experts now.