Attribution

The new era of attribution – radical approaches for ad performance and improvement

2nd October 2024

Carly Bell

Recent industry developments including the findings of the Google Antitrust Case, changes in weighting of its attribution model and the cancellation of plans to remove third-party cookies from the Chrome browser are transforming the landscape that marketers have to operate against.

In our webinar The new era of attribution, our CEO Chris Liversidge and Claud Munro, Client Success Director at Corvidae took a look at what this means for marketing performance including:

How Google was found guilty in the anti-trust case regarding ad price manipulation
How Google quietly changed its attribution model for Ads in June and what it means
The implications of Google cancelling plans for third-party deprecation in Chrome
How Gift Universe has deployed AI-driven marketing attribution to take these issues out of the measurement equation

Google is found guilty in the anti-trust case regarding ad price manipulation

First up, Google has indeed been found culpable in the 10-week antitrust case that’s been ongoing in the US for a number of months now. That case found that they were using anti-competitive practices and there was a lot of deposition from senior Google execs, particularly execs responsible for their search Ads platform. One of the big standout insights was that Google does indeed tip the scales of the auctions that are being run in their search results.

So, each quarter, if Google feels that it is not meeting its quarterly business targets – at the level its parent company Alphabet requires – then those auction prices will be nudged up. A control that Google has acknowledged and is on record as doing as part of this antitrust case. An astonishing admission and you can see those price trends when you look at the data.

So, we have a signal here in terms of the way Google is operating with its Ad platform. Ultimately it can present itself as providing solutions and options for its customers, but it has also been found guilty of manipulating those auctions for its own benefit, as a result of having a monopoly.

And so that really increases the requirement for us as marketers to have another source of truth – or an alternative source of truth, at least – as we strive for advertising improvement. To help us counterbalance the data and pricing that’s being presented back to us by Google when we are running ads on their networks.

How Google quietly changed its attribution model in June

The second big piece of news is that, in June, Google also stepped into its attribution platform – so that sits across Google Ads but also GA4 – and it decided to tip the scales again.

Now in this instance, the way it tipped the scales was by changing the weight of its data attribution model to give more credit to Paid Search. The reason it gave for that was that it was unable to track across devices in its measurement because it relies on first-party cookies. For clarity, this is a situation where Google is marking its own homework with data-driven attribution.

Last year Google also removed all the other attribution models from its platform, apart from Last-Click. So, you are really only able to look at Last-Touch or data-driven attribution in GA. And you have here a really strong signal that if you’re using Google’s data-driven attribution to measure marketing performance, you’re extremely unlikely to find any result other than that Google’s own adverts are performing best for you.

We know from our work with customers that Google’s data-driven attribution is really weighting heavily towards those last touch points – in fact, it’s more or less equivalent to Last-Touch. And the reason it does that is that people are searching and ‘in the market’ when they’re quite far down their consideration journey. And have already been heavily influenced by things like your brand marketing and so on.

And so, ultimately, this data-driven attribution model- that’s being presented to you in Google Ads and GA4 – is artificially boosting any Google Search Ads performance that’s being captured.

Google also cancelled its plans to remove third-party cookies

And then the third major piece of news is that Google has given up on trying to remove third-party cookies from Google Chrome, or at least that’s the headline.

In reality, there are a couple of subtleties to what Google has decided to do.

So a quick recap, Google owns Chrome. Chrome is 60% of the browser market by default and all other browsers essentially block third-party cookies straight out of the box. Google was going to switch to that and provide us with some tools so that you can continue to advertise to Google Chrome users using their Privacy Sandbox tools.

Those tools are opt-in, meaning in effect, you are not going to be able to advertise to the vast majority of that audience using cookie-based or deterministic-based advertising. That means that when you are using audiences to push out and target your ads, the audiences available to you are now vanishingly small. Not ideal in the quest for advertising improvement and audience development.

There’s an awful lot of evidence to suggest that the cost of that third-party data outweighs any incrementality you would get from adding that level of targeting.

Ultimately, what’s happening here is Google is stepping back from taking responsibility for data compliance after it was challenged essentially by the IAB and other major data compliance organisations in the UK and Europe. And it is going to effect a third-party opt-out that will happen at some point next year.

This has several implications including:

advertisers will have a very, very small pool of audience to bid against
this means that the planning that we have been doing with third-party depreciation coming down the pipe should broadly speaking, stay in place

But there’s another implication to this change especially when you stack it with the two previous changes that are highlighted above:

this is simply that, if you are relying on Google data to optimise your Google ads, you are operating in a world where:

your attribution model is stacked against you
- your cost per click is also stacked against you
- and, ultimately, your audience bid options are being removed

The result? There has never been a stronger case for looking at alternative paths to ad optimisation, which fundamentally is all that we are trying to do as marketers now.

Corvidae has been solving this problem for several years now. We have many customers who have gone through the process of switching away from cookie-based, Google measurement and are relying entirely on AI-based stitching. Which is something that is unique to the market and which allows you to:

send much more accurate data: back into Google’s Ad tech (and Meta and other ad tech providers) based on much longer customer journeys that are being stitched by these AI processes
reduce your cost per acquisition: even in the face of Google manipulating the auction because they’re trying to hit their quarterly targets

How the use of AI transforms your marketing attribution

So how does it work? We ultimately replace the use of cookies in all of your measurement.

AI sidesteps the issues associated with cookie-based measurement

The use of AI means that you no longer have to worry about people moving across different devices as they go towards a conversion point. Fundamentally, Corvidae takes care of all of that.

We see on average something like 28 different touchpoints in a typical shopping-based or e-commerce-based conversion journey – and sometimes quite longer journeys for some more B2B-related purchases. And things like the Financial Services products for example.

Crucially, those touchpoints never occur on a single device. They’re always multi-device. All of your data fundamentally is based on the accuracy of those journeys being stitched together. We would contend that using cookies for that – a 20-year-old technology that isn’t able to represent sessions going from one device to another – is never going to give you a good result. Thankfully, we see fantastic outcomes by moving to the use of AI.

Corvidae uses AI to rebuild customer journey data

When customers onboard with Corvidae they can get a side-by-side comparison of how much longer the journeys are that are being stitched together with our AI process.

You can see an example below for one of our customers where AI is stitching paths nearly five times as long as the cookie-based paths being stitched together by GA4.

It underpins all of our data accuracy and it’s compliant ‘out of the box’.

Helping you fight back against the never-ending rise in CPAs

And crucially Corvidae allows you to take control of your cost per acquisition.

It does that by allowing you to share these genuinely longer journeys with your Ad Tech. This enables your Ad Tech to see auctions, which it previously thought on cookie-based measurement did not proceed towards a conversion (but did!).

And by bidding in those auctions – because there are fewer competitors and most of your competing bids are being measured by that same cookie-based analysis – there are very low bids and low bid volumes in those auctions.

So, they are cheaper which typically delivers CPA reductions of up to 45%.

Deploy and test different attribution models that suit your needs

Importantly, Corvidae does not limit your options in terms of finding a model that works for you.

We are not advising that there’s one model to rule them all. Instead, we are providing a platform where we have done the stitching and all common models are available to you to test and discover incrementality with, and we have an expert team that works to help you, and your agency, discover those growth opportunities.

Why does that matter?

Above, we see models that are extremely effective in Google Ads – which may be weighted slightly more towards the conversion point – and perform really well here. But these models are outperformed in Meta – which represents social advertising – and that is often valuable when you’re trying to discover new customers. It is an entirely different use case – and a model that weights more towards those early touch points will give you greater success as you deploy automated feedback from Corvidae on the success of auctions for your ads.

So, AI-driven attribution provides a clear way to futureproof yourself by gaining control over measurement and moving away from cookies in a way that no longer relies on the type of self-reporting you get from Google.

Which is the approach that Gift Universe is taking.

Customer focus – how Gift Universe cut CPAs by 31% with Corvidae

Gift Universe is one of the UK’s leading and fastest-growing operators in the gifts and gadgets industry. It has a number of trading brands including Men Kind, The Source and Prezzybox as well as several product brands.

They had 4 key objectives when they approached us about using Corvidae to improve the effectiveness of their attribution as follows. They wanted to:

fight back against rising CPAs
gain a better understanding of the paths that customers were taking to convert
find an attribution solution that they could trust
expand spend and invest in their brand with confidence

We were able to quickly onboard them to our Corvidae platform and run a pilot programme which:

stitched together customer journeys that were up to 40% longer than previously
has increased conversions by 55%
unlocked average CPA reductions of 31%
driven an increase in ROAS of 42% on average

Start your new Era of Attribution with a demo of Corvidae now