Attribution 18 min read

How to Set Up Your Marketing Measurement for Success in 2023

By Laura Imrie 1 December, 2022

As marketers look to set themselves up for success in what is likely to be a challenging time next year, they are likely to be giving even more thought to how they measure marketing impact than ever.

In our webinar, How to Set Up Your Marketing Measurement for Success in 2023, our CEO and Founder Chris Liversidge unpicked the impact of a number of very specific challenges including:

  • Increased pressure to prove the impact of marketing spend in recession
  • The removal of 3rd party cookies
  • Illegality of GA, sunsetting Universal Analytics, the introduction of GA4 and replacement solutions for Cookies
  • How AI driven trumps Cookie-based solutions in terms of effective attribution

Our own poll during the webinar itself pointed to a 45% of marketers who identified Knowing Where to Spend Budget as their biggest in challenge.

Read on to find out how Chris helped them to address the issue in more depth.

Watch the webinar: How to Set Up Your Marketing Measurement for Success in 2023

Some background on the current state of measurement

So, first of all, a little bit of context.

We’re obviously really staring down the barrel of a recession by all measures in the UK. As a result of that and market pressures generally, about a quarter of marketers expect their marketing budget to reduce in 2023 according to the Integrate ‘State of Marketing Budget’ report. Certainly, that’s what we’re hearing from our own customer base.

And, in addition to that, data from a study from QueryClick – points to the fact that 2/3rds of marketers agree that the data they receive from analytics tools for cross channel decision making is broken.

Meaning, for a majority of marketers, we haven’t really moved on from a situation that we’ve pretty much been stuck with since the very early days of the internet. Where, as soon as an individual moves on to a different device like their phone from desktop (and maybe even into an app on their phone versus their web browser on their phone) measurement of the journey is not proving to be effective.

And, so as a result, many of us are still stuck using things like a Last-Click approach to measure this. Which is clearly suboptimal from an attribution point of view.

So, there’s pressure on our budgets going into next year being caused by recession and other effects. And there’s still this gap in terms of our ability to measure and therefore optimise our spend efficiently, particularly when it comes to allocating overall marketing budget.

With Google Analytics or Adobe Analytics, only really giving an accurate picture of short journeys. Often single device journeys, which typically represent about 20% of your total customer journeys.

So, that is only a view of a very brief moment in time, immediately before conversion.

Now, that is the sort of activity that Last-Click attribution models reflect very strongly in their weightings. And when you try and move from Last-Click to more sophisticated models (even if you’re doing that outside of Google and Adobe) what you will typically find is those models don’t become more effective or more predictive about a customer’s likelihood to convert.

And that’s because it’s data that’s only 20% accurate that you’re processing with those more advanced models.

So, Last-Click just simply is what we default to as a result of failing to be able to get something to work that’s more reflective of a complex journey.

So, how would our decisions change if that data accuracy went up to 95%? Because that’s the sort of impact you can expect to see from moving away from cookies to AI based measurement.

What will measurement in 2023 look like?

In 2023, we have a few key things that are technology changes that we need to look into when we’re thinking about our budget setting. Because they have a direct impact on the cost of acquisition that we can expect to achieve.

  • Removal of cookies
    Once again, we have third-party cookies being removed from Chrome. This is the second time that Google has delayed that by a year. It was due to occur in Q3 this year so right roundabout now and they have pushed that timeline back by a year. Now, it’s important to remember that Chrome is Google’s property. And Chrome is the majority of the marketplace but it’s not the entirety of it. And all other browsers have already removed third party cookies – more or less.
  • Google is sunsetting Universal Analytics
    This will be sunset in Q3 2023, funnily enough, when the removal of third-party cookies starts to impact us as well. So Q3 is shaping up to be a pretty transformational quarter.
  • The illegality of GA
    And finally, there are ongoing issues with Google Analytics. Not just the old version of Google Analytics (Universal Analytics) but the new version GA4. Which many of us are now onboarding. GA continues to have a few problems from a legality point of view and the GDPR point of view which has implications for media spend and efficiency.

Let’s take a look at each of these issues in turn.

The removal of cookies

So, first, the much-heralded death of third-party cookies. What exactly is happening with third-party cookies and Chrome?

As Chris explained on the webinar, the delay on this has happened a couple of times now. Google keeps pushing back its timeline.

However, we think it’s very unlikely that that timeline will be pushed back again, simply because of the potential legal consequences of doing so.

So really, we should be running as hard as we can away from the use of third-party data entirely today.

It’s true that Google has a majority of the market share. But more than 40% of that marketplace is already blocking third party cookies entirely. Meaning your use of that data – which is often ‘paid for’ data, is entirely wasted for that whole marketplace. And a different strategy that does not depend on trying to use third party cookies is required.

That principally means that your Display, Retargeting and Prospecting activity are most affected here with the following implications.

  • For many of us, Display is often a smaller amount of spend. And we are spending directly in Paid Search and Paid Social, but it’s extremely relevant. Because you are entirely dependent on those two existing channels without the ability to effectively prospect and capture new customers via Display

    And those two channels are increasingly expensive due to the volume of auctions increasing pretty strongly throughout this year. And that will continue to go up next year. So, advertisers are paying a lot more if they’re not putting money into good prospecting activity that’s outside of Search and Social
  • On the Retargeting side, it is very often accused of being cannibalistic activity. And often in Display, what we see when we have clients coming to onboard with us is that the Cost Per Acquisition in retargeting is actually propping up the Display prospecting activity
  • So, costs for those of us who are still using third party data will go up. Because our conversion behaviour will go down dramatically when 60% of the market then removes third party data

Why AI replaces cookies

If you move away from the use of cookies, your only option really is to look at a predictive technology – which is a different technology that allows you to measure across different data silos.

AI allows us to do that. It looks at the probability that a particular event in one silo is a continuation of events in another silo.

This is something that QueryClick has been working on been working on for the last eight years. Developing our AI modelling capability and investing about 5 million pounds or so in the technology, Corvidae (our attribution platform) and all the wraparound support functions required for a good quality enterprise deployment. And

AI has the ability to probabilistically join over data silos.

We’re now scoring roundabout 95% data accuracy for customers once they’re on board with us. So, 45 days after deploying a Corvidae pixel they’re moving up to +95% accuracy.

The accuracy is measured by understanding how well the AI predicts a conversion. And that’s a completely definable, measurable process. You can also measure the accuracy of the cookie-based conversion path being presented by competing platforms like Google Analytics and Adobe – which turn out to be only around about 20% accurate.

Google’s alternatives aren’t up to scratch

At this point in the webinar Chris took a look at some of the technology changes that Google are introducing in Q3 next year as replacements for third party cookies – as a means to assess the value in moving to AI driven attribution. Here is a very quick summary of the points covered (and links to more detail in each case):

  • Google Topics for Targeting – the first technology that Google is proposing to use in Chrome is to replace the prospecting use of third-party cookies via Google Topics. Topics provides a choice of 350 taxonomies or groups that advertisers will be allowed to bid against. One of the challenges with Topics is that the approach lacks granularity which means that targeting relevancy is degraded significantly. More detail on Topics >
  • FLEDGE for Remarketing (single device) – offers the potential to remarket to individuals on a browser-level basis on a single device where individuals can opt in to interest groups and be served relevant ads based on this. One of the limitations in this approach is that it is ‘opt-in’ on the browser which will potentially suffer the same challenges around opt-in rates that the IOS 14.5 encountered of less that 2%. More detail on FLEDGE >
  • Reporting API – this is the data we will receive from these new technologies. The API as it stands does not have any attribution as it only allows you to do last click measurement, which is not great. And it has no viewability control at all. So, ads that are shown below the fold, which aren’t ever actually seen by humans will be reported exactly the same way as the right front and centre screen. A major setback from a measurement point of view. More detail on Report API >

The final point Chris made is that this is also ‘on device’ only. Meaning it’s completely siloed to each individual device. Which means it’s no better than a cookie-based solution from a measurement point of view. Which will also have a negative impact on our ability to get efficiency in our spend.

The illegality of GA (Data Compliance for GDPR is not ignorable)

This is something we have been commenting on for several months now.

For the detail, take a look at our on-demand webinar ‘Are your web analytics illegal in the EU?‘. But here is a synopsis of the key points:

  • Google Analytics is problematic from a legal point of view. The sheer volume of data being processed means it has to do bulk of that processing in the US which is not GDPR compliant
  • There was a “safe harbour agreement” in place that was found to be illegal in 2020. There have been a number of rulings by the local data protection authorities in European countries, Sweden, France, Austria and Italy have all ruled that Google Analytics is illegal as it stands. European GDPR is the same as UK GDPR. So, whilst there has been no DPA ruling in the UK, we would expect to see more countries continue to rule – and that creates legal exposure for any of us with legal entities in those countries.

And GDPR fines are either £20 million, or 4% of your turnover if you’re lucky enough to have that be a higher number.

Ultimately, you do have to look somewhere else for your measurement solutions to remove any legal risks.

Which brings us back again to AI driven attribution.

How AI helps with efficiency in measurement

Here are just a couple of different ways that really good attribution makes a difference to your cost per acquisition:

Finding new areas of opportunity

In this case, which is one of our customers using our own Corvidae platform:

  • Paid Social – the impact of Paid Social revenue is being over-reported by Facebook reporting revenue (due to cookie-based measurement) at £248k when the figure as shown by AI driven attribution in Corvidae is actually £80k
  • Affiliates can be effective sometimes, but in this particular example, they’re really cannibalistic. So, they were always coming in towards the end of the conversion path. Taking more credit than they deserved for influencing the sale. So, what AI driven attribution is doing here is weighting the earlier interactions more and highlighting where in fact you’re getting cannibalistic behaviour
  • Search generic terms have three times more revenue than they are reporting in Google Ads.

And fundamentally, the other kind of control you get is a proper understanding of sources of revenue.

So, we have a recent case study from a customer who saw a 57% increase in revenue from looking at the reallocation of their revenue using Corvidae attribution. And they were finding about 11:1 ROAS on those reallocated activities, as an average.

So, AI-based technology allows us to sidestep the negative cost per acquisition pressure we’re all going to see in Q3 next year.

And this is the sort of thing that we should be planning into our budgets and our strategies.

In fact, Deloitte has found that it has had a significant impact in terms of straightforward economic effect for businesses and other areas. And it’s time for us to embrace that in the marketing world.

Source: Deloitte

There was a lot of ground to cover in the webinar. So, if you want to catch up on any of the topics raised and hear things directly from Chris – have a listen to our on demand version of the webinar.

Or, discover how AI driven attribution can set you up for success in 2023.

Want to learn more about attribution?

Identify wasted spend with our ROI calculator

Own your marketing data & simplify your tech stack.

Have you read?

See all articles