Brand measurement hasn’t changed, only the conditions around it. Its purpose has always been to measure where brand discovery takes place, how we show up in those moments and then prove the effectiveness of that. All of that is still true. The challenge that now needs solving is the evaporation of the click. When your entire measurement solution is click based - what happens when an increasing proportion of discovery and consideration can happen without a click? This is what is driving a sudden demand from clients for agencies to offer MMM.
The good news is that conversation has suddenly become a lot easier - the same clients who were reliant on last-click 12 months ago are suddenly going full steam into MMM. The problem now is that the market is flooded with solutions positioning AI as the answer to their measurement. The entry price of doing MMM, or econometrics, was historically prohibitive because it was so expensive due to the skill and resource to effectively deliver it.
Whilst AI is helping to lower the cost of entry, unfortunately it means the market is proliferated with solutions who have popped up overnight and lean entirely on AI to do the thinking and analysis for them - all being prompted by someone who doesn’t fully understand modelling to begin with. The job of agencies therefore needs to weed out the good from the bad. Ensure you’re triangulating and validating the model through tests such as incrementality experiments and hold-outs.
Last-click attribution, the analytics model that assigns 100% of the credit for a sale or conversion to the final link or ad a customer clicked before purchasing, is an outdated tracking metric. And brands continue to react too slowly to changes in the digital landscape, meaning they’re playing catch up to 5, and even 10 years ago.
When Apple introduced Intelligent Tracking Prevention (ITP) and then App Tracking Transparency (ATT), that should have been the end of last-click attribution. When half of the UK’s smartphone users are on iOS, a significant amount of the data signals brands were basing measurement and investment on were decimated. That should have been the trigger for brands to find a more reliable way to measure their advertising effectiveness. But it wasn’t. Solutions like CAPI were introduced, and brands latched on to those because it felt like a safety net - something that meant they could cling on to last-click. But those solutions only go so far, so even with them in place, your last-click attribution still has a huge hole in its data, yet that’s what you’re basing your investment decisions off. It’s frightening how many millions of pounds each day are invested on totally flawed data.
The other area many brands are blind to is their adstock and advertising decay rate. Even beyond last-click and looking at DDA and MTA models, these are poor at looking at these kinds of marketing effects. Whilst you can get a small indication with sales showing up as cookied sales coming through after activity has stopped, it doesn’t show the overall lift a campaign had on your other channels.
If you’re running a YouTube campaign, an MTA may be able to identify that an exposed user subsequently converted but it cannot in isolation tell you how many would have happened anyway or the incremental effect of demand captured elsewhere. It won’t tell you if running that activity meant people were then more likely to click and convert on one of your PMax ads.
Fame is an asset that can help navigate many of the headwinds of changing market conditions. We saw it with Covid, we’re seeing it now with cost of living, and research is emerging that brand equity is also a key driver of AI visibility. So when focusing on conversion metrics, all that’s being measured is how much yesterday’s sale cost, and whether tomorrow’s sale can cost a little less. The focus isn’t on what caused that sale to happen.
Why did that person choose your brand? Was it the price? The creative? Is it because they were already familiar with the brand? How many times had they visited before making that purchase? Was there an opportunity to convince them to spend more during their transaction? And the catch-22 is that the more focus on conversion, the less that stuff matters. Weak brand strength erodes pricing power and your category distinctiveness. This means it isn’t just the marketing strategy that suffers; it can actively damage the brand that has been built over time.
Most brands are looking for an agency to tell them what is going to happen next. Being asked to predict the future through forecasts, budget laydown and corresponding executable plan is what a lot of CMOs and CFOs expect from an agency - that’s the easy part. The hard part is getting senior figures to switch focus on the leading indicators that matter and the different reporting cadence they require. If a client is used to the sugar-high of checking yesterday’s numbers, suddenly switching to focussing on certain metrics we are going to only check every 3 months can feel jarring. The way we do it is to build an entire measurement framework that incorporates both short-term metrics, and the indicators we’re using to identify future demand.
It’s important to have both, because if launching a traffic campaign, conversion rate could plummet, or awareness activity could mean ROAS will go the same way. As a bridge between the short and mid-long term KPIs, use “in-flight metrics”. These are the indicators that all your activity is working as expected. No client wants to hear “we’ll see you in 3 months when the brand lift study has concluded” to determine if a campaign was a success.
So instead, in-flight metrics such as CPM, frequency and VTR, can serve as an indicator as to whether creative is effective and you’re going to deliver the same reach as planned. These can indicate if the campaign is on track, but also fill that “sugar-high” of being able to check some metrics on a daily basis. That means reporting on leading indicators of future demand, such as Share of Search or our own Relative Discovery Score, without having last-click revenue as fall-back,
This is where discovery being mediated by AI gets dangerous. If discovery is taking place in AI, then the obvious metric is going to be AI visibility. How much am I showing up against my competitors, how many mentions do I have, how do I compare on Claude compared to ChatGPT etc.
Whilst visibility can be helpful to determine if you are at least showing up, beyond that it is largely nothing more than a vanity metric unless you’re measuring it with context. You need to track the semantics and sentiment around each mention. I have seen this first-hand recently for clients, where a leading AI tracking tool has them as number one against all their competition.
The problem was that it couldn’t be replicated. Every prompt I fed AI, and it didn’t matter which engine I used, was returning competitors time and time again. So the client was visible, but not against the prompts it wanted to be appearing for. That’s why you need to ensure the content you are putting out there, both owned content and that of influencers or content creators, needs to help the algorithm to ensure it is correctly identifying and categorising you.
If a lot of CFOs truly knew the extent of how much of their marketing budget was being invested almost blindly, and then being reported on incorrectly, they’d likely break away from last-click attribution.
At Journey Further, we have client meetings where the CFO is invited to be in the room, because once we’ve explained that without the right strategy the well is going to dry up, they often understand it better than some of the marketers in the room. It starts with being open and honest. Start with these 3 questions:
This doesn’t mean abandoning rigour - it actually should mean the opposite as one lagging indicator is no longer a measure of success. Instead triangulating multiple sources, and moving towards predicting future growth and reducing poor investment is the new focus - all things CFOs love! Some may view moving away from last-click as risky, but last-click doesn’t reduce risk, all it does is hide where the risk is.
The good news is that conversation has suddenly become a lot easier - the same clients who were reliant on last-click 12 months ago are suddenly going full steam into MMM. The problem now is that the market is flooded with solutions positioning AI as the answer to their measurement. The entry price of doing MMM, or econometrics, was historically prohibitive because it was so expensive due to the skill and resource to effectively deliver it.
Whilst AI is helping to lower the cost of entry, unfortunately it means the market is proliferated with solutions who have popped up overnight and lean entirely on AI to do the thinking and analysis for them - all being prompted by someone who doesn’t fully understand modelling to begin with. The job of agencies therefore needs to weed out the good from the bad. Ensure you’re triangulating and validating the model through tests such as incrementality experiments and hold-outs.
But last-click attribution continues to give brands a false sense of certainty
Last-click attribution, the analytics model that assigns 100% of the credit for a sale or conversion to the final link or ad a customer clicked before purchasing, is an outdated tracking metric. And brands continue to react too slowly to changes in the digital landscape, meaning they’re playing catch up to 5, and even 10 years ago.
When Apple introduced Intelligent Tracking Prevention (ITP) and then App Tracking Transparency (ATT), that should have been the end of last-click attribution. When half of the UK’s smartphone users are on iOS, a significant amount of the data signals brands were basing measurement and investment on were decimated. That should have been the trigger for brands to find a more reliable way to measure their advertising effectiveness. But it wasn’t. Solutions like CAPI were introduced, and brands latched on to those because it felt like a safety net - something that meant they could cling on to last-click. But those solutions only go so far, so even with them in place, your last-click attribution still has a huge hole in its data, yet that’s what you’re basing your investment decisions off. It’s frightening how many millions of pounds each day are invested on totally flawed data.
The other area many brands are blind to is their adstock and advertising decay rate. Even beyond last-click and looking at DDA and MTA models, these are poor at looking at these kinds of marketing effects. Whilst you can get a small indication with sales showing up as cookied sales coming through after activity has stopped, it doesn’t show the overall lift a campaign had on your other channels.
If you’re running a YouTube campaign, an MTA may be able to identify that an exposed user subsequently converted but it cannot in isolation tell you how many would have happened anyway or the incremental effect of demand captured elsewhere. It won’t tell you if running that activity meant people were then more likely to click and convert on one of your PMax ads.
The hidden cost of optimising for immediate conversion
Fame is an asset that can help navigate many of the headwinds of changing market conditions. We saw it with Covid, we’re seeing it now with cost of living, and research is emerging that brand equity is also a key driver of AI visibility. So when focusing on conversion metrics, all that’s being measured is how much yesterday’s sale cost, and whether tomorrow’s sale can cost a little less. The focus isn’t on what caused that sale to happen.
Why did that person choose your brand? Was it the price? The creative? Is it because they were already familiar with the brand? How many times had they visited before making that purchase? Was there an opportunity to convince them to spend more during their transaction? And the catch-22 is that the more focus on conversion, the less that stuff matters. Weak brand strength erodes pricing power and your category distinctiveness. This means it isn’t just the marketing strategy that suffers; it can actively damage the brand that has been built over time.
From lagging indicators to signals of future demand
Most brands are looking for an agency to tell them what is going to happen next. Being asked to predict the future through forecasts, budget laydown and corresponding executable plan is what a lot of CMOs and CFOs expect from an agency - that’s the easy part. The hard part is getting senior figures to switch focus on the leading indicators that matter and the different reporting cadence they require. If a client is used to the sugar-high of checking yesterday’s numbers, suddenly switching to focussing on certain metrics we are going to only check every 3 months can feel jarring. The way we do it is to build an entire measurement framework that incorporates both short-term metrics, and the indicators we’re using to identify future demand.
It’s important to have both, because if launching a traffic campaign, conversion rate could plummet, or awareness activity could mean ROAS will go the same way. As a bridge between the short and mid-long term KPIs, use “in-flight metrics”. These are the indicators that all your activity is working as expected. No client wants to hear “we’ll see you in 3 months when the brand lift study has concluded” to determine if a campaign was a success.
So instead, in-flight metrics such as CPM, frequency and VTR, can serve as an indicator as to whether creative is effective and you’re going to deliver the same reach as planned. These can indicate if the campaign is on track, but also fill that “sugar-high” of being able to check some metrics on a daily basis. That means reporting on leading indicators of future demand, such as Share of Search or our own Relative Discovery Score, without having last-click revenue as fall-back,
Visibility isn’t enough: Measuring discovery in context
This is where discovery being mediated by AI gets dangerous. If discovery is taking place in AI, then the obvious metric is going to be AI visibility. How much am I showing up against my competitors, how many mentions do I have, how do I compare on Claude compared to ChatGPT etc.
Whilst visibility can be helpful to determine if you are at least showing up, beyond that it is largely nothing more than a vanity metric unless you’re measuring it with context. You need to track the semantics and sentiment around each mention. I have seen this first-hand recently for clients, where a leading AI tracking tool has them as number one against all their competition.
The problem was that it couldn’t be replicated. Every prompt I fed AI, and it didn’t matter which engine I used, was returning competitors time and time again. So the client was visible, but not against the prompts it wanted to be appearing for. That’s why you need to ensure the content you are putting out there, both owned content and that of influencers or content creators, needs to help the algorithm to ensure it is correctly identifying and categorising you.
Breaking free from last-click without losing accountability
If a lot of CFOs truly knew the extent of how much of their marketing budget was being invested almost blindly, and then being reported on incorrectly, they’d likely break away from last-click attribution.
At Journey Further, we have client meetings where the CFO is invited to be in the room, because once we’ve explained that without the right strategy the well is going to dry up, they often understand it better than some of the marketers in the room. It starts with being open and honest. Start with these 3 questions:
- Do you feel you have full visibility of how all activity you’re running is performing?
- Can you confidently say where you would invest the next £1?
- If it was coming out of your pocket and not the company’s, would you be making the same investment decisions?
This doesn’t mean abandoning rigour - it actually should mean the opposite as one lagging indicator is no longer a measure of success. Instead triangulating multiple sources, and moving towards predicting future growth and reducing poor investment is the new focus - all things CFOs love! Some may view moving away from last-click as risky, but last-click doesn’t reduce risk, all it does is hide where the risk is.
Written by
Related Insights
How to build a discovery strategy that works across every channel
How to build a strategy where consistency meets creativity
What is discovery marketing?
Unlearning marketing’s biggest myths
The irony of performance marketing: why the smartest move is to not always depend on it
