APG Noisy Thinking: What should we be measuring in 2026, and why? Or… Getting creative with measurement

I hesitate to begin this write-up with a negative, but I think I could make a reasonable bet that most planners and strategists would not put measurement and effectiveness at the top of their favourite ‘Planning Things To Do’ list.

It’s easy to shunt what seem like the difficult, data heavy aspects of Planning to another day, especially if you prefer the creative thinking aspects of the job and all the mental and imaginative stimulation that these can deliver.

This was the impetus for Noisy Thinking. Precisely to try and show that the whole game of measurement can be as rich and varied and interesting as brief writing and creative development if you have the right mindset and a suite of engaging tools and ways of thinking at your disposal.

We chose our speakers to represent the cutting edge of thinking about measurement. They focused on things we might not immediately associate with measurement, and in the end there was a surprising level of agreement about what Planners and Strategists need to do and think about when it comes to measurement.

At heart it seems to come down to this:

We need to eschew single measures and look measurement/effectiveness in the round. We have the brains and hearts of storytellers, so let’s use them. And increasingly we will need to think about engineering our brands for AI readability as much as consumer engagement.

When I write up an event I usually do so thematically. In this case, the speakers had such individual and specific ideas to convey that I have given each their own ‘word doc stage’. NB The narrative from TikTok has been amended by their internal comms

Let’s draw back the curtains and hear what they had to stay

Find practical measures of meaning for your brand: Jared Tanna of Canvas8

Jared started in the wider culture, with holistic measurement, staying abreast of trends that directly impact brands and advertising.

He cast a new and in my view hopeful light on how people will be demanding greater meaning from all aspects of their life in 2026. He contended that marketing and ad people have until recently got stuck on measuring brand purpose and values, worrying about whether the products and services we marketed really added to society and the collective good.

This has turned out to be a somewhat hollow practice, based on a rather shallow understanding of what people need, and how our brands and services could contribute in a meaningful way to their lives.

His talk focused on those measures of meaning. The things people are most concerned about are not existential but more practical, physical and substantive.  This is because we have lived through a ‘meaning recession’ when big cultural forces and experiences have increasingly delivered lack of meaning. Think doom scrolling, dead loss content, fake food, Duolingo addiction - plus bigger social forces like a disconnect with real life and its milestones. Younger people delaying marriage and parenthood is a good example of this

This has inspired a counter trend of a ‘meaning boom’ where some successful brands are helping to reconstruct meaning; around eating and drinking rituals like matcha, returning to effortfulness and analogue media, anti-slop mindsets and life fulfilment over life hacks.

It’s about human qualities and values over the ‘datified’ self.

So how do you link that back to brands and measurement?

One is personal utility. Is your brand useful? What does it add to peoples’ lives? Or are you just trying to distract, meaninglessly?

Another is social utility. How can you strengthen relationships with your brand as a tool?

And positive friction. Are you giving them a chance to stop, think, reflect, and intentionally engage?  What are you doing to help them link back to how they feel as a human being?

And the big question for all brands: How much meaning can you create in the world? What sense and substance can you bring?

Share of Model: Tom Roach

Tom’s talk focused on thinking about LLMs and share of model as an essential new approach to brand building and measurement in the age of AI

AI isn’t just providing a new toolkit to create brand content, it’s building a whole new audience to create brand content for.

So we need to start thinking about how our brands will reach, be thought of and show up in the LLMs and the many kinds of generative AI content they produce.

But whilst consumers are already being influenced by generative AI content, even in its early stage forms, for many brands this influence is currently only happening accidentally and serendipitously, rather than being intentionally orchestrated.

What is happening now is that brand discovery and preference are beginning to be shaped and influenced by LLMs. 2bn people a month are using Gemini, Llama, Amazon’s Rufus is already influencing shopping choices, and of course ChatGPT is in wide use.

Some brands are reporting significant drops in website clicks of up to 30% because people are getting what they need without having to click on a link - ‘zero-click search’.

51% of consumers are increasingly comfortable with AI curating or generating answers.

67% of users prefer summarised answers vs browsing multiple links

How does it all work?

First, there’s generative search, with AI answers embedded directly in search results, which are cannibalising click traffic. The opportunity is to be cited as a credible source within these summaries. 

Next is conversational AI on platforms like ChatGPT and Gemini, where users engage in extended research sessions averaging 8-14 minutes (source: Similarweb) versus 1-3 minutes for traditional search.

Here, brands need to establish themselves as subject matter experts so their content feeds directly into LLM training inputs. 

Finally, Agentic AI will see agents autonomously research and purchase for users, bypassing traditional discovery. Success requires ensuring your content is visible and your website offers the right tooling for agents to integrate with.

So whilst brands are reporting drops in traditional search traffic, search is in fact increasing overall. And a high search ranking still matters because many generative answers start with a web search. AI isn’t killing search, it’s reshaping it.

Before AI, searches were becoming more fragmented and distributed across more platforms, surfaces and engines, and AI is accelerating this. Thinkbox the TV industry body described this as ‘SEO isn’t dying, it’s having babies’. 

And all this really is beginning to influence brand choices by consumers in the real world. 66% of US 16-24s claim they ask AIs for brand and product recommendations. And unsurprisingly older groups are a little behind but not far.

This is where GEO (Generative Engine Optimisation) comes in.

GEO shapes how generative AI systems perceive, interpret & present your brand, so that you show up as a trusted, preferred answer when people ask questions relevant to your business. It’s essentially about creating and structuring content to be easily digestible by the LLMs to influence AI summaries to feature your brand. 

And if GEO is a new form of SEO, GEM - Generative Engine Marketing - sorry for yet another TLA - GEM is the broader discipline or perspective on how you make sure everything you do as a brand - is oriented towards the LLMs in terms of thinking about them as a new audience. 

And success in this new world is more than simple web traffic, it’s about understanding brand visibility, perception, and citation rates within LLMs. It’s essentially about knowing what the LLMs are saying about our brands and looking for ways to influence them.

Jellyfish describes this as a brand’s ‘share of model’.

Interestingly, a brand’s share of mind doesn’t always tally with its share of model. Brands need to be aware that doing things right for GEO can mean you are less well adapted to building human brand awareness. Closing your human/AI brand awareness gap could be a useful objective for brands looking to take a first step. 

‘Brand’ in this world is far from being a static set of words in a strategy document. It’s a dynamic network that simultaneously builds and refreshes human memory structures while being constantly interpreted, reproduced, and represented by machines.

Brand communication always needed to build mental availability; now it's needed for model availability too. And when physical and digital availability are also strong, you have the conditions for brand choice - by both humans and agents. 

The problem with the single measure and put creativity first:

Sarah Cornish of TikTok

Sarah contended that the dominance of the single measure as a tool for strategic decision making is a disaster for brands and advertising – focusing initially on how things have historically become muddled around the use of Brand Lift studies.

You know, the ones that take a measurement amongst an exposed and control group about their impressions within 48 hours of delivery, to get a stronger read on reactions.

When used in the right way, these can be a great tool to gain a read on brands’ abilities to influence perception, but in the wrong hands can lead you up a blind alley.

In the past, a results at all costs approach to Brand Lift Studies, particularly against core measures such as Ad Recall and Top of Mind Awareness, has led to particular ‘best practices’ becoming commonplace.

In particular, 6 words became the mantra: Put your logo at the front.

What was required of course was a fundamental re-evaluation of how to communicate in these new spaces. What ensued instead was a short-term solution that guaranteed the patina of success but delivered poorly and this conventional wisdom has prevailed for almost a decade in digital and social advertising.

In attempt to disrupt conventional thinking TikTok ran a study with System 1 earlier in 2025.

It confirmed a lot of suspicions about the methodology and found that while delivering high levels of brand recognition, slapping a logo over your ad has a hugely detrimental effect on attention. And attention is increasingly seen as the key to long-term brand growth.

And the problem is broader than just confused approaches to Brand Lift.

It highlights the fact that bending the creative to fit the measurement has become endemic on social media platforms.

And as Goodhart’s Law states: When a metric becomes a target, it ceases to be a good measure. And can lead to deeply un-strategic decision making when a single metric guides investment.

What is required is a far more rounded approach to measurement, creating a 3-dimensional perspective that adds all-important context to the discussions.

This can lead to far more creative and original solutions where a platform like TikTok could (for example) deliver not just upper or mid-funnel but also in the lower funnel using TikTok shop. And there are new ways to do this effectively too such as a live catwalk, a yoga lesson or concert? Not only could you sell product but you might be able to do something for your brand. Showmanship, over salesmanship.

It's tempting to go for simple solutions in an ever more complex world.

But ultimately, it’s the job as strategists to retain that sense of clarity and simplicity, without falling foul of Goodhart’s Law.

Holistic measurement and storytelling as a road map for success: Joy Talbot

Joy spoke about the mechanics of measurement and effectiveness; what you might cast as the hard work, getting your hands dirty with data aspects of the job. Her contention was that this is just as much about story-telling and creative thinking as any other aspect of creative strategy.

It’s a gauntlet that I think we should take seriously!

She sees measurement as a holistic framework that connects all the dots across media, brand, product pricing and external factors (human behaviour, cultural tension, the search for meaning…)

This is crucial because you have to see what is going on with your brand in the context of a full business story. If you focus on isolated metrics you will make bad decisions and it will cost you sales.

There are issues with attribution modelling whatever form it takes. It’s like looking for your keys under a streetlight – you might find them, but there’s a whole area you might miss. For example, data shows that last-click attribution can miss the mark on payback by up to 200%

Econometric modelling done properly brings real rigour; incremental impact, controls of external shocks, wargaming what if scenarios. It’s the scientific backbone that makes insight trustworthy.

By contrast there are increasing numbers of poor quality MMM providers having a detrimental impact on new business conversations. These can be cheap machine learning based analysis badly disguised as econometrics. So beware and be cautious when choosing a provider.

Which brings us to the power of story-telling and the importance of working out a narrative together. If you get the story right you can turn numbers into meaning, pointing out what has changed, why it matters and giving clear recommendations about where to go next.

It’s effectively a mindset shift from a kind of passive ‘here’s what happened narrative’ to a shared roadmap for decision making.

When you have holistic measurement your story has more credibility, your measurement has impact and bringing them both together helps you unlock investment and growth.

These 4 talks taken together give us a fantastically fresh perspective for thinking about measurement in 2026.  I hope they give new impetus and ideas for your conversations with clients and confirm that measurement can be as fascinating and creative as well.

We need to eschew single measures and look measurement/effectiveness in the round. We have the brains and hearts of storytellers, so let’s use them. And increasingly we will need to think about engineering our brands for AI readability as much as consumer engagement.

Sarah Newman

APG Director

APG Noisy Thinking in partnership with Canvas8