Signals, noise, and why accountable data and diverse planning are the only way to ‘invest with confidence’
Opinion
Media shouldn’t be seen as in competition but rather as an orchestra. Where the greatest planners and strategists identify every emotion, nuance and contribution each musician brings to the bigger whole, writes the IPA’s Simon Frazier.
There’s a familiar feeling in media right now: the faster the landscape evolves, the more we crave certainty. In a world of dashboards, attribution windows and AI-generated “answers”, certainty can feel like it’s only a click away.
Back in 1997, Clippy would pop up in Microsoft Office and confidently tell you he knew what you were trying to do. He rarely did. Fast forward to 2026, and if we’re honest, some media dashboards feel suspiciously similar, offering more data and confidence but not necessarily more truth.
That’s the central tension in Signals in the Noise 2. We’ve never had a better scientific understanding of how advertising works: mental availability, memory formation, attention, reach and frequency, the long and the short, and the 95/5 rule, which challenges many hyper-targeted approaches. But we’re trying to apply it in a data environment where the loudest numbers often come from the platforms selling the space, using bespoke definitions and opaque methodologies only they fully understand.
Or, to put it less politely: if a platform sells the media, defines the yardstick, and builds the tool that “proves” it worked, we shouldn’t be surprised when the recommendation is… more platform investment.
It’s a bit like judging a Bake Off final where each contestant brings their own scales and the output of a Betty Crocker cake mix is put on the same level as the genius creation of Nadiya Hussain.
Rubbish in, rubbish out or why Bertha the machine was right
If you grew up on kids’ TV in the 80s, you’ll remember Bertha (alright, you might not, as it only ran for one season, but it taught the vital skills of life), the magical machine that could make anything… provided you fed it the right inputs, and knew how to work with Bertha. The same rule applies to modern measurement: models are only as good as the data that feeds them and the people who know their limitations.
MMM is having its well-deserved comeback, and rightly so. But if your exposure data is inconsistent, incomparable or commercially conflicted, then the model just gives you a more sophisticated version of the wrong answer. Sure, since you’ve shown your work, you might not get fired for a poorly performing decision, but is that really what we should be aiming for?
That’s the real issue flagged in Signals in the Noise 2:
* Incomparability – a “view” isn’t a universal currency anymore
* Short-termism – we optimise what’s easy to measure, not what actually works long-term
Or, to put it in more modern terms: it’s like optimising for TikTok-length answers to problems that behave more like a Game of Thrones plotline – messy, slow, and driven by things you can’t always see immediately.
Enter the reality check we actually need
This is where Making Sense 7, the JICs, and IPA TouchPoints play a critical role.
At a time when every platform claims to have the answer, TouchPoints does something far less flashy but arguably far more useful: it shows how people actually live their lives with media.
Despite everything – AI, smartphones, streaming, whatever the metaverse was supposed to be, the patterns of media consumption haven’t fundamentally changed in 20 years. We still wake up, listen to audio in the morning, consume text and OOH during the day, and watch video in the evening, and that happens across a far broader range of channels than most platform narratives suggest.
The essential challenge today
The difference is how we do those things. And that creates fragmentation – lots of channels, lots of touchpoints, none of them dominant alone. Which leads us to the bit the industry sometimes resists, that fragmentation isn’t the problem, monoculture approaches are.
The great Sarah Gale once said, “As the media landscape fragments, diverse media planning is absolutely essential for delivering results.” And being in the privileged position of co-hosting the IPA Making Sense podcast alongside the brilliant Molly Bruce and interviewing some of the greatest planners and strategists in the industry today including Rich Kirk, Caroline Manning, Vicky Fox and many more – the unifying factor we’ve observed is they all subscribe to the principles of this fundamental sentiment, with an absolute understanding of the people they are trying to reach over just the platforms they choose to do so, at the core.
Think less “one hit wonder” and more Glastonbury lineup. You don’t go for one act – you go for the combined experience. You could see Dolly Parton in concert for a lot less than the £400 Glastonbury ticket price, but what if she’s having an off day?
Whereas with a festival approach, the upside risk of not having a great time is greatly reduced when so many other artists are thrown into the mix, whereby even if the headliner is disappointing (Dolly never would be, so probably a bad example), you can be almost certain that your investment will net out to an overall positive experience, that single act investment just can’t guarantee.
Different channels do different jobs:
- TV and high-attention environments build long-term memory
- Digital “lots of littles” build reinforcement across moments
- Audio and OOH wrap themselves around people’s lives
And crucially, they work better together in harmony than alone.
Or, we say in Making Sense, media shouldn’t be seen as in competition, but rather approached like an orchestra. Where the greatest planners and strategists work like Michael Kamen, Daniel Barenboim and Simon Rattle, identifying every emotion, nuance and contribution each musician brings to the bigger whole.
AI, efficiency and the danger of becoming a tribute band
There’s another trap here: the shift towards pure efficiency.
AI makes it easier than ever to optimise, automate and scale. But if we’re not careful, we end up in what the great Roderick of Sutherland puts it: “the exploit mindset,” chasing short-term metrics, repeating what worked last time, and turning creativity into a cover version of itself.
And as any music fan knows, tribute bands fill pubs – but they don’t move culture.
The alternative is the “explore mindset”: test, learn, diversify, and crucially question the data you’re feeding into the system. And while creativity and finance are often seen as opposed, the most successful VCs and entrepreneurs don’t put all their investment eggs in one basket but hedge across diverse opportunities.
Back to signals
So, where does this leave us?
We’ve got:
- The best understanding of effectiveness we’ve ever had
- The greatest minds working in the industry today
- The noisiest, most fragmented data environment we’ve ever had
- And more tools than ever, trying to tell us they have the answer
The answer is actually pretty simple, just not necessarily easy: use better inputs, make broader plans and approach everything with more scepticism. Embrace your inner five-year-old and never stop asking why?, why?, why?, why? And one more why? Just for luck.
Invest in accountable, comparable data (and yes, that means leaning into quality joint-industry measurement – after all, its methodologies are Joint Industry for a reason). Use TouchPoints-style reality checks to ground decisions in how people actually behave and build diverse media plans that reflect fragmentation rather than fight it
Because in the end, effectiveness hasn’t changed. It’s still about reaching people, building memory, and showing up in the right moments.
Everything else? That’s just noise with a nicer UI.
TouchPoints celebrates 20 years of investment, innovation and guidance this year. Come along to celebrate on the 1st of July.
Simon Frazier is the associate director and head of marketing, data innovation and The Making Sense unit at The IPA.