Differentiated proposition and compelling claims
TThe tools and techniques we use to turn insight into commercial propositions
The principle
Strong insights don’t optimise themselves into winning propositions. They have to be worked.
The data on this is unambiguous. Kantar’s analysis of 7,300 launches found that strong propositions achieved 47% distribution share, against 16% for the failures. Nielsen BASES found that propositions delivering on performance claims are 15 times more likely to succeed. And McKinsey and Nielsen’s joint Innovation Practice work has shown that optimised concepts generate 38% more revenue and are four times more likely to succeed than unoptimised ones.
That 4x multiplier is the single biggest lever in the whole NPD process — and it sits at the proposition stage, not the ideation stage. Most innovation programmes invest the opposite way round.
The question we ask, on every project: how do we get this proposition from “interesting” to “irresistible” before it goes to the buyer?
The human work that doesn’t change
There are three things that build winning propositions, and we don’t believe any of them can be outsourced to a tool.
Category-growth workshops. Two days, the right people in the room, structured around the consumer tensions identified in the insight phase. The job is to generate propositions that resolve a real felt need, build category-incrementality from the start, and have a margin story that protects retailer economics.
Insight-focused brainstorms with structure. Most ideation sessions fail because they’re unstructured. We use tension-led prompts, category-disruption frames, and consumer-language constraints to force the team off the obvious answers. The output is a smaller number of stronger propositions, not a longer list of weak ones.
Proposition architecture. A winning proposition needs four things working together: a consumer story that resolves a real tension, an incrementality story that earns distribution, a margin story that protects retailer economics, and some form of technical insulation against private-label copying. We build all four into the proposition from the insight phase onwards — not as a reverse-engineering exercise at sell-in.
Skip these and you’re sending undercooked propositions into expensive quant testing.
How we optimise at this stage
This is where the 4x multiplier is built. Optimisation isn’t testing — testing asks “do you like this?” Optimisation asks “what would make this irresistible?” We run structured co-creation sessions at the proposition refinement phase, using stimulus generated in the workshop and reaction loops with target consumers, until the gap between “interesting” and “irresistible” has closed. McKinsey and Nielsen’s joint Innovation Practice data is unambiguous: this is the work that turns a strong concept into a winning launch. Skip it and the proposition that goes to quant testing is materially weaker than it needed to be.
Where AI is now genuinely useful
We haven’t replaced the workshop. We’ve rebuilt it.
Live visual stimulus in the room. This is the change that has reshaped how we run category-growth workshops more than anything else. Where we used to brief a designer for two weeks to mock up six pack directions, we can now generate forty in an afternoon — and react to them, refine them, and kill them in the same session. The tools we use: Midjourney for aesthetic exploration, Flux for photorealistic pack mockups, Ideogram for any concept that needs legible typography, and our own Wan 2.2 capability for bespoke moving-image stimulus. The implication for clients: the proposition the team commits to at the end of the workshop has already been visually stress-tested, not just verbally agreed.
Semantic landscape mapping. Before the workshop, we now systematically analyse the category’s existing language — competitor websites, packaging, reviews, ad copy, social conversation — to identify where the white space genuinely sits. This used to be a strategic-planning instinct call. It’s now an evidence base. We use a combination of custom Claude-based pipelines and the narrative-intelligence layer in Pulsar and Brandwatch.
AI as an ideation partner, used against structured prompts. This is a methodology rather than a tool. Used carelessly, AI generates regression-to-the-mean ideas — bland, category-conformant, indistinguishable from what’s already on shelf. Used systematically against insight tensions, with constraints on consumer language and category disruption, it produces a usable proportion of provocative propositions that humans then refine. The key insight: AI is a generator of stimulus, not a judge of merit. The team in the room is the judge.
A note on the co-creation work specifically: this is one of the places we deliberately keep AI out of the room. The optimisation conversation is between humans and target consumers, full stop. AI generates the stimulus that goes into the session; humans run the session itself.
Where we’re cautious: claim optimisation tools
Platforms like Anyword and Persado promise to optimise marketing claims using machine-learned language patterns. They have real applications — but mostly in B2C performance marketing where success is measured in click-through rate, not in FMCG buyer meetings.
Our concern: these tools optimise for engagement signals, not for category truth. A claim that wins on click-through can still die in a buyer meeting, fail to convert trial into repeat, or actively undermine the proposition’s connection to the underlying tension. We use them as one input among many, never as the optimisation logic itself.
What Brand Development is doing today
The workshop is where most of our AI capability is now visible to clients. We bring live stimulus generation into the room, we run the semantic landscape analysis as a matter of routine pre-work, and we use structured AI-augmented ideation against the tension prompts as part of the session itself.
What this means in practice: the workshop output is sharper, more visual, and more pressure-tested than it was three years ago — but the people, the time, and the rigour are the same as they’ve been for thirty years.
The lever it pulls
What used to be a six-week proposition-development sprint can now be a three-week one, with stronger visual stimulus, a clearer semantic landscape, and a more provocative starting set of ideas.
The 4x success multiplier from optimised propositions doesn’t change — but the cost and time of getting there does.
Next
[Read about Pillar 3: Commitment to drive trial and support into Year Two →]
[Or talk to us about a proposition project →]