


Why behaviour is replacing opinion as the main signal for demand, and what leaders in innovation, product and consumer insights should do now.
22
pages
45 min
reading time
5
industries
May 2026
edition

Get the report for free
Why budget is committed long before customers act
What 32 insights leaders said at IIEX Europe
How AI, surveys and behaviour work together
A roadmap for the first six months
Form loads here
PDF, 22 pages, in English. The link is sent by email.
Foreword
Daniel Putsche
Founder & CEO, Horizon
Customer centricity does not start with an opinion. It starts with a decision.
An opinion can be asked for at any time, without a real choice. A decision has real consequences, real alternatives, a real price. The companies that succeed will not be those with the most ideas, but those that know fastest which ideas customers would actually buy or sign up for.
The decision gap
Tariffs, ranges and products are approved on the basis of stakeholder consensus and surveys. The first signal from real customer behaviour often only arrives after launch, when budget, IT and brand have long been committed.
Study
76% of consumer goods launches miss moderate sales targets. The figure has barely moved in ten years (BCG 2025).
Pulse Check · IIEX Europe 2025
82% of respondents work at companies with more than 5,000 employees. The results are consistent with the research.
Gate 01 · Method
81%
still use surveys, 75% interviews. Only 31% run tests with real behaviour.
→Gate 02 · Confidence
3.3 / 5
average self-assessment. 53% rate their methods 3 or lower, only 6% give a 5.
→Gate 03 · Outcome
31%
launched after positive feedback and performed worse in the market than expected.
The hidden finding
Validation often serves internal alignment, not a better decision.
validate to build confidence with stakeholders
44%
validate to decide faster and with more certainty
12%
Source: Horizon Pulse Check, IIEX Europe, May 2025 · N=32
The diagnosis
Surveys, interviews and conjoint remain valuable for understanding motives. As the sole basis for anticipating purchase decisions, they are unreliable, because the deviation runs in both directions. Which one cannot be known in advance.
The deviation is not a constant you can correct for. It can only be observed.
Stated above behaviour
+20%
On average, respondents state a willingness to pay around 20% higher than what they actually pay.
Schmidt & Bijmolt 2020 · Meta-analysis, 100+ studies
Stated below behaviour
−28%
In the premium segment, stated willingness to pay is up to 28% below actual willingness to pay.
Ebert, Winzer & Müller 2025
The architecture
Large insurers, retailers, energy suppliers and manufacturers are moving towards the same division of labour. AI personas simulate language, not trade-offs with a real budget. That is why purchase or sign-up intent measured in behaviour remains the benchmark.
Context
Gartner Hype Cycle 2024. Pretotyping is listed in the Hype Cycle for Innovation Practices.
AI explores
Hypotheses, personas, segments, narratives. Breadth at low cost, but no certainty about demand.
Behaviour validates
Purchase and sign-up behaviour under realistic conditions: painted door tests, offer pages, simulated checkouts. Behaviour narrows down the options.
The gatePeople interpret
Strategy, brand, second-order effects. Behavioural data shows the what, the interpretation remains a management task.

Case · Home appliances
How far can the monthly price rise before demand is lost? Virgin Pure tested it: the same ad on Facebook and Instagram, four offer pages, only the price changes. Demand kept pace up to £34.99 and only dropped at £39.99.
3,574
consumers
7 days
in field
4
prices
Add-to-cart rate per price
At +15%, the add-to-cart rate was more than twice as high as at the previous price. Virgin Pure raised the price by 15%. A survey would have shown neither the headroom nor the edge above it.

The insights provided by Horizon enabled us to implement an astonishing 15% price increase with confidence.

The consequences
When behavioural data becomes a prerequisite for investment, gates, timing and the role of the insights function change.
L1
Opinion-driven
Intuition decides, budget follows hierarchy.
L2
Evidence-aware
The gap between saying and doing is named, tests are being discussed.
L3
Experiment-driven
First tests with real behaviour, dependent on individual champions.
L4
Behaviour first
Behavioural thresholds come before every stage gate. Real demand decides go or no-go.
Three steps from L3 to L4
01
From input to gate
Set a behavioural threshold that a concept must reach before it gets budget.
02
From tool to capability
Establish test design, traffic and analysis as a discipline, in-house or with a partner.
03
From new concept to portfolio
The most expensive concepts are the ones never checked against behaviour. An audit shows the gaps.
Roadmap
The first 90 days deliver evidence. Anchoring the discipline in governance takes around six months.
Month 0 to 2
First test, stocktake
Select one or two upcoming decisions with real budget
Review the last five major decisions
Set up the first test and go live
First behavioural data is on the table.
Month 2 to 4
Repeat, define the threshold
Two to three more tests across departments
Document patterns: where does data change the decision?
Draft the stage gate threshold with the CFO and gate owners
Threshold drafted and aligned internally.
Month 4 to 6
Anchor
Make the threshold a requirement in the stage gate process
Decision: build in-house or with a partner
Set the cadence: how many decisions per quarter?
Behavioural gate and working rhythm are in place.

The most effective innovators do not ask whether an idea sounds good. They check whether customers act on it before budget is committed.
You will speak with Daniel Putsche
Founder & CEO, 30 minutes
Related