What people say, what people do: Angela Bliss and the science of behaviour

Early in the pandemic, a market researcher could have been forgiven for writing off Corona beer. Asked directly, a striking share of drinkers said the brand's unfortunate name had poisoned it for them: in one widely cited survey, 38% declared they would never buy it "under any circumstances." It is exactly the kind of clear, quotable finding that gets a product pulled or a campaign rewritten. And it was, in the sense that matters most, wrong. Sales of the beer went up.

For Angela Bliss, that gap, between what people told a survey and what they did at the till, is not a curiosity. It is the whole problem. As Head of Behavioural Science at NAB (National Australia Bank), Bliss spends her days on the premise that the tools most organisations lean on to understand their customers are quietly measuring the wrong thing.

The unreliable witness

The argument runs against decades of professional instinct. Surveys, interviews, focus groups, usability tests, the standard kit of customer research, all share a hidden assumption: that people can tell you, accurately, why they do what they do. Behavioural science says they mostly cannot. By Bliss's account, drawing on the now-familiar work of researchers such as Dan Ariely, roughly 95% of our decisions are made by fast, automatic, intuitive "System 1" thinking, below the level of conscious reasoning. We then, she writes, "make up stories that align with how we'd like to see ourselves to post-rationalise choices."

That is not a knock on customers; it is a fact about human cognition, and it applies to all of us. The practical consequence is that asking people to introspect about future behaviour, would you buy this? would you switch? how likely are you to recommend us?, collects sincere answers that may bear little relation to what happens when the moment is real, the options are many, and nobody is watching. The Corona example is memorable precisely because the stated preference and the revealed behaviour pointed in opposite directions.

"Surveys, interviews and user testing, while still helpful, need to be deployed alongside experiments that test how customers will actually behave in a service interaction."

Note what Bliss is not saying. She is not throwing out traditional research; her framing is explicitly additive, those methods are "still helpful." Her objection is to using them alone, as if a self-report were a behaviour. The remedy she presses for is experiment: designing the service interaction itself so that real behaviour can be observed and tested, rather than predicted from a questionnaire.

From experience design to behavioural science

Bliss arrived at this position by an unusually applied route. Before her current role she was Head of Experience Design at NAB, where she led a team of service, CX and UX/UI designers working on how the bank's tens of thousands of employees experienced their own organisation, internal-facing products and services, designed with data and behavioural insight rather than guesswork. Earlier still she worked as an independent CX and behavioural-design consultant (under her own Angela Bliss Consulting, and with clients including the retailer MECCA), positioning herself at the seam where design thinking meets behavioural science.

The intellectual grounding came from the London School of Economics, where she completed an MSc in Behavioural Science, the discipline that gives names and evidence to the patterns a designer might otherwise only intuit. It shows in how she talks about her craft. Where many practitioners reach for jargon, Bliss reaches for frameworks with teeth: the COM-B model, which holds that changing a behaviour means modifying at least one of three components, Capability, Opportunity and Motivation; and the B-MAP formulation (Fogg's behaviour model), in which a behaviour only happens when Motivation, Ability and a Prompt converge at the same moment. These are not decoration. They are the difference between hoping a service nudges people and being able to say why it will.

The Code of Conduct test

If the theory sounds abstract, one project at NAB makes it concrete. A corporate code of conduct is, in most organisations, exactly the kind of document the behavioural sciences were built to improve and rarely get near: a dense rule book, written to satisfy lawyers, that employees skim once and forget. Telling people the rules is a System-2 act, slow, deliberate, easily ignored. Most decisions that the code is meant to govern are made by System 1, in the moment, under pressure.

Bliss took a behavioural-science approach to redesigning that code. Rather than restating policy, the work used scientifically grounded nudges and an employee-centric lens to turn the document from a compliance artefact into something that actually shapes everyday decisions, guidance pitched at the way people really make choices, not the way an org chart wishes they did. It is a neat demonstration of the thesis turned inward: the same discipline she would apply to a customer's journey, applied to her own colleagues' behaviour.

Why a bank should care

There is a reason this matters more in financial services than in most fields. Banking is built on decisions people are notoriously bad at making in the abstract and predicting about themselves: saving, borrowing, switching, planning for a future self they struggle to imagine. A bank that designs from stated intentions, what customers say they will do with their money, is designing for a fiction. A bank that designs from observed behaviour, and tests its way forward, has a chance of building services people actually use.

That is the case Bliss carries onto the conference stage. She is listed among the presenters at The Outlook's flagship TO26 Firebrand conference, billed there in her NAB role and grouped with the designers, strategists and product leaders the event gathers. Her published work, including her LinkedIn essay "How to Combine Service Design and Behavioural Science" and a podcast appearance, "The Science of Decision-Making: Enhancing UX with Behavioural Insights", circles the same point from different angles: that good service design and behavioural science are not rivals but halves of one method.

The ask sounds modest. Stop trusting the survey to tell you the truth about behaviour; build the experiment that can. But follow it through and it reorders how a whole organisation decides what to ship. Behind the frameworks and the case studies, that is what Angela Bliss is pressing organisations to do, and the reason a bank decided it needed a Head of Behavioural Science at all.