TL;DR SUMMARY: The word “nudge” has been diluted into a catch-all for prompts and persuasion, yet a genuinely effective nudge is something more precise: contextual, behaviourally informed, choice-preserving and delivered at the moment it can help someone make a better decision.
What if almost everything we call a nudge today isn’t really a nudge? The notification on your phone, the brightly coloured button on a website, the reminder in your inbox, the pop-up asking you to subscribe, the suggested upgrade at checkout and the message telling you that “people like you” have already made a particular choice are all being gathered under one convenient word.
What happens when that word becomes so broad that it stops telling us anything useful?
The term nudge has travelled a long way since Richard Thaler and Cass Sunstein popularised it through their work on behavioural economics and choice architecture. In Thaler’s own account, a nudge is a feature of the environment that influences behaviour while preserving choice. The architecture changes, the options remain and the individual remains able to choose.
That distinction matters when we’re trying to understand what makes an intervention genuinely useful. Somewhere between the behavioural science laboratory and the modern product roadmap, nudge has become a convenient label for almost any intervention designed to move somebody towards an outcome.
A reminder becomes a nudge, a banner becomes a nudge, a prompt becomes a nudge, a sales message becomes a nudge and a notification becomes a nudge. Gradually, the concept becomes diluted.
So, what actually makes a nudge a nudge?
Perhaps the simplest way to understand the problem is to return to the idea of choice architecture. We rarely make decisions in a vacuum. We make them inside environments designed by somebody else, whether that’s a supermarket, a website, an airport, a pension scheme or a cyber security system.
Every one of those environments contains signals, defaults, friction, prompts, sequences and social cues. Some of those features are deliberate, while others are accidental, and they can either support a person’s decision-making or simply make the behaviour somebody else wants easier.
The important question is therefore less “Does this influence behaviour?” and more “How is behaviour being influenced, and what happens to the person’s ability to choose?”
That brings us back to the original idea of the nudge. Thaler describes nudges as features of choice architecture that influence human behaviour without materially changing incentives or removing options. A default pension enrolment is a classic example because the default influences what people do while the alternative remains available.
A nudge can be small and almost invisible. Its sophistication lies elsewhere, in understanding when, where, why and for whom the intervention might work. That’s behavioural science doing useful work.
The fly in the urinal tells us something important
Consider one of the most famous examples. At Amsterdam Schiphol Airport, a small image of a fly was etched into urinals, giving people a simple target and encouraging them to aim more accurately. The story became famous because of its apparent simplicity, and Redflags itself has used the Schiphol example when explaining nudge theory.
The interesting part isn’t the fly itself, though. It’s the context in which the intervention appeared.
The cue was placed in a location where a particular behaviour was already taking place. It didn’t require a training module, a policy document or someone remembering a lesson from six months earlier. The intervention appeared at the point of action, when it could influence the behaviour directly.
That distinction is easy to miss when nudge becomes a synonym for prompt. An intervention isn’t necessarily useful because it’s clever in isolation. Its value comes from how well it’s connected to the behaviour, the environment and the moment.
From System 1 to System 2
There’s another reason context matters. Human decision-making is often described through the distinction between fast, intuitive processes and slower, more deliberate ones, commonly associated with Daniel Kahneman’s System 1 and System 2 framework.
System 1 is fast and automatic, which is often exactly what we need to navigate everyday life. It’s also where many of our shortcuts, assumptions and biases operate. System 2 is slower and more deliberate, asking us to stop, inspect and reconsider what we’re doing.
A well-designed nudge can create a small bridge between the two. It can create a moment in which someone pauses, make relevant information visible and bring a consideration into the decision at precisely the point when that consideration matters.
That’s a very different proposition from simply telling someone what to do. The aim can instead be to give people more agency over a decision by helping them notice something they might otherwise overlook.
A useful intervention might help someone notice the sender of an email, recognise a default, consider a deadline, spot a healthier option or identify a potential risk. The person can then decide what to do with that information.
Healthcare shows what contextual nudging can look like
Healthcare offers some of the clearest examples because timing and context can be extraordinarily important. A reminder about an appointment is ordinary communication, while a behaviourally informed reminder can take into account what people are likely to forget, when they’re likely to act and what information they need at that moment.
UK public health programmes have used behavioural insights to improve uptake of NHS Health Checks, including reminders, prompts to clinical staff and behaviourally informed communications. Government guidance has also highlighted the importance of delivering prompts when people are receptive to them, recognising the influence that surrounding context can have on whether someone acts.
Research has gone further, with randomised trials examining behaviourally informed invitation letters and SMS reminders for NHS Health Checks, alongside research testing behavioural message content in cervical screening invitations.
The lesson isn’t that every text message is a nudge. The design of the intervention is what matters, including what is being asked, when it’s being asked, what the person already knows, what friction exists, what they might forget and what information could help them act.
The question of choice remains equally important. A useful intervention can support someone’s decision-making while leaving the decision itself with them.
Those questions take us back towards behavioural science, where the focus is on understanding how people actually behave within the environments and circumstances in which decisions take place.
The dilution happens when the label becomes the strategy
This is where the word nudge starts to become problematic. Once the label becomes fashionable, it can become tempting to work backwards from the intervention.
“We need a nudge,” so someone designs a pop-up. “We need another nudge,” so another notification appears. “We need to increase engagement,” so another prompt is added. “We need people to remember this,” so another reminder is sent.
Soon, the organisation has a collection of nudges and the person has a collection of interruptions, while the original behavioural problem remains.
That’s how nudge fatigue begins.
Redflags makes a similar point in its own approach to cyber security, recommending focused campaigns around a small number of behaviours and allowing the frequency of interventions to be adjusted. Its approach is designed to make interventions relevant to actual behaviour rather than turning every possible moment into an opportunity to display a message.
A nudge should have a reason to exist. It should be based on a behavioural hypothesis, appear in a meaningful context, be tested and refined, with its success ultimately understood through changes in behaviour rather than the number of messages delivered.
Cyber security may be where the distinction matters most
Now consider the modern workplace. An employee receives an email while they’re busy and has a meeting in four minutes. They recognise the sender’s name, hover over a link and are about to click.
This is a very different behavioural environment from an annual cyber security training session. The training may have been completed, the policy may have been read and the phishing simulation may have been passed, yet the decision is happening right now.
That’s precisely where contextual nudging can become interesting. Redflags describes its approach as delivering real-time, targeted security prompts at the point of risk, helping people pause before potentially unsafe actions such as interacting with unfamiliar links, sharing sensitive information or leaving a device unsecured.
Its case study with Accenture describes a deployment in which prompts were configured around specific behaviours, including phishing susceptibility, device security and password practices. The interventions were delivered within the desktop environment, where the relevant behaviour was taking place.
That’s a useful illustration of what makes a nudge more than a notification. The intervention is connected to the behaviour, the moment and the individual’s context, while the person remains the decision-maker.
The alternative is familiar. A warning arrives every Tuesday at 10am, a training module explains phishing and a poster says, “Think before you click.” Each of these approaches may have value, yet none necessarily arrives when the decision is actually being made.
A contextual intervention can work differently. The risky action appears, the cue appears, the person pauses and the decision becomes visible. That small pause can create a space between automatic behaviour and deliberate choice.
The future of nudging may be less about persuasion
This is perhaps the most interesting misconception to challenge. People sometimes hear nudge and think manipulation, imagining a hidden hand or a system quietly steering somebody towards an outcome.
There are legitimate ethical questions around influence, choice architecture and who decides what constitutes a desirable behaviour, and those questions should remain part of any serious discussion about nudging.
The original concept also contains a useful counterweight through the preservation of choice. A good nudge can help somebody see something they might otherwise miss, make an option easier to understand, reduce unnecessary friction, surface information at the right time or create a moment to think.
That makes the conversation less about “How do we make people do this?” and more about “How do we help people make the decision available to them at the moment it matters?”
That’s a much more interesting question, and perhaps a more human one.
Recovering the meaning of nudge
So, is everything we know about nudges wrong? Not quite. The problem is that we’ve started using one word for too many different things.
We’ve taken a precise behavioural concept and stretched it across product design, marketing, healthcare, public policy, workplace technology and cyber security until almost every prompt can be described as a nudge.
The result is conceptual dilution, while the opportunity is conceptual recovery.
A nudge is most interesting when it’s embedded, contextually aware, behaviourally informed, timely and choice-preserving. It sits inside the environment where the decision happens and understands something about the behaviour it is trying to influence.
For cyber security, that means thinking beyond the idea of simply putting more security messages in front of people. The more useful question is whether technology can recognise when a security-relevant decision is taking place and provide the right information at that moment, while still leaving the person in control.
That’s where behavioural science can offer something genuinely practical. The aim isn’t to add another layer of prompts to an already busy working environment. It’s to understand behaviour well enough to know when an intervention could help, design it around the context in which the decision happens and give people a better opportunity to make that decision themselves.
Key takeaways
- A nudge is more than a prompt. Its value comes from being behaviourally informed, contextual and timely, while preserving the person’s ability to choose.
- Context is what gives an intervention its power. A cue delivered at the point of action can be far more relevant than information someone was given weeks or months earlier.
- More nudges don’t necessarily mean better behaviour. Without a clear behavioural purpose, frequent prompts can become interruptions, creating fatigue rather than encouraging safer decisions.
- Cyber security can make nudging genuinely practical. Context-aware interventions such as those used by Redflags can create a moment to pause and reconsider, helping people make safer decisions without taking the decision away from them.
