Want more practical insights on understanding what customers really need?
Learn more about UX research, product strategy, and evidence-led decision-making.
Dr Nick Fine
Published on 9 September 2026
How to discover and develop what people really need
Establishing what people really need – and will pay for – is essential in product development. But it's often something teams can lose sight of. The challenge isn't having ideas or even shipping them quickly, it's keeping sight of those needs. The good news? There's a well-established approach that helps individuals and teams explore, evaluate and communicate to create effective products.
If your teams are creating products based on hunches, guesswork or even the HiPPO (highest paid person’s opinion), they’ll likely end up in the wrong place. All kinds of ideas can be a starting point for a product. However, no matter how quickly and effectively they ship, if it misses the mark for customer need, it won’t work commercially.
In my experience, the ability to test, learn and adapt requires the right approach and mindset. Operating within a lightweight framework gives individuals and teams a way to explore, evaluate and communicate around the ideas and insights that become effective products. Fortunately, there is a well-established approach that is perfect for the job. The scientific method.
Putting learning at the heart of the process
Being scientific isn’t difficult and need not be bureaucratic. However, it really does help and support us to navigate to the best ideas and the most effective solutions. It’s about putting learning at the heart of the product development process. Why is that important? Because when we assume we have the answers, when we’re reluctant to validate our ideas, we’re at the biggest risk of messing up.
It’s easy to make assumptions or jump to conclusions. In fact, evolution has primed us to do just that. The result is a range of potential biases that can send us off course. Science provides a way to step outside those biases.
Applying some scientific method
When we admit that we don't know something, we bring an open mind to a problem. That means we can observe and learn without any preconceptions or biases getting in the way. Confirmation bias – the tendency to search for, interpret, favour and recall information that confirms our pre-existing beliefs - is perhaps the biggest gotcha in product development.
So how do we do this? The scientific method gives us a cycle of research-design-test that works as both a mindset and a practical framework. Rather than a rigid process, think of it as a set of lenses through which to view product development. The core elements are straightforward:
- Observation: There are two basic approaches to research. Surveys and reports of problems, and actual observation of them. Observation is more reliable because we’re seeing what people really do. All the evidence shows that people don’t always do what they say they do or think what they say they think. Observation tends to be at the heart of making the right choices in the product development process, helping us gain genuine insight.
- Hypothesise: It’s okay to have a hypothesis based on reasoning or assumptions. However, it’s important to test them and reshape them based on observations and data to avoid confirmation bias. A hypothesis is a synonym for an idea and that’s a good way to talk about them. We think some people will need [something], but we need to establish a) exactly what they need, and b) how much they need it (i.e. whether they will pay for it). If the idea is proven to be needed and valuable, we can move to the next level of refinement. Our ideas become the product roadmap.
- Experiment: The key to finding the right answers is to work quickly. Test ideas and learn at speed with disposable, low-cost prototypes. We don’t need to build an MVP or a fully-working product – not least, because that’s too slow and expensive. One of the unsung benefits of AI is that we can test ideas through medium or even low-fidelity prototypes. Remember: we’re testing an idea, not the real thing.
- Analyse and learn: Review the data collected by the experiment. If the test is finding out what relevant people think in an interview, then the analysis is looking for the themes and reviewing the evidence. If it’s more of a performance-based test, get some more definitive results (were people able to do X, how quickly or easily, for example). Scientifically speaking, we don't prove hypotheses, we provide support for them (or not). We assign a level of confidence to them.
- Conclusions and communication: Arriving at an interpretation of the data and what it means for the hypothesis gives us a sound basis on which to proceed. Of course, it also means we have something structured and reasoned to share with the team and stakeholders. Making insights easily understood and shareable pays dividends. It's a critical researcher skill and activity.
How fast can we go?
Some might worry that following a series of steps like this might slow the process down, like it’s bureaucracy. However, it absolutely can – and should – be done quickly. Have an idea, test it, analyse the data, write it up and share it. Much quicker than heading off in the wrong direction, building something based on hunches and then seeing what users think of it.
We don't need to give everyone ‘scientific’ training; we just need to formalise a simple form of Agile that has those component steps. When we're testing frequently — say, every week — we can’t help but get the feedback that helps us course-correct.
Making the approach stick
My advice is to build out from ideas of Agile development. This is simpler than Agile, but it aligns with that mindset. However, it can be incredibly hard to effect a change like this without executive buy-in. If we want to do any kind of transformational change, people need permission from the top. Why? Because using data to validate ideas - before building things - tends to challenge established decision-making habits.
As well as increasing the chances of creating the right outputs - products that customers need and value - the biggest benefit of this approach is to help people, teams and organisations to learn, be productive and move in the right direction. Good evidence is the difference between winning and losing.