Yes, data can produce better policy – but it’s no substitute for real-world experience

Governments like to boast that “data-driven” policies are the best way to make fair, efficient decisions. They collect statistics, set targets and adjust strategies to suit.

But while data can be useful, it’s not neutral. There are biases and blind spots in the systems that produce the data. Worse, data often lacks the depth, context and responsiveness needed to drive real-world change.

The real questions are about who decides which data matter, how it’s interpreted – and what the change based on the data might look like.

Take the Social Investment Agency, for example. One of New Zealand’s best-known data-driven initiatives, it was established to improve the efficiency of social services using data and predictive analytics to identify individuals and families most at risk, directing funding accordingly.

The model is intended to guide early interventions and prevent long-term harm. And on paper, this appears to be a smart, targeted strategy. Yet it has also faced criticism over the risk of data-driven policies reducing individuals to measurable statistics, stripping away the complexity of lived experiences.

The result is that decision making remains centralised within government agencies rather than being shaped by the communities most affected.

What data can’t tell us

The Social Investment Agency also relies on Stats NZ’s Integrated Data Infrastructure, a database of anonymised administrative information. While a rich source for longitudinal research and policy development, this too has limitations.

It relies heavily on government-collected data, which may embed systemic bias and fail to represent communities accurately. Without accounting for context, some populations may be underrepresented or misrepresented, leading to skewed insights and misguided policy recommendations.

This kind of data is completely separate from the lived reality of the people the data describes. Māori in particular have been concerned about a lack community ownership and that the Integrated Data Infrastructure does not currently align with their own data sovereignty aspirations.

Given this greater likelihood of misrepresentation, Māori and Pasifika communities worry that data-driven funding models, on their own, fail to account for more holistic, whānau-centered approaches.

For instance, a predictive algorithm might flag a child as “at risk” based on socioeconomic indicators. But it would fail to also measure protective factors such as strong cultural connections, intergenerational knowledge and community leadership.

This is where the kaupapa Māori initiative Whānau Ora provides an alternative model. Instead of viewing individuals in isolation, it prioritises the needs of families to provide tailored housing, education, health and employment support.

Change from the ground up

Funded by Te Puni Kōkiri/Ministry of Māori Development, Whānau Ora has been criticised in the past for the lack of measurable outputs data-driven systems can offer. But research has also shown community-led models produce better long-term outcomes than traditional, top-down, data-driven welfare and service delivery models.

A 2018 review found Whānau Ora strengthened family resilience, improved employment outcomes and increased educational engagement – for example, through supporting whānau into their own businesses and off social assistance.

Whānau Ora’s work strengthening community networks and building self-determination migh be harder to measure using standard metrics, but it has long-term economic and social benefits.

Similarly, data-driven approaches to disease prevention can fall short. While governments might rely on obesity rates or physical activity levels to shape interventions, these blunt measurements fail to capture the deeper social and economic factors that affect health.

Too often, strategies target individual behaviours – calorie counting, exercise tracking – assuming better data leads to better choices. But we know local conditions, including what financial and community resources are available, matter much more.

An example of this in action is Health New Zealand/Te Whatu Ora’s Healthy Families NZ division. With teams in ten communities around the country, it works to create local change to improve health.

Instead of simply telling people to eat better and exercise more, it has supported community action to reshape local environments so healthier choices become easier to make.

In South Auckland, for example, Healthy Families NZ has worked with local businesses to improve access to fresh, affordable food. In Invercargill, it has helped transform urban planning policies to expand green spaces for physical activity.

Data in perspective

Such initiatives recognise health is about more than just individuals. It is a shared outcome that results from systemic processes. Data-driven approaches by themselves struggle to capture these less measurable pathways and relationships.

That is not to say government-led, data-driven methods don’t often diagnose the problem correctly – just that they frequently fail to provide solutions that empower communities to make lasting change.

Rather than over-relying on data analytics to dictate funding, or on national health targets to guide the system, cross-sector and place-based initiatives such as Whānau Ora and Healthy Families NZ can teach us a lot about what works in the real world.

Data will always have an important role to play in shaping policy, but this requires a broader perspective. Data offers a tool for communities, not a substitute for their leadership and voice. Real system change happens when we fundamentally rethink how change happens, and who leads that change in the first place.