Post-digital govt: how data and analytics will change

After 20 years focused on administration and citizen experience, digital government has evolved. It’s time for Australian government organisations to prepare for when digital itself is insufficient to deliver improved outcomes.

Many government organisations are now moving into a ‘post-digital’ era, where the business case for further digital investment reaches beyond improving experience and administration, focusing on delivering enduring, mission-relevant outcomes instead.

Gartner predicts by 2026, over 75% of governments will gauge digital transformation success by measuring the enduring mission impact, rather than only looking at hours saved, efficiencies or citizen satisfaction.

To generate critical insights for core mission needs, post-digital governments are rapidly evolving their data and analytics investments. ‘Simple’ extrapolation from past experience is no longer sufficient to address government challenges due to the increasing levels of global, economic and environmental instability.

Australian government organisations must make informed decisions using real-time data from multiple sources — whether for an individual regulatory action or to react to a complex developing situation, such as a natural disaster or a pandemic. The ability to respond flexibly is critical.

Integrating insight and prediction with real-time response

Traditional analytics approaches in government have been reactive in nature. The more mature organisations have adopted predictive, or even proactive, approaches in determining policy and guiding actions. A good example is CSIRO’s SPARK system, an open framework for fire prediction and analysis to predict bushfire spread.

Government actions can have significant impact, which includes the potential to create change in a situation. To avoid triggering secondary problems, managing the timing of any regulatory action and the potential consequences is important, whether intended or not.

Embracing empathy with citizens and other stakeholders is critical, leveraging insights to both anticipate optimal engagement and effectively react to the unexpected. Just as important is managing feedback, monitoring the impact of the action as it occurs — not months or years later.

Delivering the power of analytics at the operational sharp end, with critical actionable moments, is only just becoming possible with emerging technologies.

Using multisource data

Most governments currently don’t have large enough volumes of data to create value for their missions. Outside a narrow set of specific use cases, government data is complex rather than extremely large.

Historically, data collection has been very costly, with different parts of government collecting data for separate purposes in varying formats. This is one of the challenges that makes sharing government data difficult. Aggregated data may provide more value, but it increases the cost of management and the risk of data exposure.

Post-digital governments acknowledge this and use multisource data — some complex data in smaller volumes and some higher volume, less complex data captured as a by-product of normal operations.

As examples, Finland’s Real Time Economy project and TicketBAI in Spain are at the upper end of this range. They aim to collect retail-level transactions and use these to fulfil several capabilities, including transactional taxation, real-time financial reporting, compliance by design and so on. Australia’s Single Touch Payroll system provides a similar function for payroll data.

This approach is unlikely to be feasible in all economies. Many governments will need corroboration of data between sources to enable a real-time view even with gaps in data. This was in part the approach taken by the U.S. Department Health and Human Services’ HHS Protect project during the early phases of the pandemic.

Using multiple techniques and tools

Aggregating data may provide a scale base for machine learning, but it can have significant flaws. Even with the highest scale approaches, such as those used to support large language models like OpenAI’s ChatGPT, there are inaccurate results — ‘hallucinations’.

In a changing environment, scale data will not be available in the early stages of a developing situation. An illustration of the difficulties faced by governments in responding to natural disasters came in the wake of the 2022 floods in the NSW northern rivers.

An example of a post-digital response might be the use of parametric insurance — instead of attempting to process claims for support after a disaster, payouts are automatically calculated based on the likely needs of the relevant population.

This requires pre-analysis of impacts on a community. Payments can then be triggered easily by a much smaller and simpler analysis at the time of the disaster, such as river levels reaching a certain height, significantly reducing the administrative burden and delays in economic and community recovery.

This is an example of rebalancing the natural urge for increased data for more precise decisions with a need for quick decision-making to alleviate problems and achieve the policy result.

To address these risks requires an approach that combines data analysis with human intelligence, the use of predictive analysis and local operational support analysis. This combination has the effect of increasing the timeliness and the value of any regulatory action, without forgetting the human element.

Ultimately, for government organisations to deliver on their mission in the post-digital era, data collected from regulatory action must close the loop. No government response will ever be perfect.

The key factor is that governments use their advanced data and analytics capabilities to achieve the desired effect. Also, use near real-time feedback to enable continuous tuning of the action to optimise a government organisation’s ability to deliver the desired policy outcomes with the resources available.

Image credit: iStock.com/olaser