Given all the hype about a "digital revolution", how come the complex organisational structures, functions, processes and systems of government remain largely as they were decades ago? It can't be through any lack of ambition or vision - politicians have understood since the early 1990s the potential role of technology to help reform and improve our public services.

The sheer scale of this digital transformation - the redesign and re-engineering of government on every level - presents some uniquely challenging problems. That's why in the late 1990s the business case was made for a transformational switch, away from traditional department-focused silo working towards a cross-government approach. Without this common approach, each part of the public sector would pointlessly duplicate expenditure, processes and systems - and continue to provide a fragmented citizen experience.

From 2000 onwards, a series of common platform components were developed spanning content management, identity, transaction handling, payments, secure messaging and alerts. They formed part of a cross-government IT architecture constructed on open, interoperable standards and the design of services focused on the citizen - not the owning departments. These components could be utilised by all government departments and agencies, making it easier to stitch together the right blend of services to meet their specific needs. It would be easier, less expensive and more efficient for government to transform and improve its services.

Yet this transformational switch never really happened. So what went wrong? In reality building a set of central platforms was never somehow magically going to transform government. The components failed to transfer across at scale in the form of improved organisational processes and functions. Such a failure is not unique to government - many commercial start-ups often make the same mistake: they develop a "build it and they will come" group-think and create some new shiny technology, only to find users never take it up.

This is why Tim O'Reilly's 2009 vision for "Government as a Platform" (GaaP) is potentially significant. It broadens the focus away from defining a set of narrow "things to be built", recognising that platforms themselves have little or no intrinsic value: value is only created when users interact. Instead it emphasises transparency, participation, and collaboration and the ability to identify and define the value to be created or consumed.

Successful implementation of GaaP will require innovative discovery processes, first of all identifying where and how value can be better created for citizen and government alike. This will involve mapping the various ways that different functions and processes work and the extent of their duplication, assessing how they might be improved, streamlined and aligned, identifying and implementing organisational improvements and understanding the impact on existing information systems. Only later can decisions be taken about what technology might be needed and how best it can be provided.

The smart adoption of GaaP could finally help produce the fundamental transformation of our public services foreseen back in the 1990s, providing the opportunity to nurture and grow an entirely new operating model for government. However, if GaaP simply becomes a list of technology to be built, it will illustrate once again that those who fail to learn from history are doomed to repeat it.