A recently released report auditing myGov has made a number of findings and recommendations for the future development and improvement of myGov. The report identified a number of positive aspects of myGov, however, a number of severe shortcomings were also identified, particularly around personalisation and the orchestration of user journeys. The report has been widely discussed in the media and wider web circles. In this article, I’d like to address the results of the adoption of Adobe’s product, especially as it relates to personalisation.
Background
In September 2022 Bill Shorten announced the formation of a panel of experts, headed by David Thodey AO, to conduct an “expansive user audit of myGov”. The audit was intended to “identify the changes needed to improve myGov’s reliability, functionality and deliver a more user-friendly experience” as well as to provide a roadmap to the future.
For those interested in the source documents please refer to the final report. The Report summary (PDF) myGov User Audit January 2023 provides the top level findings. However, much of the relevant detail is contained in Volume 2 (PDF) Detailed analysis of the myGov user audit which identifies the “numerous challenges” to the current state of myGov.
Personalisation use cases
The report concludes that the personalisation goals of the project have not been met.
Life journeys
One of the goals of the project was to improve citizen’s user experience for undertaking key life journeys. Finding 6 of the report state that myGov has fallen short:
myGov could be considerably smarter and more helpful; trust and autonomy for Australians will continue to be critical.
Today’s reality is a long way from that vision. Some progress has been made – notably around the birth of a child – but life-event services remain more hope than near term expectation. Part of the problem lies in the challenge of coordinating multiple agencies across multiple levels of government. But part also lies in recognising that issues which look small when sitting inside government, loom much larger for individual Australians when dealing with government.
It would appear that very little progress has been made in stitching together user experiences across agencies and services.
Personalisation based on profile
In the detailed report it was found that the personalisation technology backing the system has been unable to find traction:
Services Australia has bought a cloud software product, Adobe Target, for tailoring myGov’s content and functionality based on an ‘enhanced profile’ of a user. This software has not yet been used for live users. myGov has developed a ‘personalisation strategy’, but the audit was not made aware of any specific use cases identified for the tailoring functionality.
The failure to identify any use cases for personalisation is surprising as the choice to go with Adobe Target presumably would have been made based on its ability to deliver on content personalisation.
A background to the thinking and approach taken by Deloitte and Adobe can be found in A blueprint for enhanced citizen experiences. The blueprint outlines a number of approaches to personalisation based on user properties which can be derived from user behaviour. The document also presents a case study where a user undertakes the first part of the journey on an (anonymous session) Government website. The blueprint clearly identifies the user flow (from anonymous to authenticated), however, there is no clear articulation of how content might be personalised based on the profile.
A generalised recommendation engine is possibly not the best technology for orchestrating user journeys for key life events. Wiring together a journey requires logic, rather than a process of building a user profile based on their history. Successfully mapping out journeys will therefore involve an analysis of services, eligibility and user properties. This will involve domain experts across departments to define what the journeys should be and how they are orchestrated. In this light Adobe Target is not the best tool for defining user experiences for authenticated users.
No legislative basis
The report goes on to note that there is no legislative basis for the collection and retention of user data.
Currently, there is no legislative basis that underpins myGov’s collection and retention of information about individuals. This limits the extent to which personal information can be stored, regardless of whether consent has been provided. In the absence of a legislative basis, the protections for Australians are inherited from existing privacy legislation and are not tailored to the myGov environment. Without a clearer legislative basis, the future collection or storage of data in myGov, including for tailoring, will be limited.
This represents a fundamental shortcoming. It would appear that a clear understanding of the legislative basis for the retention of data would have been a foundational concept for any improvements made to myGov. Presumably this aspect can be improved though clearer legislation and guidance on what can legally be collected.
Where does GovCMS fit in?
The InnovationAus article, cited above, makes a case for GovCMS being “shunned” as a suitable platform for myGov. This can be considered a slight overreach. However, it doesn’t mean that GovCMS should be a natural choice for delivering myGov. As currently configured, GovCMS is suitable for hosting client-facing “anonymous” digital experiences for citizens and as such is not a suitable fit for the delivery of authenticated sessions such as those required by myGov.
GovCMS can play its part in the broader landscape for the orchestration of user journeys and personalisation. The key insight here is that user journeys will often start from agency websites (for anonymous users) before moving through to an authenticated system such as myGov where there are strong controls on user identity and privacy. This is where GovCMS can play a key role - GovCMS sites can be uplifted to be more integrated into this user flow.
The Blueprints document from Deloitte identifies this process in one of its examples of a user being guided through a process which involves learning and refinement resulting in the preparation of documents which can then be submitted in an authenticated session. Improving the “pathway” from agency website to an authenticated system will be a requirement for many websites which deal with citizens and industry as they navigate regulatory frameworks. This is an area where GovCMS sites, and indeed any government website, can be uplifted to better support user flow.
Avoiding privacy concerns
It is important to note that personalisation and better user experiences can be achieved for anonymous user sessions such as those on GovCMS. Privacy concerns can be avoided by not collecting, retaining or transmitting any personal data about the user. Personalisation systems, and tools such as decision trees, can be designed to operate so that non PII user data is only retained on the client machine. Avoiding the syncing of data to a CRM is one way to protect individuals’ privacy as well as avoid any risks of other user’s spoofing the identity of another.
Such an approach has been taken at Morpht, where anonymous personalisation can take place without the need for a centralised data store. We have taken a privacy first approach where a user profile is built up over time and stored in the localstorage of the user’s device. There is no need to share this information to third parties.
See our demonstration sites at Convivial for Civic and Convivial for GovCMS where a number of personalisation approaches are being taken to help support the user journey. We have also implemented “decision trees” which are able to lead users through regulatory requirements, landing them on an “outcome” which can then be used as a springboard into authenticated systems such as myGov or others.
Conclusion
myGov plays an important role in the digital lives of Australian citizens. The recently released report has identified a number of positive aspects as well as a number of areas which need fixing or improvement. The report has pointed strongly to failures in personalisation efforts and the orchestration of user journeys for key life events. In particular, the technology choice of Adobe Target was identified as a tool which has been unable to find traction.
The problems faced by myGov can be put down to a number of factors including the complexity of Government services, the lack of supporting regulatory frameworks around privacy and technology choices. There is a need for improvements to the system as it evolves into the future.
The tender process for the myGov project has been questioned, with a technology choice being determined before going to market. This necessarily led to a limited pool of respondents who were familiar with the Adobe platform. For important projects such as myGov, a more open tender process would encourage a wider range of solutions which may have been better suited to delivering on the requirements of the project.
In this article I have identified areas where future improvements could be made:
- Strong focus on identifying and mapping the user journeys across departments by domain experts.
- The benefits of taking a simple approach to mapping the event to next best action rather than relying on systems based on more generic derivation of preferences.
- Identification of how platforms such as GovCMS can help in user journeys from anonymous to authenticated systems.
The future success of myGov will rely on many aspects coming into alignment. Technology plays a role, however, the coordination of government agencies as well as the uplift of their websites will help ensure that users are presented with the best experience as they navigate their interactions with the government.
