Chris Buxton says the technology team at Stats NZ have completed numerous innovation activities including sponsoring its own public hack events.
These activities ranged from large scale vendor engagements to smaller scale explorations.
One of these was exploring the way they engaged through government procurement to evaluate commercial options, called project Wero.
This was a wide-ranging activity that engaged vendors to understand our organisation’s problems, strategy and customer needs and for them to identify opportunities to innovate to help us achieve success.
Initial submissions from 42 vendors was reduced to six to carry out proof of concepts. These included a big data solution, Customer analytic capabilities, process robots and innovative data collection partnership.
The proof of concepts also allowed us to explore how we approach this sort of innovative market approach.
He says big data solution was one of the most successful delivering change benefits, outside the technology.
“It has enabled us to promote different ways of thinking about problems which improves solution designs, with software explored being used to modify our data ingestion capabilities,” says Buxton.
“This is enabling significant reduction in the cost and development effort required to respond to external data changes. It now means that we can start to explore data capture in different ways to improve the quality and variety of data we provide to customers.”
He says another innovation exercise explored the use of AI and real-time language capabilities to improve customer engagement. He says this is now in the form of a chatbot that is now live on the Stats NZ website.
He says the project developed as an idea coming from a five-day hackathon that Stats NZ teams participated in.
The project brings together Microsoft Azure, cloud based capabilities in delivering the chatbot that is then combined with an enterprise search capability to greatly improve data discovery.
This also enables the user to search using real language rather than the legacy, statistical terms,he says. For example, instead of searching for “employment” a user can search for “jobs”.
The solution was initially developed as a MVP within five days, and was then given approval for further work to refine as a beta product. “This is a great example of us exploring a new technology and then being able to start to gain value from the solution,” says Buxton.
We have carried that on and are now working on an internal capability to improve the information and support we provide to our organisation.
He says an example of their work on machine learning is going to be used for the first time in the 2018 Census.
This is a workload allocation capability that combines various data sources to make the most efficient allocation of field staff for data collection.
It is integrated with Salesforce, geospatial systems and other staff management data systems, to make the best allocation based on the operational need, staff location and the geography to make decisions, says Buxton.
The solution was developed in partnership with the University of Auckland.
This new approach has been enabled through the digital first approach taken for the 2018 Census, with an integrated processing and analytic system. This is enabling near real-time insights into response rates, never before been seen, so we can target our activities to gain the best outcome for those hard to reach areas.
“We continue to explore other technology innovations to help the organisation to understand its new cloud based delivery model and the opportunities that it is providing,” he says.
He points out a major cultural challenge through much of the innovation was overcoming the organisation’s desire to over-engineer solutions as well as specify what it required up front. The organisation’s approach to major activities like this in the past has been to run large up front procurement processes, that tend to be slow and deliver inconsistent results.
“To overcome these cultural issues, we maintained tight control on the initiatives, actively monitoring and managing scope to keep the innovation moving,” says Buxton.
“I recognised that once innovation activities gain broader business engagement they can tend to turn into large, complex investments quickly, trying to satisfy everyone and lose sight of the objectives. These then take longer and longer to deliver. By keeping tight control, I was able to keep the initiatives focused and on target.
“The other area was to focus the innovations on an outcome or specific problem statement. That meant we could also empower the teams to explore options while maintaining the clear objective. This again ensured that the reason for the work was consistently communicated irrespective of the technology or solution being explored.”
Focus on business continuity
Some of the lessons from these innovation programmes were incorporated into the delivery of the new Stats NZ website.
“The feedback from customers and our own staff was that our website was difficult to navigate and find content, manually heavy to maintain and less than robust,” says Buxton.
The Kaikoura earthquake resulted in the loss of the service for two days, while the many components were recovered, it was a less than ideal situation, he adds.
We transformed the Stats NZ website from an internally developed, hosted and delivered service to leveraging a modern as a service web hosting platform, says Buxton.
User experience (UX) was a key focus of the redesign, as well as the establishment of an agile baseline so Stats NZ can respond to customer expectations.
The website has gone live and is proving to very successful with the organisation’s customers, he says.
“For the best outcome, we wanted to ensure that our Channel development team was built into the projects in a co-located multi-disciplinary model. For this piece, we established business leadership but then co-located the teams to ensure greater engagement.”
“We were also able to take some of the capabilities developed through our innovation/hack events and incorporate them into the solution,” he adds.
The best example is a natural language chatbot with integrated enterprise search. While still in beta, it has the potential to transform the way that users discover data. This ability to have co-located teams, but also leverage the larger Digital Business Services team proved extremely valuable.
“Co-locating the teams meant that we had to take some elements of the technical team outside of my area and embed them within the web publishing team. This was a new approach for the organisation, leveraging a previous move to a DevOps delivery model to move to the next level.
He says the structural challenges of the past have started to fall away as the DevOps model matures in the Stats NZ environment.
Buxton is involved in multiple leadership and governance groups and uses several avenues to collaborate and influence both within Stats NZ and the wider government sector.
He says Stats NZ has a core leadership group (CLT), made up of senior leaders, that meets weekly to discuss operational matters and strategic initiatives, “bringing together all elements of the organisation”.
The CLT meets with the executive leadership team (ELT) every month to discuss relevant topics and review organisational performance. I also sit on the Enterprise Prioritisation team, evaluating initiatives against our strategy to give the organisation guidance on relative priorities.
Buxton says to develop their teams, Stats NZ is investing in a guild-based model.
In any outcome focused group the maintenance and sharing of skills and IP is critical, he says. “We have a series of technical ‘Guilds’ or communities of practice. These bring staff together so they can learn and collaborate.”
“Through these guilds we also provide mentoring and leadership to our staff,” he says.
Across government, Chris plays a key role in a number of all of government groups, for example he is a member of the GCDO Technology Partnership group and the accelerating cloud program steering group, focused on the acceleration of cloud uptake across government.
He is bringing a community based model into the broader government sector. Buxton established and is the chair of a government community group, the “Cloud” center of excellence. This group is bringing together practitioners to share their experiences of cloud adoption and helping their colleagues avoid the pitfalls.
“This has involvement from all agencies and brings together those implementing cloud systems, so they can share IP. This exposes my teams to other agencies and their experiences. This is adding real value through creating new collaborative opportunities.”