Wrap-up of LabPlus: GaaP tranche 1 (April - June 2017)

Watch this video presentation by Pia Waugh (25 mins), or read the transcript below. Please go to the slide show or the mind map of this report and presentation to click through to all the outputs of the first tranche of LabPlus.

Video transcript

Introduction

The Lab+ Experiments: what the future of government services delivery could look like

What we’ve done in this experiment is we wanted to explore what the future of government service delivery might look like, feel like, be like, when we start looking at a couple of specific assumptions.

There has been a lot of people who have contributed in some way, shape or form to this and it wouldn’t be as amazing as it was if it wasn’t for all those people.

The other element to this, which has been a little but new for some people to get used to, is rabid, open transparency about everything we do. Making sure that we publish everything is not a matter of fun. It is not a matter of being nice. It is about taking a scientific approach to the design and delivery of government. If we push out the information, if we push out our outcomes, our discoveries, our hypotheses and our output, it gives us a chance to test and validate with people, for people to say “actually there is a better way of doing this”, and the amount of goodwill is created and people have been willing to come and share their perspective to actually improve the outcomes, which is amazing.

So a lot of that draws from a few of our experiences in the open source community, that methodology of open source, science and peer review is a critical part of this and it will continue to be so moving forward.

Systemic challenges in driving integrated services

The first one is that there are few tangible of published examples of what good looks like, particularly relating to integrated services. Everyone has a difference perspective. Everyone has a different view, different lens that they bring to it and when you actually get in the room and try to design or implement something, quite often that disconnect or different approach gets in the way.

One of the things we wanted to achieve from this work was to explore what good could look like. Not what it does look like, but some ideas of what it could look like into the future.

Secondly, iteration has become an appealing way of improving the status quo, but it runs the risk of reinforcing the status quo. What we wanted to do was ignore the status quo. Ignore the technical environment, legislative environment, ignore all of it and actually take a user-centred design approach to the nth degree where we could look purely through the lens of the user, and designed a future state that could be private sector, or public sector or community sector, or even more interestingly, it could be all of them. The idea that we should be able to design our way towards the future rather than just limit ourselves to the constraints of the present.

The third major systemic constraint, as we’ve seen internationally and to some degree locally, when agile and user-centred design methods have been adopted, they are often constrained. They are constrained to the agency view, to the mandate, the legacy technology limitations and other things.

Finally, this concept of government as assumed to be the role of sole provider. This is a big systemic challenge if we’re looking to the future because government will only ever be able to do so much for so many people. It can’t be everything to everyone. And yet there is an increasing complexity and diversity in the needs of the communities that we provide for, that we support. So the idea of opening up government components so that government provides a service, but opens it up for the private sector and community sector and even citizens themselves to be able to get that serendipitous innovation only enabled by open access.

The hypothesis

• Does government as a platform provide a practical path to the future state of government services? A secondary part of that is does it help us tackle those systemic challenges that we identified.

And the secondary hypothesis was:

• Can the neutral permissive environment of a service innovation lab support service delivery teams to build people-centred integrated services, to build better things?

It wasn’t just about providing a lab and people coming and using it and watching. That was one part of the experiment. Our team (Lab+) was about being service deliverers in designing and trying to implement and seeing how that would work and what is actually needed to support that across the system.

What we did

We then conducted independent user research specifically so that we could test if we took that no-agency lens or even a no-government lens, did it result in any change? And then could check that back and validate it.

We discovered and designed some informed future state concepts, which was purely driven around the user needs.

And finally all these concepts were tested with users and we came out with insights and validations. Analysis was done across the team, we had major support from Creative HQ and Assurity on that and is in the process of being published. Obviously none of that is personal information as you would expect but all of the insights we will be publishing that we have created, everything that we're able to publish. Quote:

At some point in your life, if you’re lucky, you get to design the way in which things evolve.

I don’t know why Daniel Day Lewis said that quote, but I thought it was a lovely idea of the mentality we brought to this.

Future states

blog post, "Lab+: Potential future states for government service delivery

[Pia switches to a report showing the results of the user research. Then to the Lab+ future states: conversational video on YouTube. The video is playing in the background with the sound muted during this slide.]

In this example a person has turned 65 since last logging on to their online banking. They get a new notification about not being eligible for NZ Super and the user has a chance to explore the matter and get it resolved through a transparent and multi agency conversation. The conversation could also be sparked through the person initiating a transaction.

[A mock-up web page appears on the presentation screen behind Pia.] So when the person logs on to their bank and gets a “We noticed you've turned 65, happy birthday. By the way you aren’t eligible for NZ Super” and the person says “hold on, why is that?”. And it says “you aren’t eligible because you haven’t been in New Zealand for 5 years of more since turning 50”. The person knows that’s not true, I want to resolve this matter. So they click into a conversation and are asked by an agent how they can help. [Pia clicks a link on the mock-up web page which starts an online chat session]. The person says they’ve been told they aren’t eligible for this reason but actually I have been here. And the agent says “how can you help us prove that”. In this particular case they say “I’ll upload the record from my passport” but keep in mind you wouldn’t necessarily need any uploading or those traditional ideas of copying and pasting stuff around government [Pia closes the first chat and starts a new one] but the concept of the agent being able to say “do you mind if I validate that you’ve been in the country for more than 5 years since turning 50 with Immigration?”. Immigration joins the conversation, with permission from the user, and Immigration gets to say “yes, we can validate that’s fine”. The service then effectively gets updated, the status, your eligibility criteria and now it’s resolved.

So you can see there is a lot going on there, but there are a bunch of concepts and then we did a bunch of A/B testing of these future states to test with our users.

[Pia switches back to the presentation.]

We experimented with a few different concepts. [Pia switches to the Lab+ future states: Proactive delivery video on YouTube.]

Becoming a senior ended up being the lens we did all of this work through because it was such a good cross-agency, cross-life event, cross-private/public sector example. One of the proactive examples was the idea that you get a notification that you are now entitled to free travel, and you go onto a bus and it just works from your phone or credit card or whatever. This was an automatically pushing service. This example was more permission-based, to enable the services.

[A mock-up web page appears on the presentation screen behind Pia.] In this case, you get a notification from Auckland City Council saying “hi, we’ve noticed now that you're eligible for the rates rebate. Are you happy to authorise us to verify that your income meets the eligibility test with IRD”. So there is a permission there. There is no copying and pasting of content, there’s no “here’s how much I earn” or having to provide that. [Pia clicks an option to grant permission, and updated rates information appears on the screen.] Rather, there is a validation going back to a trusted entity who is the authoritative source of that information for government. The person says yes, the application gets applied, you get a win for the Council, a win for the person and a win for Central Government. Again, there are some interesting concepts in that.

[Pia moves back to the presentation.]

Some of you by now will quite clearly be thinking “well that’s all fine, but this is all make believe”. The whole point here is that we get to reverse engineer these and say what would actually enable these kinds of future states. And then, based on user testing, on engaging with industry, and based on understanding the composite view of service delivery across some of these areas, we can identify key areas to invest in experimentation and see how it works.

We had a lot of our 55-70 year olds hadn’t even started on that journey. They just wanted to understand what the choices were, what the ramifications were, how they could make choices over the years leading up to their retirement.

So the idea of discovery of information and services through self-selected context, current or future, to plan one’s life and engage with government once comfortable or prepared. This is taking a more dynamic approach that what we’re traditionally taken in government but of course there are excellent examples of this already. SmartStart and NZReady are excellent examples of this kind of concept. We took it a little bit further. We took all of the business rules and worked closely with Becky from MSD (a huge thank you to Becky), and what we ended up doing was being able to come up with a demonstrator of the idea of making the business rules code.

[Pia switches to the Lab+ future states: Help me plan video on YouTube and shows some code.] What if we could identify what all those conditions were across different rules, different agencies and different services, and then you can start to do some very clever things.

[Pia switches to a prototype tool. Visible are the words "What can I help you with?" and "Tell us about your circumstances".] The clever demonstrator of this, and please don’t get caught up with the colours or those sorts of things. The functionality of this is fascinating.

If I’m a 55 year old moving to New Zealand because my daughter is about to have a child and I need to become the primary carer. I’m interested in immigration, retirement and health. Am I a citizen or resident of New Zealand, I’m going to say yes for the sake of this. I’m going to say yes and no to a bunch of these questions to just show you. You can see the idea of being able to validate certain criteria with source agencies.

[Shown on the screen behind her, Pia is providing mock answers to a series of questions. As she does so, a list of services that she may be eligible for is changing on the screen.] A lot of the criteria down the bottom (of the demo) are showing requirements are or aren’t met showing the person “I’m not eligible for this one, I am eligible for this one”. What you end up seeing is the list of services available to you down the bottom is responding to the questions, and you’ve obviously need to be smart about the questions to make it as streamlined as possible, and this is just a demo. But let’s show you where it went.

[Demo screen has 2 rows of boxes with green banners - meant to indicate eligibility, and 1.5 rows of boxes with red banners - indicating she is not eligible for these.] Our users didn’t really mind if they were or weren’t eligible, but they hated the fact they had no transparency. “What am I not eligible for and why? If I am eligible for something, why?”. They really wanted that information to understand what was going on. So be able to say “look the parent or grandparent visitor visa, 7 of 8 requirements met. Which one isn’t met?

[Pia clicks on a service she is eligible for, but for which all criteria haven't yet been met. It shows a list of the criteria and highlights the unmet ones.] Applicant sponsor is a relative. Yes, actually I’m not a relative” but if the condition was “applicant intends to meet visa conditions, then they could say “oh I didn’t know that was a condition, I’m going to choose now to change my planning and then be eligible for it.”

So this is kind of powerful in that a person could look at their answers and review the yes and no, to better understand their eligibility. [Pia brings up screen showing how she answered the questions.] At some point they would want to apply, but they are told they aren’t eligible for something which I should be, so now they can go and resolve the matter, perhaps through the conversational service.

So you can see how these 3 modes of service delivery might blend into each other. In a couple of weeks time we will have a beautiful future state example for you based on an actual service that does blend the 3 together in a meaningful way.

[Pia switches back to the presentation.] So these 3 concepts are quite big, quite chunky, quite interesting, and the idea of making those business rules available has such wide ranging possible applications not just for service delivery but for a bunch of other needs such as regulation or compliance.

Reverse engineering the future states

Lab+: Reverse engineering the future states

[Pia switches back to presentation.]

Reusable components

[Pia opens a page on data.govt.nz that provides the services register in machine-readable form.] So one of the things we did was, again we worked with MSD, to take their Family Services Directory and create a machine-readable version of the exact thing they already have on data.govt.nz and the machine-readable one just means you can now talk to it via an API (Application programming interface). It’s only a JSON API it isn’t a spatial one in this case (yet) but it’s an interesting start because then we got to do some interesting things with it. We took all the DIA registers we had access to and built a first version prototype, completely experimental — none of this is real — to see what all the services are. When you have machine readable access to information like this that is reconsumable, you can start to do some clever things.

[Pia displays an example pie chart representation of the services register data.] So this is literally about half an hour of play time just to see what you could do with it from a data visualisation perspective, but we’ve already got 3 life events wanting to use it and 2 external organisations wanting to use it and we haven’t really even been talking about it yet.

Being able to say what are the services by agency, and which agency is providing all the services, or I want to look at payments by agency. [View has changed to an interactive data visualisation showing payments grouped by agencies] and as you would expect there are a lot with Ministry for Social Development, but being able to see some of this information about where the payments are is a good way of understanding what’s going on. And then finally a simple one, given we were looking at seniors, just understanding what services are relevant to seniors helps us to present that information appropriately.

So the services register has become a fairly obvious thing that needs to be built but there is no natural home for right now, so that’s one of the things we’ll be experimenting with next.

Finding diamonds

We’ve also written a quite amazing case study on MyMSD, not just MyMSD but also the digital transformation journey of MSD. Again there is some great lessons learned in there, some examples of good, but also good examples of where it needs to go next. One of the dangers is that if you see something better than what you have, and assume it to be sufficient, then you miss the opportunity to leapfrog, so there are some good opportunities there.

And I love this quote for this kind of things:

The ecosystem

We went in genuinely with the idea that maybe this idea is not of interest, maybe it is, but let’s actually see. The result was quite promising. They were very keen and when we started actually building things, we started seeing all kinds of clever things could be built, including online contributions.

[Pia switches to a blog post and scrolls to pictures of Lego models.] The interesting thing about these is that all of them actually have both government and private sector components, so they started to identify those and in assessing that work we can identify common components that if we built should actually get reused as a test of whether it would actually get reused.

It is easy to build Lego but let’s see what happens in the real world.

[Pia switches back to the presentation.]

Supporting agencies

Creative HQTECIRD

[Pia switches to the blog post on the Web Toolkit.] Kudos to them for writing a fantastic blog post about the experience, about what worked and didn’t work, and what the outcomes were. And they ended up with a very shiny clickable prototype [Pia shows the prototype page] of a student trying to make better choices. There is a video now up about that with a transcript available. [For more details, see the blog post Lab+: Working together to improve outcomes for students".]

We also had a SmartStart team exploring potential improvements and a future state being finalised at the moment.

Value proposition of Gov as a Platform

Treasury

Service Innovation Toolkit 2.0

Next steps

There is a lot of reading material there and a lot of interesting things to delve into. The next steps is worth covering. The hypotheses have proven reasonably successful as a first key thing. Next steps would involve further validation and research but taking a pragmatic next step. What can we design, build, test, both in terms of what would help service delivery teams in agencies, and what helps private sector and community sector, and most importantly what helps people that need to be able to interact with government either directly or indirectly.

This first 10 weeks has been about a small team with diverse backgrounds testing some concepts. Now that we’ve done that, now is the time to engage with the broader public service. There wouldn’t have been much of a point in doing that unless the outcomes of this had been a little bit promising, but now is the time to figure out how to do that better and we want to work very closely with all of you to do that to create more of a virtual Lab that people could drop in an out of as they like so that we’re not limited to the physicality of one or two locations. And we look forward to working with all of your to take this idea forward.

Thanks

For a non-traditional outcome you need a different approach and we’ve had some good discussions with some of the contractors involved, some of the public servants that have been involved about the experience they’ve had, and just that idea of partnership. If we’re both trying to do something meaningful, how can we do that genuinely collaboratively. That’s been a part of this experiment and a very important part of it’s success. So the next part of that is that there has been a lot of people involved.

Agencies and companies in no particular order

• Department of Internal Affairs (Service Innovation, GIS, GCIO, SmartStart)
• Creative HQ
• Ministry of Social Development
• Assurity
• Tertiary Education Commission
• 3Months
• Inland Revenue
• NZYM
• Commission for Financial Capability
• Felicia Yii
• PHARMAC
• Rabid
• Also a thank you to CatalystIT for sharing their technical lessons learned.

People

Boss: Darryl Carpenter

Service Integration Lead: Pia Waugh

Core team: Michelle Edgerley, Lee Dowsett, Tegan Bertram, Merridy Marshall, Michelle Shannon-Smith, Jean Johnston, Susan Carchedi, Kate Forward, Bill Young, Taher Amer, Jennifer Geard, James Fuller, Tinus Wagner, Matt Lough, Felicia Yii, Brock Jera, Yasmine El Orfi

Agency support and sprint teams: James Collier, Christine Bennett, Murray Johnson, Glen Thurston, Kate Nixon, Becky Cassam, Melanie Turner, Clare Toufexis, Leigh Vollans

Additional support: Rachel Prosser, Chris Pollard, Ben Hayman, Stefan Korn, Mark Pascall, Jennifer Nichol, Brett Calton, Phoebe Kwan, Anne Davidson, Yvonne Tse, Kelly-Ann Hubbard, Derek Moore, Nadine Henderson, Josh Forde.