Transcript - 'Podcast with Paul Shetler: a digital architecture for government,' The Mandarin (July 2016)

Tom Burton: Hi, I’m Tom Burton, publisher of The Mandarin, and today we’re speaking to Paul Shetler. Welcome, Paul. 
Paul Shetler: Thank you very much. 
Tom Burton: Thank you. Paul, you’ve had this unique position to watch government and it going through its digital transformation processes. We know all governments around the world, they’ve given a lot of priority to digital. Today, I just wanted to talk through a little bit your observations, having watched it from the cockpit of the positions you’ve been in. How do governments start to think about really accelerating their capability, given the Internet’s here, it’s arrived? 
Paul Shetler: This was one of the major things we were focused on in the UK, and it’s one of the things we thought an awful lot about here in Australia as well. It’s very difficult to actually transform your organisation if you don’t have a clear idea of what it is you’re trying to do, or how to do it - if you don’t have the skills, right? So the first thing really was trying to understand what capability people actually were actually needing in their different agencies. If you were looking at a delivery agency, for instance, or a very delivery-focused agency, like say the Ministry of Justice, or the Department of Work and Pensions, or the Home Office, or something like that, you would very typically see that they might have had an awful lot of traditional IT skills, but they didn’t necessarily have the skills in human centred design, or product management, or delivery management, or agile development, or things like that, to be able to take these things forward properly. So building up those skills is massively important, right? Also, in many cases, I know this is probably not too much of a surprise, but a lot of government agencies, pretty much around the world - there’s nothing specific to any place here - typically tend to have relatively weak commercial skills as well, and to some extent that’s a result of many years of outsourcing. So knowing how to build things, understanding what should be built, and understanding how you manage that process - whether you do it yourself or through a vendor - is massively important, and, if you can’t do that, you can’t move forward. Those are also the kind of skills you have to have pretty much from the bottom all the way through the very top of the organisation. It’s really kind of no good to have a digital team, which can even be a pretty great digital team, if it’s not connected to the top leadership of the organisation. If what it’s doing is not aligned with what that leadership wants, it won’t work. And if the leadership itself doesn’t understand what it wants done, there’s no way it can transform itself either. So you really need to go from top to bottom. 
Tom Burton: Right. And in terms of, if you like, building that capability, what are the types of propositions from what you did in your own experience, if an agency really wants to try to accelerate that capability, what are the sorts of thing should they be doing? 
Paul Shetler: There are a few different things. So, from the standpoint of practitioners, the most important thing is to learn by doing, right? It’s one thing to produce lots of documents, or manifestos, or Tweet all over the place, or whatever, saying how you should do stuff. But it’s very different, you know, - it’s hard for people to pick up on that and say, ‘Well what does this really mean?’ For instance, when I first came into the British government, we were talking about Discovery. Okay, that’s wonderful, we’d say, ‘We’re going to do some Discovery’. What does that really, really mean? That’s why when we came here in Australia, we tried to be very very clear about that. We were very very, prescriptive about that. But learning by doing, actually working in teams, with other practitioners who had done this before, who were delivering things according to a particular set of standards, and sharing that learning with other teams, in an environment not unlike an Accelerator or incubator, is a great way of skilling up large numbers of people, and, when they are skilled up, having them move back into their agencies as functional teams, which was really important. 
Tom Burton: This was your technique with the prototypes, wasn’t it, with the DTA? 
Paul Shetler: Yeah, this is what we did with the exemplars. 
Tom Burton: Exemplars, yeah. 
Paul Shetler: But it’s also what we did with the exemplars in the UK. In the UK, there were 25, in Australia there were 6. And the reason for doing that was twofold. One was to show that actually, yep, you can - government can deliver a great digital service, and give examples of what those looked like. And the other one was to build the capability to do so across the other services as well. 
Tom Burton: Right. A lot of these skills are jobs that traditionally haven’t been understood in the public sector. How do you build that sort of understanding of even what those skillsets are, and, if you like, the taxonomy sitting behind, the classifications, if you like, of these jobs. 
Paul Shetler: Yeah, so we worked really hard on that. We worked really hard on that, both in MoJ, and GDS, and then later on in DTO. And in a number of - in some cases, the same people did the same work, where we defined out what you would have in the typical digital delivery team. One of the things which I was always pushing very very hard from the very beginning is delivery teams need to be small, and focused - focused on what they do, so they don’t get sidetracked - it’s very very easy to get sidetracked when you’re working in an organisations that works in a slightly different way, which is in the process of transforming, and small enough that they can actually be really lean and effective. So there’s nowhere to run, and there’s nowhere to hide, and you can deliver things quickly. Now, we defined a series of roles like Product Manager, Web-Ops Engineer, Delivery Manager, User Researcher, Developer, Interaction Designer and so on and so forth and we defined them with the role specification, and we then said: this is what each team needs to have, almost like a cookie cutter. 
Tom Burton: Right. 
Paul Shetler: Now, the thing is hiring the right people to bring them in. Now, on the one hand, you can train people up, and we can can talk about how that can be done, but also, in many many cases, these are skills that just don’t exist, like product management is not something people normally think of when they think of government services, right? And yet, when the public consume a service, that’s very much what they think about. They think about it in the same sense in which a financial service they would be consumed, which is also a product by the way. So getting those kind of skills in is really really difficult, and I think you need to have a clear focus on hiring externally to bring some of those in, to help skill up your internal people. 
Tom Burton: Right. And presumably what we’re seeing here is a shift, if you like, away from the traditional ICT into the digital world. 
Paul Shetler: Yeah, it’s a shift and it’s an addition. So you’re always going to need to have traditional ICT skills, absolutely, because the problems that government is dealing with are not only end-user-facing problems. So there’s many things that have to happen in the back office, and all the way throughout, where you still need to have traditional ICT skills. But when you’re talking about things which are dealing with end users, which are typically, these are relatively novel, unknown kind of situations, you might think you know the solution at the very beginning, but almost always you find out you don’t once you start releasing the code - then you need to have people who are used to dealing with that kind of ambiguity, who are Product Managers, User Researchers, who can actually pull out what the needs are, and so on and so forth. 
Tom Burton: Right. And in Australia, you started an Academy, is that a proposition you would recommend? 
Paul Shetler: So in the UK, there is an Academy. The Academy was initially in the Department of Work and Pensions, it was pretty successful. I have to say that when I was first at the Ministry of Justice, I was a little bit skeptical about it. But I was made into a believer when I went there and I saw it. I went up to Leeds and I saw the teams that were working there, and these were all long-time public servants, right? So these were public servants, these were not, you know, latte-sipping hipsters or anything like that who’d been hired in by the Guardian, these were people who were long-time DWP employees who had a real sense of mission of what they were doing, and who also, very importantly, understood the problems, because they were right in at the coalface of delivery, they understood what their users were going through, and they also realised that there was a path forward as the organisation transforms itself and has the need to take that knowledge and put it into sort of new kinds of roles. So they were training up their people, they were building up their existing capability, in these Academies, and what they were basically doing was they were defining out according to the same taxonomy I just mentioned earlier - ‘OK, so, you want to be a product manager, this is what you have to learn, here’s some courses on how to do that,’ and then they did the same thing by learning by doing. They put them on projects very very quickly, and they mentored them through. 
Tom Burton: Right. Just thinking about the, as you said, the leadership issue, top to bottom, you know, how many agencies, if you like, excite themselves about this technology piece, how do they start - because if the Internet’s here to stay, how do they get that sort of commitment and institutional momentum and focus around technology? 
Paul Shetler: I think it’s really - it’s really - you have to sort of set - it’s like a marketplace, it’s almost like a Darwinian kind of marketplace. You have to set the rules that it operates within, you have to say what is good and what is bad. Public servants, just like anybody in any large organisation, you know, they operate according to the set of incentives that they’re given. You give them one set of behaviours, you give them another set of incentives, you get another set of behaviours. If you start telling them that the way you’re going to get ahead, if you’re an ambitious, smart public servant, that the way you move ahead in the organisation is by finding ways of making complex things simple, is by knowing - being familiar with - digital, by understanding the possibilities of technology, by promoting that change within your organisation, by being an advocate for this within your organisation, then you’re going to get one set of behaviours that are highly desirable. Now, in the UK, the way they did that, which I thought was really quite brilliant, was they formed a group called Digital Leaders. Now, Digital Leaders were essentially the various sort of smart, ambitious, Directors-General who were probably going to wind up being the next Permanent Secretaries of our departments in the UK. And Francis Maude set up this group. Francis Maude also made it very, very clear that digital was the future. It wasn’t about disruption. It was about fixing things, it was about making the Civil Service capable of operating with its head held high in the 21st century, with the same kind of efficiency, the same kind of effectiveness as we see in any green-field organisation. A quite ambitious thing, right? So when it became clear that, if you want to move forward in the Civil Service, you have to adopt this, and you have to push this, well, that’s of course what they did. 
Tom Burton: Right, right. And you mentioned avoiding the word disruption, what do you mean by that? 
Paul Shetler: Disruption is a word that’s often times, that’s sometimes been applied to the work DTO’s done, which I’ve always found sort of funny because I’ve never used it to describe anything we’ve ever done. The reason I don’t particularly like the word disruption is that when you’re dealing with an organisation like government, which is in many cases dealing with very critical things, right, if government is disrupted, you can have riots, you can have, you know, no food in your supermarkets, you can have financial crisis. These are typically not good things. And nobody actually seeks out disruption, even in the private sector. It’s not like people say, ‘Oh yeah, you know what, I want to disrupt my organisation.’ No! The reason you go with digital is so you don’t get disrupted by others in the marketplace. You want to be the disruptor of them, not them of you. 
So what we’ve always felt, our view, what we’ve always felt we were about, was fixing things, making things better, making them more effective, making them more efficient, making them - making it so they actually met the needs of people who were trying to use them, right? That’s how government avoids being disrupted, by actually improving itself. And that’s what we thought our role was. So it’s actually not really about - it’s not about disrupting the public service, it’s about making the public service a 21st-century public service. 
Tom Burton: And fixing things. 
Paul Shetler: Absolutely. 
Tom Burton: Delivery being the proposition, if you like. 
Paul Shetler: Absolutely. And I don’t think you can separate delivery from policy, right, I think that’s a very false distinction. 
Tom Burton: Yep. Tell me about, if you like, the sort of an architecture that governments can think about. We’ve seen a lot of digital transformation in the private world. Governments are now moving into that world. How would you recommend governments think about this? I know that’s a big question, we might need to break it up into pieces. But what’s your - if you’re trying to describe the way governments, state governments, federal, and local, for that matter, around the world, are thinking about this, what’s your thinking? 
Paul Shetler: I think the first thing we have to do is realise that we’re not a greenfield, right? So were not the sort of sprouting up from nowhere and able to sort of create anything ex novo. If we were, that would make it so much easier, right? Our problem is more complex because we’re a big brownfield. In fact, in many ways, in some places, probably the brownest of the brownfields, right? We are the institutional origins out of which other things came from. Look at the UK, for instance, and then look at anything that came after that. It was the government, in many cases, that allowed the things that emerged in the first place even existing as brownfield industries. So what we can do, and, I think, what we have to do, is look at, ‘What are other brownfields doing? How are the ones that are succeeding in the face of internet doing? What are the approaches that they are taking?’ And apply this basically commercial best practice to what we’re doing. We don’t need to try to go way ahead of them. And when we talk about digital transformation of our services, we can talk about dancing holograms and all kinds of stuff like that. Actually, if we just meet the standard of, say, a really brilliant bank, we will already have moved things - we will already have already pushed the dial way forward. So let’s look at how companies like that are actually starting to shift their businesses, and apply some of those lessons. 
Tom Burton: Right. So trying to understand that - how do you see that? Presumably there’s sort of like almost a user layer, the layer that users start on. 
Paul Shetler: Yeah. So if you look at, say, banking, or brokerage, or any of those kind of industries, what you’ll frequently hear is that people talk of ‘on the channel’, right? Which is that you should be able to dip in and dip out of even a relatively long-running business transaction at any point at any particular channel and not have to enter from scratch, and basically know where you stand. You’ve progressed things. So for example you might want to start something out on the web browser and then find out later on by a notification that’s come through at an ATM where you stand, click on the notification and it takes you to maybe your app or whatever, doesn’t really matter, but every step along the way, that transactional information is consistent and consistent across all the different channels, and it’s also orientated towards what you are wanting to accomplish. 
Tom Burton: Right. As opposed to the organisation - ? 
Paul Shetler: As opposed to the organisational specifics, right? When you think about - think about if your bank or somebody like that tried to expose all its innards to you and said, ‘Yeah, you know what, we’ve got all these different lines of business, we’ve got all these different little groups, each one has its own website, this team here of six people, they got their website, you can deal with them,’ you would never know how to navigate through any of that. What they do try to do is they try to say, ‘What is your goal? You’re trying to buy a house? Trying to send your kids to school? You’re trying to do X, or Y, or Z? Here’s a whole set of propositions we can take you through to get you to that point.’ Now, when you think about government, we have very similar kinds of problems. If you want to start a business, you’re probably going to have to deal with two, three or four different agencies at the federal level. You’re probably almost immediately going to have to start dipping into the state level, when you’re dealing with another set of agencies. And then probably at some point, assuming your business is in a particular place, you’re going to have to get licences, and so on and so forth. Typically, each one of those transactions, or interactions, is treated as if that’s the only thing in the world that you care about. The only thing you care about is getting that office space! That licence or that office space, from the perspective of the public service operative. Actually, what you care about is starting your business, and knowing that you’re being compliant across the entire process - that you’re not missing important steps, and that everything you’re doing is moving you towards that goal, and, by the way, where do you stand? So thething is we should be modelling our services around what the users are trying to do, around that entire user journey, making that accessible via any different kind of touch point that you want to come into, it could be things like a conversational agent like Alex, which the ATO uses, or Watson, or Siri, or a web browser, or your phone, or an app - even if you need to, in certain cases, going into a call centre or even if you really have to going into a shop-front - but at every step along the line, you should have that same idea of the same thing you’re trying to do with the same information, like what you did last. 
Tom Burton: Right. So, that’s the front end. Now, understanding the back-end, if you like. I think you’ve said before, I’ve heard you say this before, governments need to move away from this concept of case managing everything. What do you mean? 
Paul Shetler: Yeah. Well, okay, for that, for what I just talked about to happen, you cannot have a back end which is highly, highly manual, and which is made up of an awful lot of very disparate, broken - not broken, that’s the wrong word - very disparate IT systems, which are almost chopped and changed and moved around from department to department as administrative orders change. You just can’t have that sort of incredible web of complexity. What you need to have is you need to have a really, really highly automated back office that can respond to requests coming into channels very very quickly, right? So think about what happened to financial services - when you had to move from t+3 to t+1 to t+0. What did that do to the back office? It did a lot to the back office. It meant that they had to automate the hell out of it. That’s what they had to do. They had to automate the hell out of it, because otherwise it wasn’t going to work. The same thing here. You can’t talk about a user journey and joining all these kinds of things up if you actually don’t even know how to map that in your back office, right? If you haven’t automated it. So a large driver of complexity in the public service is the idea that pretty much everything we do is case management, right? Sort of follow everything through, we record everything, we treat everything as if it was more or less an edge case or an exception. It’s impossible to automate that way. What we need to start doing is to pretty much remove that paradigm from our thinking and realise that actually much of what we’re doing, in terms of legislation and the policies that we’re implementing - because that’s what they actually are - can be codified as code, right? And basically think of your back office now as being a place where you implement that, run that through a rules engine, the contacts that I have with you I can use a simple CRM to manage that, and then the whole idea of case management sort of goes out the window. What I’ve replaced it with is straight-through processing, which is a concept familiar to most people who work in financial services, and for those actual particular cases where I cannot handle them through those roles, because I’ve found something that does not work that way properly, then I actually have people, right? I have people dealing with the exceptions, trying to fix things there. But that shouldn’t be more than 1 or 2% of my cases ever. 
Tom Burton: Right. So that’s a big task, isn’t it, because in this - particularly in the Australian government context, state and federal, I’m going to say very few agencies, let alone the system, have that kind of straight-through type thinking. 
Paul Shetler: I think it depends on where you look and where you talk to. I think some states and agencies have been looking at that. I think at the federal level some agencies are looking at that in particular ways, and are actually doing a fairly good job in a particular couple of cases. And I think also depending on the state, some of the states are actually looking at that pretty intently, and trying to implement some of the things, so I think that’s a massively important thing. I’ve been talking about this sort of from the standpoint of federal government. It does become more complicated as I start pulling things down. Nevertheless, I still need to be able to have some kind of that user journey, and linking that back to the various different back offices to support it. Until I can do that, it really doesn’t make sense to even raise people’s expectations too much about it too much, right? Because, if I’ve got all - if the government thinking is manual in my back offices, I’m not going to - 
Tom Burton: To achieve the first goal, which is to - 
Paul Shetler: Yeah, which is for government to start doing what people are asking it to do, right? Rather than people do what government asks it to do. 
Tom Burton: Right. So the - presumably there’s a huge drive in productivity as well, because, if you can automate a lot of this, there are big savings that have been promised, just simple cost terms, start to pop out? 
Paul Shetler: Yeah, exactly. And this is exactly what every bank and brokerage and insurance company have seen, right? Precisely that, where they used to have huge operations departments of people who were basically just inputting from one green screen into another green screen, printing out documents, annotating those documents in pencils, handing them over to the person sitting next to them, reaching into another screen, and so on, and so forth, and when you then settle trades the same day, all that stops, all that stops, it all becomes very efficient and delivery-focused. 
Tom Burton: Right. How does - to an agency lead that says, ‘Yeah, I get that,’ how do they start? What’s the process for that? 
Paul Shetler: You need to start - the first thing you need to start doing is looking at - I think there’s an awful lot of complexity embedded in their back offices as a result of legislation, right? So, for instance, you look at some of the stuff that actually has to happen if you want to make a benefits payment, for instance, there’s an awful lot of manual steps that you’re forced into, so the first thing that you really have to do is understand why you’re doing this thing in the first place. You do need do a Discovery on this, you actually do need to understand why is this actually even happening, and one of the first things you’re going to want to do is to get rid of those manual requirements. 
Tom Burton: Right. So a pretty hard-headed review. We’ve had these sort of red tape reviews, but it’s not red tape in this case - 
Paul Shetler: It’s a hard-headed digital review. You’ve got to say, ‘If we want to bring government into the 21st-century, and we’re really serious about this, what do we have to do?’ We can’t continue going down the same path. We actually have to show some leadership, and do the hard work to make things simple. That’s the phrase we used to use a lot at the GDS, you know, ‘We’re there to do the hard work to make things simple,’ because it’s not simple today. It does require hard work, it does require analyses, it does require, perhaps, changing laws. You know, there’s no way that we’re going to reach that point unless we do that. And we have to do that. It’s our duty to the people of the country that we’re - 
Tom Burton: And presumably you need a pretty strong sense of the user cases as well - 
Paul Shetler: Of course - 
Tom Burton: So that’ll prioritise - and then government needs to be working together in a way it probably doesn’t at the moment. 
Paul Shetler: Indeed. So what you want to do - you basically need to do a review of exactly what is driving this back offices, why they are the way they work today, how much of that is legislative, how much of that is policy, understand that. Change the legislation, change the policy, as appropriate. I’m not saying in every single case, because there may be very good reasons for certain things, but, as appropriate, do it, and have that be your default position. And then go through all of your back office processes after that, and optimise, optimise, optimise, optimise, so they’re automated by default, right? We were earlier talking about automated as appropriate, but they should automate by default. Find the reason not to do it. That’s the more important thing, right? Even for that to work, you’re still going to need to have a set of simple functions that are going to be pulled across all of these different back offices. You need to have a set of government platforms, which, by the way, don’t need to be operated by government. 
Tom Burton: Yeah, I was going to ask about these platforms, because, you know, they’ve become a bit of a fad word. What is your advice around government thinking on platforms? 
Paul Shetler: So it’s important to keep in mind what a platform does, right? A platform is not a big application. A platform is not a big complex system. A platform basically exists to hide complexity, right? A platform basically is a simplification of all the complexity that lies underneath it that allows for the development of high-order systems on top of it. So when we’re talking about platforms in government, we’re talking about things in some cases like, obviously you want to have some way of somebody proving they are who they say they are, like an identity platform, you’re going to want to have some way - some very standard way of letting users know things via whatever device that they want to, whether that’s their phone, whether that’s their email - 
Tom Burton: Some sort of notification? 
Paul Shetler: A notification, you want to have some way of notifying users of where things stand, so they known in their user journey, ‘I’m at this point, and I need to do this next,’ so they don’t have to call your call centre or go to a shop-front and wonder, ‘Hey, how come it’s been six weeks since something has happened,’ right? So they’re kept fully in the loop. You’re going to want to have a standardised way of taking money in, because people make payments into government, as well as receiving money from government, and so on and so forth. Now those are all things which currently already exist as commercial products, right? There are companies - their entire business is based on that. So why would anyone think that they’re going to build something better than that? Why would you want to? Why would you want to? Just use the commercial offer, and provide a simple government API over that, so that you can then decide, ‘I don’t like that commercial offer anymore, there’s something better, I’ll switch out that substrate, and I’ll keep the same way of accessing the service, but I have a different provider underneath it’. 
Tom Burton: Right. So you could work with - I’m just making this up, but - you could work with PayPal for a number of years as your collecting system, and decide that you want to move on or stay, and, as you say, turn it on and off. 
Paul Shetler: Absolutely, right? And supporting all of that, you need to have some set of common reference data, right? Some people call it a register, other people call it a reference data, it depends on where you work. My background is in financial services, and we used to call it reference data. It’s essentially, what is the data that defines [indecipherable], that defines entities, and so on, and so forth. You have a simple way of doing it, so you avoid things breaking when a piece of information might be one thing here and another thing there, or one entity thinks this thing is that, and another entity thinks it’s another thing, that they have to reconcile later. So having a very simple set of information across everything that supports all standard business processes that government operates. Ultimately, all those things, that back office, your platforms, any kind of data, should be able to feed a set of analytics, so that you can then understand, ‘How well are you government services doing? We’re spending X amount of money, we’re spending X amount of billions of dollars on IT supporting the public, how well are we doing? What is our time? How long is it taking to do stuff? What are our call volumes? What is the impact of these changes we make in these products? Reporting this out in a transparent way that means the taxpayer understands what they’re getting, right? Because they are paying. 
Tom Burton: Yeah. And also understanding performance, and user pain, and all those sorts of issues. 
Paul Shetler: Absolutely. If we were making a change to a particular product in the user journey, what impact did that have? Did it mean we were having less calls coming in? Did it mean we were providing a more efficient service? Did it mean our user satisfaction numbers were going up, because we should be measuring user satisfaction directly, right? Or, did it mean the opposite? And provide that, again, in real time. That should not be information which departments or agencies or anyone else is editing. It actually should be coming directly out of the systems, so that you have it in real time, because the Internet operates in real time, and if we want to have a digital government, it needs to start operating in real time too. 
Tom Burton: Yeah, I think it’s a really important point, too, isn’t it. This sort of aspiration for - at the moment a lot of discussion in digital tends to be very structured, whereas, as you observe, it’s a real time game, and government needs to move there, yeah. And presumably a lot of this is then feeding back to users, isn’t it? 
Paul Shetler: Absolutely. So the basic idea is that the user should always know. They should never have to wonder. Are they doing the right thing? Are they being compliant? Is government doing what it wants to do? The user should always know what is the next thing they need to do, if they need to do something else, and, by the way, if government is supposed to do something, when is it going to happen? So keep the user notified at every step along the way of what needs to happen, and also, what is happening, so they don’t have to call to a call centre or go to a shop front to find out whether something’s happening or not. 
Tom Burton: Yeah, yeah. So just last area, then, you’ve worked in government, seen it work, you know. What are the models - you know, political systems, bureaucratic systems - what are the sort of models you’ve seen that work? In Australia, we have our executive essentially sits in parliamentary structures, and bureaucrats... Do you have any observations about that? 
Paul Shetler: What I found that worked really really well, and it was when I was at the Ministry of Justice, and this is true for pretty much all British agencies, was - there’s a pretty tight connection between the Minister, the senior bureaucrats, and then the rest of the organisation. There was also - there was no real split between policy and delivery either, right? So really, really important that you would have you Secretary of State - in my case you would have your Secretary of State for Justice, and he was sitting, you know, about 50 feet from our permanent secretary Ursula Brennan, you know, they could see each other all the time, I would run into the Ministers pretty much every day. We would have conversations. So we were always aligned between the politics and the organisation. They were providing us the head cover, so we could do our jobs. And our job was, of course, to implement the policies of the government of the day, and to provide them with our unbiased advice. 
Tom Burton: So a much closer alignment. 
Paul Shetler: A very, very close alignment. 
Tom Burton: Digital needs everyone to be in the same room. You know, that’s a metaphor, but it’s a proposition that you need to be working closely. 
Paul Shetler: You need to be working closely. So we need to be working closely with the nation’s political leadership, we need to be working closely with the policy-makers, as well, and the delivery people, right? You need to have that very tight feedback loop all the way through. Now, if you don’t have that tight feedback loop, if you separate out policy from delivery, you’re never really sure if the policy you’ve put in place is actually delivering the thing that you think it is, particularly if you don’t have any kind of analytics capability, or any kind of reporting on what’s happening with those services, which we don’t today. So you need to have that - you need to have a tight feedback loop between policy and delivery, and, frankly, it seems kind of crazy to me that you would have, for instance, digital teams who are developing services without policy people sitting on them, because there’s going to be sometimes policy constraints that we have to make sure we meet. In other cases, there may be cases for making changes to the policy as a result of what we’ve learned through user feedback or user research, right? And that’s the standard way we used of working in the UK. I’d like to see that here. 
Tom Burton: Yeah. I think - I’m not saying this critically - certainly in Canberra where there’s a lot more focus on policy, that sort of distance between policy and delivery is a practical issue, if you’re going to deal with this issue. 
Paul Shetler: Yeah, it very much is. And when you think about how digital companies work, they don’t have this distinction between the big brains who think of everything, and then the worker ants over there. You just don’t have that. You have people working in multi-functional teams, focusing on the same goals, and all trying to do the same thing, all trying to reach the same end. Yes, they have different roles within those, but they’re all working together as part of one team. And that team bears the responsibility for the outcomes, right? We used always say, ‘The unit of delivery is the team’. You could never come back and say, ‘Well, you know, my product idea was brilliant, but the team stuffed it up. No way! That would be totally acceptable. 
Tom Burton: Okay. Well, thanks very much Paul, and best of luck with your next adventure. 
Paul Shetler: Thanks very much. Cheers.