James Riley: James Riley here, from innovationaus.com. With me is Paul Shetler, the former Chief Digital Officer of the Australian Government, but, prior to that, the Chief Executive of the Digital Transformation Office. Welcome, Paul.
Paul Shetler: Really good to be here, James. Thank you.
James Riley: Alright, now, talk to me about the difference between, you know, the obvious difference, between a brownfield site and doing smart things with completely new technology in greenfields.
Paul Shetler: Well, if you’re a greenfield site, you don’t have the option really - or, you do have the option, but you’d be kind of crazy to take the option - of using old technologies. If you’re a greenfield, what you probably want to do is you want focus on your business, as a start-up, probably the last thing you want to do in the world is buy off a lot of servers and pay off a lot of Oracle licences, and a lot of Microsoft licences, and a lot of Solaris licences, or a lot of whatever licences, you probably just want to go on your credit card to Amazon, spin up a few instances of whatever it is you want to do, and start working immediately. So you - the way that you even think is very different, right? I had a couple start-ups when I was in London, and, you know, it was wonderful. We could do things we could never have done even five years before that for the simple reason that we could just go direct to cloud. It meant that all those things that you normally, maybe ten years earlier, five years earlier, you would have had to do, like buy servers, just didn’t happen. But government doesn’t think that way, see? In government, we have all these servers, we have all these data centres, and we have all these people who know their data centres and who know servers, and who have built an entire career on that. So you know, people tend to do what they’re paid, and people tend to like to keep their paychecks. So if you’ve got a certain skillset, and a certain way of doing things, it’s going to propagate itself in your organisation.
James Riley: OK, well, look, let’s just jump onto that in government. I just want to walk through a timeline very quickly. The Digital Transformation Office, set up or announced in January 2015, you came on board July 2015 --
Paul Shetler: July 13th 2015, a date that will live in infamy.
James Riley: Alright. We’re not talking ancient history, though, are we?
Paul Shetler: No.
James Riley: So now. And then, the big change, the Digital Transformation Agency, as the DTO was transformed into, was October 2016, when you were made Chief Digital Officer. And then you resigned from that position in late November last year. Now, in the intervening period, lots of issues with myGov, there was the CensusFail, as it’s heroically known now, ATO problems in December, when they were shut down, I think for a few days, and obviously there’s some there’s issues around Centrelink and ongoing thinks in DHS and DSS.
Paul Shetler: Yes.
James Riley: OK, so, so that’s kind of - that’s the life you’ve been living for the last 15, 16, 17 months. You’ve been out thinking on these, you know, big, brownfield issues. So what’s a way forward an organisation, be it government or private, large, legacy-based: how do you start unwinding that stuff?
Paul Shetler: So, in all the conversations we’ve been having so far, we’ll go into more detail about some other stuff a little later on, but there are three big points I want to make first that are kind of overriding. And everything else will be to enable those or make those happen. The first one is you have to have people with the right skills, skills and mindset, to do this work properly, and one of the problems that we have in government, and one of the problems you probably see - you know, I used to see it in banking, and in exchanges and other places where I used to work - was that, you know, people fell in love with the first technology they learned. The VAX was the most amazing computer ever made and there was never anything anywhere near that. And so when I was, you know, at Republic Bank, we inherited a number of VAXen, we used them, we had a couple nodes, and we had to do stuff with them, and it was difficult to get people off them, it was difficult to sort of wean staff off them. So skilling people up so people actually have the skills in product management, in user research, in design, in web-ops engineering and so on and so forth is massively important. And it’s massively important for Australia as a country. Because, if you want to pursue a digital economy, you have to have these skill to create the products and deliver the products in the first place, otherwise you can’t even do so. The talent pool in Australia is very very broad, and very very shallow, very very distributed, but very very very shallow, so it’s very hard to find a critical mass of people. So if government, let’s say were to, sort of as an industrial policy almost, to stimulate the demand for these skills, that would be a tremendous thing, not only for delivering better services, but also better outcomes for Australia as a country. So that’s one thing, we need to upskill, so we can actually feel comfortable using 21st-century technology, just like everyone else in the world does, and so that we can deliver and design the right things. Secondly, we need to - we keep on saying this, you know, it sounds like a mantra, it sounds like a truism, but it’s true, it’s absolutely true - we have to put our users first. Now, if you look at, say, Centrelink, they obviously didn’t. They put their government need to collect debt first, and it’s a legitimate need, no one’s denying that. But they put that need first, they took the processes that they had been using to reconcile the differences that came out with those errors, the 20% error rate, that they were doing with their algorithm and with their mismatched data sets, and they pulled those people off. And they said, ‘We’re just going to dump those results right now onto our users, let them do the hard work’. No, that’s completely unacceptable, and it just, you know, it shows complete failure of service design. Thirdly, you need to get rid of the split between policy and delivery, where you have one organisation that sort of, in a very command-and-control way, has these big brains, you know, and they sort of think these big ideas, and they give these big ideas to these sort of little proles, and they say, ‘Oh prole, you go and you implement this wonderful idea I had.’ You can think of that in the sort of DHS/DSS split we have right now. And all that leads to is finger-pointing and blame-shifting, because you can - you know, the delivery agency can say, ‘Well, I did a great job of delivering it, but the policy was rubbish,’ and the policy agency can say, ‘Well it was a great policy, but the delivery of it was just crap.’ And because there’s no feedback loop between the two, nothing ever happens. Now, no modern organisation works that way. There’s always - the whole idea of any kind of lean or agile organisation is to have a very, very tight feedback loop. And what you’ve effectively done with those types of things if you’ve severed it entirely. So those are, like, the three things which have to happen, have to happen, and everything else which we’ll discuss after this are sort of ways of making sure that actually does happen.
James Riley: Right, OK. I’m kind of interested in this split between policy and delivery especially. So unless you’re going to get some Cabinet officers sitting on project delivery teams, it’s kind of tricky. So how does that work inside a government organisation?
Paul Shetler: So there’s a few different ways it works, right? There’s a few different ways you can make that work now. I think, longer-term, the idea is to - is to reorganise government. I mean, if you’re talking about a fundamental transformation of government to operate in - with internet time, and with internet quality, we need to start organising ourselves that way, and stop this kind of insane split. And I think that there are, you know, a number of people in Canberra who understand that this is a problem. This is not just me saying this. This is - a lot of people believe this. It does need to be fixed. That would require changing the structure of departments. Changing the structure of departments is not the same thing as changing government. It’s not a constitutional matter. It’s a very different thing. It’s a question of changing the way the bureaucracy works. But even before that can happen, when we are developing policy, there’s no reason on earth why we could not start using some of the same techniques that we use when we are designing services. It’s very very similar. Policy design is really service design writ large. It establishes the parameters in which any kind of service operates, so there’s no reason why we shouldn’t be talking with users, why we shouldn’t be interrogating users, no reason why we shouldn’t be looking at quantitative data as well as qualitative experience of people while we’re putting these things together. And there’s no reason also while we’re delivering services that we shouldn’t have policy people on our teams. Now we did that when we were doing the exemplars, we made sure we had policy people involved in the delivery of the exemplars, because often times when you’re trying to deliver a service, you’ll find out, ‘Oh, but we can’t do this because, ‘Policy’,’ or, ‘We can’t do that because, ‘Policy’’. In some cases, having a policy person there is helpful because you can either change the policy, or you might not understand, ‘Actually, that’s not even the policy at all, it’s just folklore’. So getting that - getting those - even a mixed team on a delivery team, having - having policy people embedded into it, and using service design techniques when putting together policies, is a step in the right direction, even now, that we could take.
James Riley: So, obviously you spent 16 months or so in the DTO/DTA, you’re now not in that organisation, I have to ask you - your reasons for leaving are obviously that you’re frustrated, you seem frustrated, so these are thoughts that have come after your resignation, so what was the, you know, the genesis of these sort of frustrations, and how did they play out?
Paul Shetler: They have everything to do, really, with this policy-delivery split. It’s my bugbear. I just don’t believe that you can deliver a service - I don’t believe you can create a policy having a feedback loop to the actual results on the ground. To me, policy is not a document, it’s also what happens in the streets. And, if you don’t understand that, if you can’t do that, you have no idea what the effect is that the policy’s actually having, you don’t know if it’s actually good or bad, or if it’s doing what it’s supposed to be doing or if it’s not. And on the delivery side, if you try to just sort of do delivery - well, what are you delivering? You’re delivering a service - OK, well, why are you delivering a service? You’re delivering it because, ultimately, there’s some sort of government policy involved there, right? So you’ve got to tie these things together. And my concern was, pretty bluntly, that, DTO - you know, we say, ‘The strategy is delivery’, which is an overstatement, but sometimes you have to make overstatements to make a statement at all, to be heard - the strategy is not only delivery, there is more to it than that, but, if you don’t deliver, there is no strategy, you may as well not have one. And I believe the same thing with policy, if you don’t deliver, you may as well not have one. We were very much a delivery organisation. We were focused very heavily on that. We got some of the best people in Australia working with us, we recruited international talent, we recruited talent from the open source community here, and we got people who really cared deeply about our mission, which was to make life better for Australians, and to use government to do that, and to be able to work in a modern and lean and agile way, which you know, even in many private sector firms you can’t do. So we were doing all of this, and I think we were quite successful, in fact so successful that DTA had additional responsibilities given to it based on the results from DTO, but, unfortunately, from my view, there was a shift - it was decided, at the government level, to pull back from having a small central agency do as much delivery. I think also, being inside the portfolio of Prime Minister and Cabinet, they probably didn’t feel that comfortable with delivery, PMC is not traditionally a delivery portfolio - and in general, in the Canberra culture, delivery is something you ask other people to do for you, whether they’re vendors, or whether they’re delivery agencies, it’s something that, you know, the big-brained bureaucrats don’t do.
James Riley: So, I know you’re not a massive fan of the word, ‘Disruption’, in what you talk about -
Paul Shetler: No.
James Riley: So what about - let’s talk about the nature of transformation, when we’re talking about transformation in government, or in a particular part of it, what are we talking about?
Paul Shetler: We’re talking about a few things. In the case of government, from the outside, from the outside what we’re really just talking about is providing clear, fast, simple services. To the outside users, that’s all that government transformation means. And it’s really easy to tell whether you’ve done that or not, because you can compare yourself to, say, Uber, and say, ‘Well, how easy is it to use your service?’
James Riley: What about on the inside?
Paul Shetler: To do that - on the inside, there are a lot of things blocking us from doing that. The ability to work to deliver things quickly, to design things around actual user needs, is stymied at almost every turn. And, you know, when I joined GDS - not GDS, the Ministry of Justice, before I joined GDS - I was told by people at GDS that part of the purpose of the exemplar programme, initially, was to have people like me come along and to try to deliver a really good digital service inside the current government context, and they likened it to running through a minefield and seeing what was there. They said, basically, you know, ‘Paul, we want you to run through this minefield and see what you find!’ Because they don’t know. They don’t know everything that’s happening, right? And what that allowed them to do was to see the things that needed to be changed. So when we’re talking about transformation of government, we’re saying on the one hand, ‘Yeah, we want to deliver these brilliant services, cheap, good, fast, clear services,’ but we know that, to do that, there’s going to be a lot of things from procurement to HR to technology to governance to et cetera that need to be fixed.
James Riley: Alright. And this is not something you do - I mean, to follow up from that, this is not something that you do as a compromise, it’s an all-or-nothing, it’s a root-and-branch, or it’s a slow-and-steady, what, what --
Paul Shetler: I don’t believe that - I think that because of the nature of particularly older institutions, where there is a strong sense of history, heritage, whatever you want to call it, you can call it inertia also, if you don’t keep on making steady progress, you will be enveloped slowly by the bureaucracy and it will congeal around you and it will stop you, so you need to move quickly, perhaps not always as quickly as DTO did, we moved very quickly, but you need to move quickly and you need to keep that pace, you can’t slow down. When you slow down, typically, is when you will lose. Because the resistance will be very strong, you know, there’s a lot of - nobody wants to transform the way they work from one day to the next. It’s not something where people wake up one morning and say, ‘Gee, I want to transform the way I do my job today.’ No one does that. People do it because they have to. In this case, we have to because what we’re offering is not good enough. But, if you’re living in an echo chamber, where people are saying, ‘Well, what we’re providing is great,’ then there can be all sorts of reasons for slowing down.
James Riley: Alright. Well, let’s - let’s go back to those kind of big examples from 2016, The difficulties with myGov, Census, ATO and then Centrelink. So I mean to my mind, I mean - look, large organisations have trouble with IT, it does happen, but to my mind 2016 was a big year for it, within a government that’s put up a sign saying, ‘We will be the government that transforms these things’. So what - what is happening that the frequency of these issues is growing?
Paul Shetler: OK, so when I came in in 2015 - I think it was July 23rd or 25th or something like that - I gave a speech after having been here a lot less --
James Riley: In IPAA?
Paul Shetler: Yep.
James Riley: The Public Administration --
Paul Shetler: In IPAA, in Canberra - I’ll send you the link, James. But all you have to do is look up, ‘Paul Shetler says, “We are failing”’ on Google and you will see the articles, they all come up. And I pointed out that the public service was failing the people of Australia, and that we could expect more failures if we did not change the way we worked. And I’d been saying this since day one, I’d been saying this when I was in the UK, because I could see it there, and what we were doing a lot to fix it, we made lot progress, but I could also see in Australia, you know, while I was interviewing for the role here, there was a whole series - I think it was May of 2015 -- there was a whole string of articles about myGov failures, and it was kind of shocking, shocking to me what I was reading, and also the response from DHS, which I didn’t think was really impressive. So I thought to myself, ‘You know, this is not - this has to be flagged up as a real problem’. Since then, you know, DTO, we tried to fix where we could, but our remit was limited. DTO’s remit was not to go into the back office, DTO’s remit was not to change business processes, DTO’s remit was not to change policy, DTO’s remit was to change the user’s experience on the front end. But you can’t change the user experience without actually going deep into the back office, because the service itself isn’t just the user interface, the service itself is the people, the processes, the policy, the systems, and the interface, and everything else that supports that service. The service can be the buildings in the shopfronts, that’s part of the service, right? If they’re rat-infested, or they’re nice and clean, those are two very different experiences, right? Even if the system experience is lovely. So all those things come together when you’re talking about a service or user experience. We had control over a very small part of that, and so my goal was always to try to flag that we need to do things better. Now, ultimately, DTA got given responsibility for the back office, you know, we pushed like hell for this, and we got it, and that was really important. I think it’s just really important that DTA also then is able to make sure that happens. Now, the reason why these things happened, is, again, you know, getting back to it, that these things need to change. You have a very de-skilled public service, very big, big IT departments but not a lot of IT skills in those departments, facing off against big vendors who provide lots of - lots of bodies, so lot of cost, not an awful lot of capability, and very little willingness to use modern technology. And I think if you look at both the CensusFail, and you look at ATO, you can see at the very bottom of that was an unwillingness on the part of ABS to actually take ownership of what they were doing and to sort of try to outsource the risk to IBM, which is wrong, and never works, and I said that when I was - I said that even before I came here, I said that throughout the whole time I was here - you can’t outsource responsibility, it just never happens, right? You can never outsource the risk. You’re the government, you’re the ultimate holder of the risk. You can’t outsource it - it’s a fool’s errand to try to do it, even though a lot of bureaucrats think they can. And vendors who are just basically you know, selling us incredibly proprietary and retrograde technologies that, you know, no start-up would be using. Now consider, you know, we keep on talking to the ‘innovation ecosystem’, or whatever it is you want to call it, here in Sydney and other cities, and we’re saying, ‘Yeah, yeah, yeah, you go guys,’ now how many of those guys are spending their money you know, buying HP services, IBM kit, WebSphere or anything else? They’re not. They’re going to Amazon, or perhaps Google or Microsoft, they’re getting their infrastructure as a service, they’re getting a few other additional services, and they’re spending all the rest of their time and money that they otherwise would have spent on all these huge capital expenditures, they’re spending it on building better and more competitive services.
The Australian Government should take a leaf from that, should follow the example of our start-ups. They should not be spending a huge amount of money on capital expenditure for stuff which is, frankly, yesterday’s technology. They shouldn’t be doing that. They should just be going cloud by default, they should be spending the rest of this money on service design and delivery.
James Riley: So what’s your - your prognosis for, you know, for the public services that you’re involved in or that you looked at as a member of the Australian government, what’s your prognosis?
Paul Shetler: I want to say -- ah. My prognosis is that things won’t change until the government takes it seriously. We can talk about transformation and innovation until we’re blue in the face, we can talk about ideas booms, and blah blahs, and, you know, some of these, you know, dancing holograms, all this kind of stuff - none of that makes the slightest bit of difference until we actually get serious and go to the root causes. We have a deskilled public service that is afraid of technology, that doesn’t know how to design, that is dependent on large vendors who are selling us retrograde technology and using retro methodologies. We’re not actually able to update our systems in response to what we know about users, when we find out more, we’re not able to actually update it, so we say, ‘This is not meeting the user need? Well, you know, too late, you’ve got another six-month cycle before you can update whatever system, you know, myGov or whatever’. You know, that’s just ridiculous. No company operates that way. So getting the basic skills in place is absolutely essential. Putting in a cloud by default policy, you know, you keep on hearing ‘Cloud first, cloud first, cloud first,’ and yet we keep on building data centres and we keep on having ATOFails and CensusFails, let’s just stop that nonsense, let’s actually get serious and stop the word games, and let’s just actually do it.
James Riley: So, structure of the DTO? What does that look like to you?
Paul Shetler: DTA has taken on a lot of responsibilities, which are good - I was really delighted that DTA has taken on responsibility for a program office, has taken over responsibility for ICT procurement reform, these are things which I pushed for for a very long time, so I was very happy they came in. I think we’ll have to judge by the results what comes out of DTA. I think that there’s a real risk that in taking on a lot of these responsibilities, the sort of [indecipherable] from Finance and other places with existing staffing and with an unclear mandate, that things won’t necessarily change. I don’t think it’s entirely clear what the program office is going to be doing, or how it’s going to be doing it, or how it’s going to actually impact service delivery, yet for us to know whether it’ll be successful, and I’d have to say a similar thing about ICT procurement reform, it’s not really clear yet what’s going to be happening, I know it’s been delayed quite a bit, so perhaps something will happen. Hopefully something will. There are some very smart people inside DTA, so there’s some incredible talent there, and I wish them all the best.
James Riley: So, tell me, what about for yourself? What are you up to?
Paul Shetler: Right now, I’m relaxing from having banged my head against the wall for sixteen months. So, the bloodstains and bruises are slowly going away. I’m enjoying going back to the gym, and giving me a chance to think about things outside of the bubble, think about other things as well. And I’m considering some opportunities here in Australia, and potentially other places as well.
James Riley: But this is your area, whether it’s public service or private sector, large organisations would be your --
Paul Shetler: Yeah. I’ve done the small start-up thing and I absolutely loved it, but so much of our lives, you know, is dependent on these very large entities that have existed for a long time, whether they’re banks, whether they’re brokerages, or whether they’re governments or hospitals or whatever. And, you know, those aren’t going to automatically just going to sort of become digital [click of fingers] by willing it. It’s going to be a lot of work. And it’s important that that happens, because they play such a big part in our lives.
James Riley: Alright, look - thanks for the chat. Was there anything you wanted to talk specifically about?
Paul Shetler: Well I guess there’s a couple of things. So digital transformation is, it’s something which people sort of like to talk about sort of in the abstract. Concretely, it means making lots of changes to the way that an organisation works, and those changes will not necessarily be pleasant for the people involved at all times. The pay-off is for the users, and ultimately for society as a whole, so that’s why it has to happen. But --
James Riley: Sorry, so you’re saying that it won’t be pleasant for the people --
Paul Shetler: Not necessarily for everyone involved. Yeah, because it involves change. And although we say ‘Change is good,’ and ‘We like change,’ actually, nobody actually wakes up in the morning and thinks, ‘Oh, I want to do everything differently today,’ and if they do maybe one day, they certainly won’t the second day and the third day and the fourth day, but actually, when we’re talking about this, it is that kind of level of change, it’s a level of structural change, level of change in responsibilities, level of change in terms of how we hold people accountable, how we are transparent, all these various things. And that means taking away some autonomy from some organisations to do whatever the hell they want --
James Riley: So you’re talking departments or --
Paul Shetler: Departments or branches or whatever you want to call them, right, within an organisation. In government, it’s going to be departments. So it means removing autonomy to sort of do whatever the hell they want and to say, ‘No, actually, we’re going to do things in particular ways. It actually means the centre being able to mandate and enforce ways of doing things. It doesn’t necessarily mandate what is done, and I think that’s a really important thing. The department still have to - you know, they have to decide what is done. They are the ones who have the users, they’re the ones who are closest to them, they’re the ones who have to deliver it at the end of the day. But how it’s done, how it’s governed, how it’s funded, and how we make sure it’s actually meeting whatever KPIs we’ve set up, in terms of satisfaction, cost, time and so on, that has to be centralised, and that has to be pretty much unsparing. That is not a change management process, that is not a thing where we sort of compromise, that is something where we hold very very high standards, and we have to insist on them, because the bureaucracy always resile from that, right? It will always try to go back to its safe space. The goal is to try to remove that safe space.
James Riley: OK. Well, look I’ll probably finish up with this, but those departments, or those branches, in any large organisation, very powerful in their own right.
Paul Shetler: Yes.
James Riley: Now, you’re going to have a very powerful agency. I wonder if you can talk to the importance of, ultimately, you know, what we would describe as political, whether it’s the executive government or whether it’s a board of directors.
Paul Shetler: It’s the single most important thing. If you don’t have the political will, because, again there will be costs, there will be screaming, there will be people saying, ‘This is not fair,’ people saying, ‘Wow, well, everything was OK until we made this change,’ and you will have fails, there will be things which go wrong in this, there will be things which hit the newspaper. You know, you will have to have somebody in charge who will say, ‘I understand these things. I understand all of this, and I’m going to make it happen anyway,’ and who’s willing to expend whatever political capital is required, probably because they also understand that you don’t keep capital, and you don’t get capital, unless you expend capital. You can’t just sort of miserly hold onto it and expect to keep it. You’ve got to actually expend it, and then get something back for it. That’s the way that, you know, the economy works, and it’s just sort of the way politics works. Now, I think in the UK we saw that with Francis Maude. He was there for five years in the Cabinet Office, he was uncompromising, he took on the Permanent Secretaries, he took on the management consulting companies, he took on the big vendors, and he also championed civil service reform, took on the unions, took on pretty much everybody. He also wound up having, you know, almost fanatical support in parts of the civil service, because we were allowed to do the right thing, we were encouraged to do the right thing, and we were held to do the right thing. We had to do the right thing. That was brilliant, you know, that’s why we came in there. It was difficult. It wasn’t complicated. Something Mike Bracken likes to say, ‘It’s not complicated, it’s just hard.’ It’s just hard. You’ve got to stick to it. Because there are a million things thrown in your way, and you just have to say, ‘I don’t give a fuck. I’m going to go through it, and I’m going to do it.’ And you need that there. And I think the same thing with any board, any government, just look at what works and look at what doesn’t work. We know that transformation by consensus does not work. You know, if an organisation could just sort of - if you could just sort of pass round a paper, say, ‘Hey, think about this.’ ‘Oh yeah, wow, wouldn’t it be great if we did it that way!’ And then wake up the next morning and the whole organisation was different? Wouldn’t it be a wonderful world. But it’s not that way. So, unfortunately, you know, we have to look at what works and what didn’t, we should not be repeating the mistakes of the past, but we should be looking at examples of what worked and we should go with that.
James Riley: Alright. Obviously the pay-off is huge, both economically and socially. Paul Shetler, thanks very much and we’ll look forward to hearing what you do in the future.
Paul Shetler: Thank you very much, James.