Bratislava World Usability Day 2016 and government IT

I wrote about sustainability and design takeaways from Bratislava World Usability Day 2016 in my previous post. World Usability Day (WUD) 2016 was organized on November 10th, 2016 in many places around the world. Theme for this year was Sustainability, but for us, working with and for the public sector, it was even more attractive thanks to the guest from UK and Estonia government agencies that implement or oversee the government services – services for real people, citizens. Services that should serve – just like the state itself should. And that is very touchy topic here in Slovakia.

Videos from Bratislava event can be found here, the page is in Slovak, but videos are easy to find and are in English.

Estonia: pathfinder or e-Narnia?

Risto Hinno came to us from Estonia, the state renown for it’s level of e-government. But if you imagined their world as a perfect place with flawless services you’d be wrong. Risto came to talk about their approach to the services and the problems they had to overcome and are overcoming.

Estonia and Slovakia are both countries from the Eastern Bloc, Slovakia is the successor of Czechoslovakia, while Estonia is one of the post-Soviet states. Both states are in NATO and EU and both use Euro, but there are also some important differences. I may not be historically accurate, but while in Slovakia we still have plenty of “old guard” people in their posts (like judges) and plenty of old-thinking politicians, many of them previously members of Communist party, now often using the sticker saying “social democrat”. In Estonia most of these were Russians and they simply were gone after Estonia became independent. And that allowed for deeper change, change that is much needed here in Slovakia but haven’t happened. Some ask: “Will it ever?”

But back to the services. As Risto put it, what we (citizens) want is value, but what we typically get is bureaucracy. The answer to this problem is to make everything smaller and simpler and really focus on the value.

Problems small and big

But just as with value-vs-bureaucracy problem there are opposite forces in play here. Even when the stakeholders agree on delivering maximum value for the money they often don’t agree on how to do it.

Very often the expectations are big and the budget follows them. Very often we don’t respect the systems our users already work with. And very often we deliver little value for a lot of money afterwards. Or worse, we often make the life of our users harder and they simply can’t understand what are the new system advantages we are talking about.

It is very important to understand that we need to deliver value in small chunks. Many times in my career I’ve heard: “…but we can’t deliver this useful system in small!” Really? How do you know you can deliver it on a bigger scale then? History shows us time after time that megalomaniac plans crumble. And, to make matters worse, they crumble often over many, many years.

Managers often expect that developers can plan their work while the developers have trouble to account for all the complexity in advance – often the accidental (that is “not essential”) complexity. And the accidental complexity always gets higher with bigger system, there is simply no remedy for that. Analyse as much as you want, you find out something unexpected the minute you start coding. Or when you meet with a customer. These are truths known for decades now, but still they seemingly make no sense to many managers and other key decision makers.

And so far we’ve only talked about mismatch in beliefs how to build complex systems. What does it matter whether you want to “build it” or “let it grow”, whether you are forced to “fixed time, fixed price” contract or you can do it really incrementally using whatever agile is currently chic – this all is not important at all when the true reason to spend the money is… well, to spend the money!

Yes, public money, aka nobody’s money – who cares? People care, of course, people who are in the chain somewhere. People who decide who should participate and have some piece from that big cake – competent or not, doesn’t matter. There are always subcontractor that will do it in the end. Money talks. And value is just standing aside. Just as users and their needs do.

It can be scaled down

Of course, it can, the question is whether we dare to be accountable and flexible to deliver clear value for the money. Value that is easy to see and evaluate whether it’s worth it or not. In Estonia they are also far from perfect, but they try hard to keep it small and simple (KISS principle). They limit their evaluation/analysis projects to 50k Euro and implementation projects to 500k.

I saw people laugh at this but 500k in these countries is a reasonable cost for 8-10 person team for a year. Yes, you have to mix juniors and seniors, which is pretty normal – and no, you can’t pay for 3 more levels of management above the team. Get out of their way and they will likely deliver more value than a similar team in a typical corporate environment that has to spend 20% of their time with reporting and other overhead (and that’s a low estimate).

If the cost calculation doesn’t work for you, take less people and make the project last half a year, not full. I’m not to be convinced that there is no way to deliver visible value within 500k Euro.

Risto Hinno also mentioned another very interesting thing. They decide how many services – or how much work if you will – they want implemented at a time. This way they prevent IT market in Estonia from heating up too much because that leads to very low quality. Companies start hiring everyone and anyone, a lot of code is written by external workers who often don’t care and everything is also done at way too high pace. These are all recipes for disaster. Things they seem to know in Estonia, but not here in Slovakia.

Problems with services

Risto talked also about typical problems they faced. The learned the hard way that services must have an owner. He also presented the maturity model of the services. Using my notes and not necessarily exactly his words the levels are:

  1. ad hoc services,
  2. simple list of services is managed,
  3. services have their owners,
  4. services are measured (including their value),
  5. service management is a daily routine.

He talked about building measurement in the services. This part of the talk rang a lot of devops/continuous delivery bells. And he also talked about the future visions they have:

  • Base future services on life events. This makes them more obvious to their consumers, that is citizens.
  • Aggregated services – many simple services can collaborate to achieve more complex scenarios. Risto actually mentioned some crazy number of services they have, but also noted that many of them are really simple. Still – it’s easier to put together simple blocks than to cut and slice big blocks.
  • Link between public and IT services.

So Estonia seems to have started well and they keep improving. I wish they keep on track because I loved the ideas presented – and many of them were familiar to me. I just needed to hear that it actually somewhere works. And now it’s time to get to the next level.

Designing the next generation of government services around user needs

That was the title of the presentation by Ciara Green who came to tell us how they do it in the United Kingdom. She works for GDS, Government Digital Service and she talked about the transformation of government services that, if we simplify it, started around 2010 with quite a short letter (11 pages) by Martha Lane Fox after she was asked to oversee a review of the state of the government services at the time. Sure, 11 pages seems long for a letter, but it was short in a world where you likely get hundred(s) pages of analysis that is not to the point in the end. The letter was.

After this government services all came under a single domain and many other good things happen. UK is way ahead of Slovakia, historically, mentally of course (despite the Brexit and all the lies leading to it) – so it doesn’t come as a surprise that they decided to focus on value and they also used current agile methodologies.

They knew what happens if you deliver over many years and then surprise your customer or users – invariably not a good surprise. So they started to deliver fast and often, tested a lot, tested with real users including seniors, focused on UX. Just as Risto, Ciara too argued for making things simple. It is very easy to do things complex and longer and we should do the opposite. We should start with needs, real world needs, remind us these often. And we should do less    (reminds me the powerful “less but better” mantra).

Another interesting point was Good services are Verbs. Bad services are names. Of course there are also other components, various registers, but in the end the focus should be on services and on the activities (e.g. life situations) they cover. Sure, the verbs are a bit unusual sometimes. One very important service is called Verify and it verifies the identity of the user with various partners (banks, Royal Mail, and more) because in UK there is no central register of citizens. So they can do this without keeping personal data (I don’t know the details) and here in Slovakia we build various registers for years and they often add more problems than they solve.

Funny enough, when she talked about some services and the name was used, it still functioned as a noun in the sentence – quite naturally. So I believe the word class used for names may not be the most important thing but using verbs may remind us what we try to solve.

Back to Slovak reality

Ciara’s talk was pure sci-fi for us. She works for government agency where they develop services in an agile way. How crazy is that? Pretty much, if you say it in Slovakia. Slovak services are portioned between many companies, most of them with political background (not official, of course), and we spent around 800 million Euro for government IT that looks like a very bad joke. Each ministry takes care of its own sandbox and if there is some initiative to unify how IT is done it is executed extremely bad.

For example, there is some central portal for public services that acts as a hub and connects various parties in government IT. However, this “service” is good mostly for the provider of the service, not for its users. The protocols are crazy complicated, if you need to connect to it (you want to or is forced to, which is more likely) you need to conform to some strict plan for testing, etc. There is no way to do it agile, it only separates you from the service you want to talk to. It adds another barrier between you and the other party, not only technically but also organizationally.

It is said that one minister mentioned to a young woman working at the ministry, horrified how the state works, that she should not be naive, that sometimes things are as they are and we have to be realistic. He, reportedly, pointed at government IT and the bad companies who suck money out of it. Now this is all a matter of speculation, but the words could have been said. The tragedy is twofold.

First: The companies do what they are allowed to do. It is not that bad companies do whatever they want, they do it with connections to the officials of the government and various bureaus. As crazy as it sounds, stories that someone who worked for some company now works for the state and manages projects his previous employer delivers. Stories like this are uncovered on virtually a daily basis now.

Second: Even if it was true and the bad companies did whatever they wanted… then the state totally failed to do its basic job. It actually did fail in the first case too, but here it seems to be a very weak state, not the state our officials depict to us.

Final words

While the Slovak reality is pretty bleak, it was very refreshing to see that it can be done and how it’s done somewhere else. It’s nice to see that agile can work, even more – it can work in a state agency. And that state agencies can deliver true value, when they really focus on it. We have also learned that state can regulate how much he wants at once. This can – and should – be done in IT, but also in infrastructure projects like motorways (another anti-example from Slovakia). It gives you better quality for lower price and surprisingly it still gets done sooner in the end!

In any case, there is a long way for Slovakia and Slovaks to get to the point when we can focus on value and don’t have to fight with elemental lack of political culture (to which I include wasting/misusing/frauding public money as well).

Neither Risto nor Ciara brought any political topics in, but some of the Slovak political “folklore” obviously affected the questions that were asked afterwards. Corruption was mentioned not once. But these were areas where our speakers couldn’t help us (oh, how I envy them).

The presented topics were so interesting for us that UX parts were often left aside a bit – although focusing on value and user from the start is pretty useful recommendation. But as with anything simple, it is much harder to do it than to do something complicated and big.

AWS did it again, this time with Lightsail

I still remember how AWS (Amazon Web Services) re:Invent 2015 conference impressed me. AWS is extremely successful but they keep pushing innovations at us. I played with their smallest EC2 instances couple of times, considering whether or not I want my own server.

This year at re:Invent 2016 happening during this week Andy Jassy, CEO of Amazon Web Services, announced many new services in his keynote. One that resonated with me immediately was Lightsail. It is extremely affordable server with plenty to offer even in its weakest configuration. For $5 per month you’ll get 512MB of memory, 1 vCPU, 20GB SSD and 1TB data transfer. See also this blog post for more information.

With such a reasonable base price I decided to try it – by the way, the first month is free! My plans were nothing big, I actually just wanted to move my homepage there. But you have that flexibility of a Linux box ready for you anytime you fancy.

I spun my Lightsail instance in no time, choose AMI Linux just for a change (considering my Ubuntu preference) and with Google’s help I got nginx and PHP 7 up and running very quickly indeed. I used the in-browser SSH but because it’s not quite the console I like (but close, even Shift+PgUp/Down works as expected) I wanted to put my public key in ~/.ssh/authorized_keys. I didn’t know how to copy/paste it from my computer, but when you press Ctrl+Alt+Shift in the web SSH it will open a sidebar where you can paste anything into the clipboard and right-click will do the rest.

I really liked this experience, I like the price as well and in-browse SSH comes very handy when you are at a place where port 22 is blocked for whatever reason. (I’m sure it has something with compliance, but don’t want me to understand.) I’m definitely keeping my new pet server although I know that cattle is more common now. Especially in the cloud.