 Okay let's get down to it Drupalist. Welcome to an analysis of government science in Australia. In this presentation today you're going to see a comparative analysis between government sites in Australia and in other jurisdictions as well as in Drupal and other CMSs. Hi my name is Murray Woodman and I am the managing director and solution architect at Morft and most of the time I'm doing Drupal-y kinds of things in my day-to-day life but occasionally I get to work on an interesting passion project and this is very much one of those. The results you're going to see today have all been generated as a research project at Morft and some of the results we've got are very interesting and it gives me great pleasure to introduce them to you. All right before we get into it let's have a look at the methodology that was used in this research. We went out and we collected 1600 websites. Now these websites were as I said base in Australia and overseas but basically they were designed to represent in good numbers the various categories that we were interested in. We have content management systems, design systems and various segments as well that we're looking at so each site has been placed into one or more of these for analysis purposes. We're using different tools to do the analysis. It is automated and we're using Lighthouse to measure SEO, accessibility, performance and best practices and we're also using Wappalizer to measure the content management system. We've also written some custom code as well that will determine a security score based out of 100 and that's determined by the various headers, security headers that can be added to the response and we've also got some custom design system detection in our code as well. How has all this been put together? It's been put together with Drupal of course so we're using Drupal to migrate sites in from a spreadsheet and it's storing the data, analyzing the data, preparing statistics, doing the presentation layer as well as providing an API to talk to the outside world and we have a couple of processes that are running in Docker containers, Node.js, Lighthouse doing the scores of course, and our little scraper which sort of does Wappalizer and some other custom things for the security. So those two processes are running all the time and continually updating the results. Some technical caveats before we actually jump into the results. The first couple are around statistics. We understand the importance of getting good numbers in each of the categories. We've aim for like 50 or 100 as many as we can get reasonably get. The sampling we've done is not a random distribution across the entire population so we're not saying this represents all sites, you know, that are Drupal sites for example. What we're doing is we're searching for the most popular, the most noticeable, or the most relevant sites. So generally these are going to be high profile sites that are either doing well or are the flagship sites for a certain category. So we're really picking the best sites here. Lighthouse scores can vary so we're using some virtual machines to run this data. We haven't seen too much variation there though to be fair. Not all sites can be processed with Lighthouse. If they aren't processed and there's a problem they're not included in. And finally our design system detection is just based on string matching around classes and things like that. Your mileage may vary there but you know we've done our best. Also quality is many things. Now of course this presentation is basically presenting the sites and the findings in terms of quality but we're only measuring five aspects, those ones on the left hand side. But of course there are very many others such as design, the content, the popularity, return on investments, user satisfaction, these kinds of things. These are all very important when you're building a site and yes you do want to optimize those when you're building a site. So we are only using what is easily measurable, what is objective and what has value. And they're those five things that we've found and that's what our analysis is limited to. So let's go in and have a look at the results and rankings. Firstly the distribution of the scores. It's basically following a normal distribution which is good to see. If you remember back to school and university days you have pass and fail and you have credit as distinctions and if you're lucky enough high distinction. Basically we kind of see that distribution here so that's worth remembering when we're looking at the results. Very few sites are in the you know the 80 to 90 bracket and there's only one site in the 90 to 100 bracket so if you're up around in the 80s you're doing pretty well. Let's turn to the CMS averages. So this is probably the first and really the most relevant slide for most of us. These are the average scores we get for content management systems. What we're looking at here is a stacked bar chart of the different dimensions that we're interested in. So the longer the bar the better the score and you can see here that Drupal's doing particularly well and it's actually gaps out ahead of a lot of the other CMSs. Most of the CMSs do do quite well on accessibility and SEO that's the blue and the green columns there that's good to see that's the bread and butter of sort of web development but it's really around best practice performance and security that Drupal is shining here and that's what's allowing to take it out to first place. Looking at some other things Open Cities seems to be doing pretty well on their accessibility winning that score. Let's see Adobe Experience Manager not so good on performance there maybe that's a reflection of the number of scripts that are running on the page and you know most some other CMSs here not scoring so well in on the performance and the security side of things which is where Drupal seems to be doing relatively well. We also look at standard deviations and that measures the variability within a certain category so for instance here we have Open Cities with very low variability and the smaller the better so Open Cities got very tightly packed sites there. The other other CMSs more variability you might expect Drupal to be right down the bottom because it is a very wild system you can put it together in many different ways but yeah Drupal's coming middle of the pack there once again the you know it's around the performance and the security where we're getting sort of big variations in in terms of what the CMSs are doing. The design system so we've managed to detect a few design systems Australian government design system Gov UK and Ripple are the ones out in front. We do have some sites in New South Wales design system but not enough and we'd love to get some New Zealand design system ones in the WA as well but at the moment we don't have enough to to the show. Gov UK is the gold standard in this area and the Australian government design system is doing very well. Gov UK wins on accessibility Australian government design system just a little bit behind. Ripple overall is doing very well that's from a Victorian single digital presence they just seem to be getting punished on the performance side of things. I looked into that and view websites can get bad scores in in White House so that is something to to think about when you're looking at particular results. The standard deviation is very interesting Ripple out in front right very low variability similar to open cities it is a system that is regulated from top to bottom from design to theme to you know code base to the platform to the delivery all very consistent and you can see that it's doing very well there super low scores on variability around accessibility as well. Australian government design system though quite a bit more variability not too much more though than Gov UK but I think that is sort of illustrative of the fact of the variability of the different implementations that you'll see of the Australian government design system. The segments now now we've pumped in quite a few segments here Gov CMS and Commonwealth are quite correlated and so they're leading moving all the way down to more decentralized systems such as local governments and universities and in those cases there is no overarching design system that will bring it all together so you can really see that the consistency there is you know the platform the design system is playing quite a big role. On the distributions a little bit harder to sort of pick out patterns here I will just pick out Gov UK down the bottom now that's not the Gov UK design system that's all the government sites in the UK so you can see in the UK they've still got a super mixed bag so the design system may be doing well but the Gov UK sites still a lot of variability there you know Gov CMS is doing you know quite well up the top and Gov music NZ those sites are doing well just in terms of variability at least and we'll be picking out a few more things out of this the segments a little bit later so let's have a look at some of the findings and fun facts that we've seen along the way we looked at Drupal services panel and Drupal agencies top 40 on the Drupal marketplace Drupal services panel seem to be doing pretty well which is good to know just beating out the top 40 there it's a case of the cobbler's shoes though if you're aware of that old saying the cobbler's shoes has holes in it something along those lines you can see the Drupal services panel the people that build Gov CMS sites are not even scoring as much as Gov CMS sites so that would say that maybe Gov CMS is lifting their game or perhaps the various agencies are just putting more flashy stuff on their home pages you take your pick on that one the regulation I've mentioned this a few times I would say a unified stack from top to bottom certainly has beneficial impacts for the overall scores as well as reducing the the variability yeah local governments and universities are reinventing the wheel falling into the same old problems each site build and not getting any of that those gains from iterating on on a better system design systems represent an opportunity that's what I would argue here in the top line we have the different sites bucketed for the various jurisdictions and in the bottom line we have the corresponding design system variants you can see the UK to Gov UK design system huge jump there 64 to 77 and it's a similar story for New South Wales design system but very very few sites there and not so much for gap for the commonwealth but certainly yeah the design system is a determining factor there I would say I haven't focused on leaderboards I didn't want it for this presentation but for a bit of fun here are the top 20 commonwealth sites for the ones that I've reported on if your site's in there congratulations but what do these sites look like are they boring or are they interesting well they're really a mixed bag the winner for all the sites is trees boy tarts blogger that's the one on the left now that is very plain right no images scoring in the 90s it is the winner well done trees but the second site the attorney general so on Gov CMS very closely aligned to the strain government design system it is only coming a little bit behind and that site there has you know got you know graphics and hero graphics and whatnot so it shows you can have a good-looking site and still get good scores and a shout out to the Gov CMS site who's also coming in the top 20 there that's got pop-up videos and you know javascript and you know various things so it too is is doing well so it shows that you can get good scores you know with graphical base websites so in conclusion let's run through them automation for the win you know thank you lighthouse and wapalizer your tools have been very helpful Drupal performs very well it is a wild system there's a lot of variability a lot of degrees of freedom but it does score well and it does okay on the consistency side of things most CMSs are doing very well and accessibility and SEO this is the bread and butter of web development but performance and security really are making the difference here I think there's easy wins on the performance side of things and I think it is the dedication and skill of the team to bring best practices and performance to the mix to lift it up to that high level the consistency of the platform designed to delivery determines the overall quality and the consistency of the results that's one of the main takeaways of this presentation if you need to know more it's easy to find out all you've got to do is go to optimal.site-showcase.com you'll see all of the results you've seen here today plus a whole lot more including leaderboards go there click around have some fun I'm sure there's something there you will find of interest. Phew that's me time's out battery's done over to you guys for questions thank you. Hey Marie one of the questions that has come in the live Q&A is best practices have you got an example of what was as bad? Yep hi there everyone well the best practices are the ones that are reported right now so what I'd really encourage you to do is that as a little you know browser extension in in Chrome you'll then be able to to see the the reports there so you know if you've got a site that you want to test you can very easily do that with you know the lighthouse extension when it does that it will give you a whole stack of recommendations the best practices is really interesting you pick up on that I did some correlations between all the different factors it's actually best practices which is very strongly correlated to the security and the performance so it shows that if you have a team that's interested in best practice you're probably going to do well on security and performance as well to answer your specific question it's a bit of a grab bag there in in those you know best practices and really coming to mind they they tend to be quite finicky little little things where you really need a team who knows what they're doing to to tweak it all up so yeah just go install lighthouse and have a look I would say there is another one in the discussion forum any specific examples of Australian government sites that do everything very well yes so it's so funny that question came up and then there was a slide with the top 20 leaderboard there of Commonwealth sites but if you go to that optimal.site-showcase.com you can browse all of those different categories that you just saw and there'll be the top you can go see the top 20 GOVCMS top 20 Commonwealth even the top 20 Drupal agencies there that's that's um and the top 20 Drupal service panels as well that's interesting reading if you want to check that out okay but yeah to answer the question I mean you know the Attorney General did very well you know the GOVCMS site there uh did pretty well as Australian government legal service and you know a smattering of others as well so I'll leave that to you to browse at your leisure all right I don't think we have any other question either in the discussion forum or in live Q&A um last chance people ask your questions if you have any come on guys like I I recorded that at superfast speeds I'd have five minutes at the end so and we still have let's use the last minute 48 yeah yeah we still have one minute 45 seconds so if we have some questions showed away otherwise we'll just end the session all right most important thing on a great website sorry what's the most important thing on a great website well you have those five things that we were measuring but then you had all those other things as well so I would argue I've presented you know five very important things for quality and you know lighthouses based around those things because that directly translates into better user experience and those numbers get pumped into their their search engine algorithm right so they are you know considered important certainly accessibility right you got to look after your users and SEO you've got to bring the users in and you know that's why the CMS is a concentrating on that um but of course you know you've got all the other things there as well I would love to try to incorporate some of those other things into the analysis um guys just as well like I will be popping over to the morphe booth after so if you want to talk about this in more detail please you can pop by there and we can have a chat and there's one question how long how long will you maintain the site very good question again um it was a pleasure to make that site and it was like a little sort of showcase experiment thing right you know where but you know it's turned out pretty well if people want us to investigate a category let us know and you know we might be able to put that in and you know augment the results uh that are there or if we want to investigate other things so I'd say we would still be keeping it going for a while that's the aim you know let's uh let's let's try to improve the quality of sites out there and there's the last question we are calling actually oh we can still go okay any government sites assessed during using static web any sites used any government sites except using static web I don't know now we there WAP Eliza reports is no CMS category which you know it's quite a number of sites in there so you know you can't detect everything so of course there's going to be you know a lot of variation out there in CMSs used so you know I can't really say how how the site was generated that that's a bit hard for us to detect I'm sorry but we'll have to cut short now um thanks a lot more it was an awesome presentation okay thanks everyone thanks for questions bye bye