 Mae'n meddwl i'r llwyth ar gyfer Calimera a Calyspera. Felly, rydyn ni'n meddwl i Calimera, ac rydyn ni'n gwneud yn y lluniau. Rydyn ni'n meddwl i'n meddwl i'r llwyth. Rydyn ni'n meddwl i'r llwyth, ymgyrch, ymgyrch, ymgyrch, mae'r 50-person digital agensu ar hyn yn Bryton yn y DU. Rydyn ni'n meddwl i'r UK, ymgyrch, ymgyrch ar y stryd yn eu. Maen nhw'n meddwl i'r compares defineiddiau siaradus Wel, rydyn ni'n gweithio i'r ddechrau o'r ddigon o'r ddigon, ar y drafodau yn ymddiadol, ar y cyfnodau PDI. Dyma'r ddigon rydyn ni'n gondol o'r ddigon, o'r ddigon o'r cofi, o'r newid ymddydd ymddydd fel y cyfan. Rydyn ni'n gweithio'r Lundon, Ac rydyn ni'n meddwl i gael glas, steel a'r fawr, yna'r ffiannwch yn ddechrau Llywodraeth, ac rydyn ni'n meddwl i'r gofyn ar y ffiannwch, nad yw rwyf wedi'i gweithio'r busnes yn ei ddweud o'r gweithio'r gweithio'r gweithio. Rydyn ni'n meddwl i'r gweithio'r gweithio. Rydyn ni'n meddwl i'r gweithio, mae'r gweithio gweithio, gweithio'r gweithio arall, a ddweud o gweithio gweithio'r gweithio, Felly, yw'r hyn o'r gweithio, dwi'n ddataeth, dwi'n ddau'r ffordd, a dwi'n adfotio. Felly, dwi'n adfotio'r platform o'r cyfnoddau yn y sector, ac y sector yw ymddangosiaeth ymddangosiaeth, oherwydd ymddangosiaeth ymddangosiaeth ymddangosiaeth ymddangosiaeth. Felly, mae'n ffordd gweithio'r ffinans. Mae'r ffordd Ynw'r Ymddangosiaeth, a dwi'n gweithio'r ten oed, a fyddwn ni'n adwsio 4 o farwy ypaf, a fyddwn ni'n fathio'u ffordd dwi'n dod, mae hyffordd ymddangosiaeth, ond mae'n ddyliadau cyrraedau, maeswrfa mewn, a dwi wedi'i bwysig iawn y dweud y brans, yw'n ddweud ar gweithio'r ei gwentau, products, a bydd ynglyn â hwnna'n gweithio. Mae'r marketing e-mail is also particularly important for them. There are important things that happen in the industry and people want to get notified as quickly as possible because it affects the market, so sending out e-mails with alerts and updates as well as regular newsletters is super important. So this is, to set a bit more context about this programme, Mae'n ddweud beth sydd wedi bod yn cyfnod o'r site ar gyfer y cyfnod. Mae'n ddweud beth sydd wedi cael y grwp o'r cyfnod o'r ffordd. Mae'n gweithio newid nhw, ac mae'n gweithio'r gwaith sydd yn y program yw'r byw. Mae'n gwneud fy ngyfans yn awr. Mae'n gweithio'r website ac mae'n gweithio'r gwahanol, yn cyfnod o'r newid yma yn ei wneud y gwaith. cyfarallydd. And make a plan to get there. They want to achieve significant growth, they've just had a round of investment and of course what investors want is their money back. With some more money on top. So I can do that we have to achieve growth. So making sure that we take their current business, which incidentally is already doing very well. Felly byddwch is often seen as hace un industry that's in trouble because it's hard to monetize and if you're in a business of producing magazines everything is online. Publishers son seem to be in trouble a lot of the time. This company isn't, they've been going superfast anyway so a key job for us is not to screw it up and then deliver growth on top. The premium products piece is worth noting as well. Their clients are some of the most wealthy, influential people around the planet. These are the folks that have their own private jets. So if we produce a product that looks like it's cheap and tacky, they're going to lose market share to their competitors. The final part of this is that right at the start of the programme, my key contacts vision was very much the same as ours, which is that there's really a kind of polarisation now in enterprise software. Either you go with a typical enterprise software product, so Adobe or Salesforce or Sitecore, one of these kind of huge, big pieces of software that can do everything, and then you learn how to configure and drive it better than your competitors. And that's a really valid strategy for a lot of businesses. But there's another way, and that other way is to design and build your own digital infrastructure as part of your business. So you have to embed that digital capability, the knowledge, the know-how within your business so that you can start with a business requirement, identify the technical requirements that are needed to solve that business requirement, go and find a solution and then work out how to integrate it with the rest of your digital systems. And obviously that is where WordPress and Open Source software can really come in because it's cheap, fast, highly scalable. So what I want to do is walk through the technical design of how we sort of implemented this new platform and I'll look at the kind of technical architecture, but also the business requirements that stood behind that and what the end result is. And then I'm going to talk a little bit about the operational design, so how do we actually work with the client to get this programme done? So this is what we think of as WordPress, right? You have a content management system and you publish a page and that's what most people use WordPress for. I'm going to add a bit of structure here because this is the way that bigger businesses think about their digital estate. So they think about it as systems of record or data, and this is stuff like finance and inventory research data, stuff that is core to the business and add strategic value and it's powered by big platforms, CRMs, ERPs, custom databases that you don't want to change out very often. Then at the far end of the scale you've got the experiences, so these are the web apps or the mobile apps or the web pages or the email or the social content, all of the different touch points that both people and machines can use to interact with your organisation. And those do change quickly, right? You can tell when a website is more than a few years old because it looks dated, it's not responsive, it's slow, etc. In the middle we've got this layer called architecture which is what we use to connect those two systems together with and these systems are things like WordPress and email management providers that change on maybe a sort of two, three, four year kind of cycle rather than every year every month or every 10 years. And even this basic implementation of WordPress is a content management system. Here's a front end was a significant advantage on what they had before. They had enterprise software and it was pretty awful even to correct a spelling mistake for many years they had to republish the entire site. So thousands of articles, that was an overnight job. So if you made a spelling mistake in an article as a journalist and bear in mind they're publishing tens of articles every day, if you made a spelling mistake it would be there for the world to see for 24 hours until the site was republished. So it's a classic example of enterprise software that's feature rich but actually falls down for simple day to day tasks. So we're already winning with WordPress which is what we do. So we're now going to add in WooCommerce to PowerSummyCommerce. Now again straightforward, we've got a content management system where we can add like a picture of the product, we can add some content about it, we can add a price and it displays out to the front end. And that's powerful because it means that like one set of training will serve both the people that need to publish articles and create products and they'll both look the same on the front end as well. So all pretty straightforward so far and then we go and do something different and a bit crazy. So we stick this platform called Blaze in front of WordPress and so all of the webpages are now served through Blaze. Blaze is, it's not a product that Pragmatic have built, it's a SaaS that is provided by a technical partner in this program and it does basically three things. One is that it caches all of WordPress's output. The second is that it has the ability to swap parts of the pages on the fly and the third is that it provides an API that has users details in. So what Blaze is all about is identity management and access control. So if you want to sell access to your content rather than providing it for free, you could use a system like Blaze to do that. So here just by putting Blaze in front of WordPress we can go from here's all the content on the web page to you must now register and or sign up, sign in to see the content. So that content is now restricted. Another cool thing that Blaze does is like you use Slack here in Greece, I think everyone uses Slack now. Slack has this neat kind of magic link feature where instead of having to remember your password you put your email in, it sends you a link, you click it and you're logged in to the app. And Blaze lets us do the same thing. So remember here we're not actually registering or signing in with WordPress at all. WordPress doesn't know anything about who's looking at these pages. All of that identity management is done by Blaze. So users can fit in their email, they can click, sign in, go to their inbox on that device, click the link and they'll be logged in. Which is actually quite powerful in terms of making sure that people aren't sharing user names and passwords. So something that might have happened previously is that a company would pay for one subscription and then they'd post the user name and password on the company internet and everyone could use it. Here the only way of accessing the content is to click a link that's gone to your email app on the same device. So you have to authorise per device and that makes it a lot harder for people to share. You're not going to share your email password. So there's lots of things that we can now do just by this simple change in architecture. Let's have a look at Google. So one of people's experiences with your brand is for a search engine. People go looking for content, they see a result from a website that you own and then they decide whether or not to click through. So if you work with digital publishing you might know that for paid or restricted content there's this idea of first click free. So if you want your content to be ranked by Google in its search engine you have to let somebody click through and see that first article for free and after that you can restrict it through registration walls or pay walls or do whatever you want. So that analysis of where has a user come from? Is it a search engine? Have they viewed other articles on this site within the last 30 days or not? That decision engine is powered by Blaise and it will either serve you the content or it will put a registration wall up. So if you remember at the beginning I said that PEI do four things. They do editorial content, they do research and analytics data, they do events and they do advertising. Research and analytics is really important. Who's invested and which fund? How full is it? What investments has that fund made? What's the performance of it? Who works at which firm? Which firms commonly collaborate on funds? What are their interests? All of this is really powerful information for their customers because it lets the market know what's happening. But their data lives in custom Microsoft SQL database in a totally different VPS. So how on earth can we expose that onto a WordPress site to give them this feeling of an integrated product? And the way that was done was by throwing all of the RNA database through Elasticsearch. So there's a process that just dumps the data into Elasticsearch every night. And we then built a React app that sits on the front of WordPress and ingests that data from Elasticsearch's API. So we've now got this website that has editorial content from the WordPress CMS. Also rich tabulated data from the RNA database as well. So we can do things, for example, like dropping a graph with live interactive data into the middle of, you know, here's a text block, here's an image, here's a video. So it starts to become a very slick and powerful platform for PES customers to find out what's going on because they can read a story and then they can explore some of the data that sits behind that. For PEI, this data is only available to higher tiered subscribers as well. Certainly not to free ones, not even to the kind of lower tier of paid. So it's really important that we can drive lots of exposure and awareness of this data to encourage people to subscribe at the highest tier. Cool. So this is what the platform looks like at the moment. WordPress and WooCommerce is still sitting right in the middle of everything but we've started to augment what WordPress could ever hope to do on its own without writing a heck of a lot of PHP that would probably run really slowly and be awful. So now we are going to add in CRM. So this is one of those systems of record and again this is a hosted service. It's hosted by a different infrastructure provider. The technology doesn't really... There's no kind of off the shelf integrations to pump that into WordPress. So we need to connect it through Blaze to handle when people buy a new account online the customer services team need to know that. Likewise, if somebody takes a call over the phone, buys a new subscription then they want to get the access rights on the website as soon as possible. And this plays out particularly with WooCommerce. So with WooCommerce the way that usually works is that you register with WooCommerce and then WordPress knows your details and when you go to a checkout page it's pre-filled some of these rows with the information that WordPress knows about you because WordPress and WooCommerce share that data. Now here because people aren't registering or logging in at all with WordPress WooCommerce is basically stateless. It doesn't know who you are. All it's doing is presenting a checkout page to an anonymous user. So here one of those components of Blaze which is the API comes into play and when somebody hits checkout and we know information about them we can pull that information from Blaze via the API and inject it into the WooCommerce checkout. So as far as WooCommerce is concerned this is a guest checking out but the only way that you can land there is if you're logged in through Blaze already and so you hit this kind of integrated seamless experience with the data coming from different points. And I'll illustrate how that works a little bit more closely in a minute. But basically what this gives the business is the ability for any subscriber that has bought subscription online or over the phone to get their entitlement straight away and for the customer services team to be able to support that customer straight away as well. So this saves a lot of time for both the sales team and for the customer services team. So analytics. Analytics are super important for business. They show them what the opportunities are for improving the return on the investment. Obviously doing all this stuff costs quite a lot of money. They want to see business value back and we do that by providing rich analytics that feed into business goals and key results. So in this case we use Google's Tag Manager product to feed data back to amongst other things Google Analytics and then back from there through to business intelligence platforms. So let's go back to that checkout page and figure out what the data is that we're going to be sending back. Okay, so on the left hand screenshot we've got pointer. Here we have colour coded the different components of the page and here is the Google Tag Manager data layer. Many people come across a data layer like a GTM data layer can I get a show of hands so I know. Cool, thank you. All right, so if you think about Google Tag Manager and a data layer is sort of Google Analytics on steroids it means that you can customise exactly what data you want to pass back to other analytics platforms and you can construct these kind of much richer data models. So in this case we've colour coded it. The WordPress part of the data layer matches the page so WordPress produces a page that hosts the WooCommerce checkout and it says things here like here's the Google Analytics account for this page, here is the URL, here's the name of the thing, the page, here's the time that it was accessed etc. Then we've got WooCommerce which provides a basket so in the basket are products that have got prices and different currencies for different, and the user is at a particular stage of the checkout journey and then in AMBA we've got the user's data which has come from blades so up here it says this is the user's ID, this is how many pages they've seen etc. So this whole combined data layer then gets extracted from these two or three different systems and presented back to the business intelligence platforms which lets us do some really interesting stuff. It gives us much more control about the data we can send back and therefore the models that we can do. There's some really interesting services around there for things like propensity modelling and big data for publishers. So if you have a bunch of subscribers it's really vital that you know roughly how many you're going to resubscribe without any extra effort because they're using the service every day and knows how many use the service for a couple of months and might need a sales call. It's about efficiency of the marketing and sales team and that leads to things like marketing automation and better business projections. So you'll be glad to know there's not much room for many more components on here but I'm going to talk about the email marketing part of it which is vital to their business model and you can see now that we've got this kind of very rich set of data flows around the architecture. At its simplest you can think of it as a kind of virtuous cycle so the job of the architecture is to expose enough of these bits of data at the front end to personalise and optimise the experience for that particular user at that particular time and encourage them to take an action which adds value back to them and the business. The data then from here flows back into these systems of record here and adds to the strategic value of the business. So around all these connections the aim is to make it go like that. Although it looks complex actually a platform like this gives us some real power so for a premium content business like PEI they have one click for free and then you can register and you can get three articles for free because they know who you are now and that's worth money to them. The next step is if you try and visit a fourth article and you hit like a paywall you now have to pay for a subscription because you're using their content enough. So it's critical that on a tactical level we encourage people to click through and read another article as often as possible. If we can encourage 20% more people to click on the fourth article every month we'll actually drive significant revenue. And so we end up looking at related articles which are kind of the typical way that people go through from one article to another and it becomes clear that it's worth investing some time and effort to make sure that these articles down here are not articles that users already looked at this month that they're going to be relevant and the architecture that we've built can do that. We know enough about the user from Blaze we know which articles they've looked at and what tags and categories they seem to be interested in and we have a content API again from Elasticsearch which means that we can mash those two things up and use a React app to kind of display the most specifically relevant next stories for that particular user. So we're giving that user the most interesting possible stories in the business of best possible charts of presenting users with a paywall. Cool, so that's the architecture. There's actually quite a little more that goes on as well and we've built very similar architectures with other clients as well so this isn't something that's sort of dreamt up this is kind of a proven model for how you use WordPress in the middle to bolt on a whole bunch of different components and interfaces and then model out the workflows that are going to add value to a business. How do we do this? How do we get from a conversation through to a year later, thousands of man hours a functional platform like that? So we did this with Agile, so a scrum team and in fact the other technical partners were working on scrums as well. Last time I gave this talk all the questions that I got about it there's nothing to do with the technical architecture it was all about how on earth did you manage to sell scrum into a client so I guess I'll talk a little bit about that now. The client knew that they wanted to use Agile technical methodologies to deliver this. It's like a large backlog, it's complex they know that they don't know all the details and so they were prepared to, within reason use high level estimates to work out how much time, how much money to do different things and evaluate their options and then run a nice scrum process to make progress. So we had to align very early on on what the end state of that vision was we had to do some discovery time to work out what all of the high level epics were and to do that estimating so that we could then start running as a scrum team. To start with it was a bit of a leap of faith to the client it always is with Agile because what you're promising is to make progress against these stories the best of your ability but as the client sees delivery coming in you build that trust and they give you a little bit more room to manoeuvre, a bit more trust, a bit more budget and you can carry on and ramp up so it's really important early on to deliver value back to the business and to do it in a visible way because we've communicated widely within the business as well so from that base of trust then we can ramp up to a good working speed and eventually everyone starts thinking about stories and points and which sprint this is going to get delivered and it's really quite powerful to see that digital transformation happening. Now of course to do that if you're coming from an organisation that's quite traditional and you've got managers and you've got hierarchies that don't necessarily align with what the business does and how this new integrated vision for the business needs to work there's a certain amount of coaching, nurturing, teaching, new skills so learning what it means to be a product owner or what an acceptance criteria is bringing in external expertise business consultants, scrum masters on both sides, we did this, the client did that you can't expect the same people to come into work the next day and do something totally different than they've done before you need to inject some training some know-how, some expertise into that really worth doing. In terms of working with stakeholders a typical project kind of works like this it always has when we were smaller you would have quite a few meetings at start you would agree what the project was and then you'd wave goodbye and maybe see them in a month or two months when you've built the thing and hope they like it this was very different this was weekly meetings and very intensive meetings every two weeks as well proper sprint planning, retrospective very open, honest conversations this isn't exactly what we intended but that's what the ticket says we interpreted it like this so together you kind of understand how not to make the same mistake with your communication or with assumptions next time and after a few sprints you really do clear that stuff out and it's resulted in fantastic delivery so what all of this has to be about is not writing code, not doing clever architecture it's about how do we deliver value back to the business so ideally they'll want to keep spending money with us because they can see that we can keep adding value back we can help them with that delivery but we can also help them understand what else is out there in the industry and how they might be able to profit from it as well so more and more I think the role of digital agencies is to be that interface between business over here which is getting on with being their business and the cutting edge of technology over here and trying to close that gap where it seems like a good fit and so that kind of ongoing inspiration, ideation creating a backlog that you can then work on it's really powerful synergy so what did all this end up with I wanted to talk through a programme status report which is maybe a kind of controversial thing to do typically not that exciting but really important to think about how when you're working with these kind of longer term programmes you're actually talking back to the people in the business and not just your immediate stakeholders but the people that they report to so how do we make this programme a kind of a success not just with the output but with the propaganda and the PR campaign internally and these status reports are one of the ways that we've done that so it's a really long page basically there's quite a lot in it broken down and I'm just going to like pull a few key sections out and say why I think they're important and interesting so at the top we've got like this high level roadmap so there are different work streams there are different projects happening within those work streams and those little red dots there triangles are milestones so these are the dates that we know that have a business impact they need to do lots of testing for a site to go live or they need to be prepared to do a content freeze or they've got an event coming up or something like that but it gives the business a really nice overview of what's been done and what's coming up next we've got this sort of overall programme status so these are high level questions at the top I've blurred it out because obviously it's kind of like a live programme but these questions say things like are we delivering to schedule are we delivering to cost is the quality good enough have we got enough type of resource so they kind of play in English open honest questions to which the business expects clear open honest answers back well it's kind of okay but blah blah blah yes we're doing okay and it's looking like it's going to be okay or actually no we're in a bit of trouble we need some more resource because we've got all these requirements coming in and we can't write them into stories quickly enough for the team to run through them then if you work scrum you'll recognise the one on the left is like a burn down chart so actually it's a bit of an embarrassing one to show the grey line is over a two week period that shows how many points you would burn so on the left you start with 40 points and every time you complete a story or a task you complete that complete that story so you burn the point that's the kind of the terminology behind it and so the flat bit in the middle is the weekend we don't work over the weekend unless there's a crisis so in an ideal way the red line which is the actual burn down would follow that grey line now obviously we've left things a little bit to the last minute on this sprint perhaps there's valid reasons for that but it kind of then it falls off a cliff but our aim always is that the red line and the grey line meet at the bottom here are some other charts as well these show things like quality of stories in the backlog and the progress through all the stories for a particular release and these charts and the ones before come straight out of JIRA so if you have used JIRA it's a super powerful piece of software and this page is generated in compliance so the two link together one's like a wiki one's like a software development tool they're really quite complex to get your head around if you used to like Basecamp or Asana or a piece of paper but when you invest in putting the data in and using tools properly the reporting that you can get back out is crazy powerful this one I really like as well I'll talk through basically these are this is like a financial model that sits behind everything so every time we agree to do a sprint like a two week chunk of work we agree which bits of work they're going to be so which stories are coming into that sprint and those stories all have like a value a nominal value which we call points so we know that if we've got four or five people working full time on a sprint for two weeks we should be able to burn 40 points and what these graphs do is they track what is the actual cost per sprint so we timesheet everything and at the end of the week if we've only done 20 points then it's going to look like the value per point is going way down if we've done 50 points the value per point is going up so this is a way for the business to make sure that they're getting delivery but we're also staying at a constant rate of how much that delivery costs and over weeks and months and years that becomes super important because it avoids conversations like I don't think you guys are working as hard on this as you used to or the client in turn trying to introduce bigger stories and pass them off as this is only one point when you know it's really a two so it's a really good check and balance and at the bottom here is like a high level risk register so what are the things which could still go wrong which could impact this program mean that we don't hit our delivery dates or costs or quality and just a quick bit of quick look at the results and then let's do some Q&A so we've now completed the migration actually it's been a pretty crazy year that says like we started running on this in January and we completed the migrations I think about a month ago and the business has kind of reformed around this whole idea of delivery so they now have new roles they have people who are subject matter experts they pass stories into backlogs they all vote on them and this idea of the business product this kind of the digital product kind of has a seat at the highest table in the business it's no longer like it's the editorial team and we've got a website it's like what is this product how does it impact our customers how is it adding value so it's a real realignment they now have faith in digital again having used some really awful software before they were pretty skeptical cynical about digital projects that's gone away and we've got through doing things like this passwordless authentication they've now got a much clearer and more honest view of who's actually using their service you know how many users in this organisation are using this username and login so yeah we've migrated all eight sites across there were ten couple of them being consolidated or ditched so eight sites and a platform build in a year feels like quite good work and the business results are pretty clear as well so total traffic's gone up nearly a quarter organic above that average and mobile traffic in particular has increased so in fact this site was on WordPress already so the difference is slightly less than on this site which was on this kind of legacy architecture and you know the overall traffic is much higher there I'm a bit over time so I'll just pick a couple of these out the transitions between different bits of a program are the most difficult and the most risky so how do you get from that conversation and the initial meetings to the first bit of paid work like a discovery thing that whole piece takes a heck of a lot of effort and care to make sure it happens and likewise once you've finished discovery transitioning from there to well we think this is what's involved and how much it's going to cost can we have the money to do it now please that's also a really critical period and it's a very time intensive and also quite emotionally intensive as well so once you've done that with a bigger program and you can kind of earn that trust and you reach momentum but building a momentum is really difficult and I'd really encourage you to invest personally in the relationships that you have with the stakeholders at those points spend time outside of meetings spend time outside of that that you have to spend with them often I mean one of the things I've always loved most about running a digital agency is meeting people right people often like oh clients this and clients that I won't have it impragmatic our clients are a the people that pay our wages but b someone the most clever inspiring people that you'll meet you know they've done well enough with their business that they can afford us to build them a website we should be very respectful of that if they don't know stuff about how digital works that's our fault not theirs so get to know these people they're running successful businesses and they're going to help you as well this kind of whole stack of tools that we was pretty painful to invest in I mean I spent a lot of kind of 2am mornings building backlogs in Jira and reassigning things into different releases and components but the reporting that we now get out of it is like super slick and the same with things like continuous integration so it's a bit of a faff to set up but once you get to the point where all your code has to be pull requested and reviewed and as soon as it gets accepted into a branch it triggers a whole bunch of build and test processes and ends up on a staging server for the client to review it saves so much time and hassle going forwards it makes you look like a much slicker where you are a much slicker operation yeah questions then I guess thank you what do you think is the role of WordPress in the architecture that you followed how big the role of WordPress actually could it have been done with any other CMS that's a great question Tak is it definitely going to be done with other CMSs so our initial involvement was responding to an RFP like a request for proposals yeah and we had to score WordPress against a whole bunch of general content management requirements I think I think we were the only open source solution there there were some other ones like proprietary ones I mean one could have done the same thing with Drupal probably probably with some other open source CMSs as well yeah parts of WordPress like the REST API or commerce as well yeah so not so much in this program although we've done quite a lot of I don't think we did it in this case I think we did it with sort of database level migration but we've definitely used the REST API for content migrations in different programs so scraping sites and injecting it through the REST API that's been really interesting so back in the architecture we saw this sort of split of WordPress here in the middle oh it's gone WordPress in the middle and then the front end website on the right hand side now if that was going to be kind of a true separation of concerns we wouldn't be using a WordPress theme we'd be using WordPress as a headless API and we'd build out a like a custom front end that was totally standalone so that the business can swap out the CMS for a different one and that's kind of the benefit and the idea behind this kind of architecture is WordPress might not always be the best content management system so from a kind of responsibility point of view we want to design an architecture that if the business is a requirement it's changed they can swap it out for a different one same with the front end same with identity management so a business needs to be able to swap some stuff out faster than other stuff for lots of different reasons and that's one of the things that becomes super difficult to do if you've got like a big monolithic piece of enterprise software you have to do all or nothing a lot of the time or it's very expensive bespoke development which then kind of takes away the value of using something which does everything in the first place okay hi, well I would like to ask a bit if possible to elaborate on why you needed agile for this project because it was it was a project that it's one off right, like I mean why did you need the different sprints for that after the first sprint you delivered the product that the company was using but it didn't have CRM for example or it didn't have email right so we used agile for a bunch of reasons one was this is being quite long project so we've done I think 25, 26 sprints this year and that gives us 25 or 26 opportunities to reprioritise review what's in the backlog and make the best possible decisions it's hard to know where the business is going to be in six months so alone what the most pressing requirements are going to be we knew there was quite a lot of uncertainty some of the components in here haven't been finalised in fact some still aren't finalised so it gave us the flexibility to head towards a vision without having to have to find every detail about it and obviously this is quite a big project and so the other alternative is to do kind of a waterfall type approach but to do that relies much more on knowing knowing more at the start here we knew that we were going to answer a bunch of questions along the way also it's worth noting that we could do agile on this because fundamentally we're building a software product the platform that underpins these brands is a product and they can onboard new brands and new websites a typical website project for you know most freelancers or small businesses like you know 5000 euros or something there's no point doing sprints for that you can take a lot of agile methodology into your approach but you know that is a sprint or it's two sprints so you don't get the benefit of the kind of continuous cycles of sprint and optimisation a longer program does is that kind of answer cool thanks hello and thanks very much for the great presentation congratulations and I had a question since you're integrating a couple of systems with WordPress and WooCommerce this should bring to you some sort of a high level of support you need to do when things are changing to the solutions that you're integrating one and second how if there's a process that you do to stay up to date with those changes in those systems so that your solution could be always up to date to the changes of WordPress and the other solutions as well yeah that's a good question so the first question was how do we support that when there are so many integrations right because the whole platform depends on multiple different pieces of software for multiple parties that's something we're still working on actually we've got at the moment we have multiple support desks from multiple providers and we pass tickets around if their business opens the ticket in the wrong service desk at some point we're going to want to put in 360 support service desks everyone just comes to us or another provider first and they follow a triage decision tree and farm the tickets out to the other suppliers one thing that we have done is have you ever seen status page it's a platform where you can put front and back end testing into one big page so that's a really useful way of diagnosing so here we can tell whether staging and live for all the different brands is up at lots of different levels so as the database up is the web server up is blaze up et cetera so that's the first one how do you support a stack like that and the second one is how do we stay up to date with all the changes and I think we go back to agile there as well which is that the business understands that this is open source software there are going to be updates so same question kind of applies to tech debt as well sometimes we'll decide to do something quick and dirty for the sake of hitting a business but we'll put it on a tech debt list and at a reasonable pace we'll ask to draw down some points from the tech debt and or maintenance and or any other technical improvements and so whether that's somebody doing some research us doing some testing they can all be pointed bits of work in our experience it's better to do little and often with that because it's much easier for the business to explain why 5 points out of 40 have to be tech debt reduction and maintenance points rather than just feature delivery than to say at the end of the year oh now here's a big bill to do tech debt and maintenance it doesn't work like that so being open and honest and making sure that the client understands from the start that running their own platform means making responsibility for those aspects of doing it as well yep one more hi David this is Chipriam from APTICOLS thank you for your presentation have you felt moments when you were micromanaging especially with those weekly meetings with the stakeholders and how can you get the stakeholders to a monthly meeting for example that's a good question it's about stakeholder involvement I firmly believe that the regularity of the meetings is key to the success and actually if the client wanted to go back to monthly meetings I'd start getting really nervous now I hasten to add this isn't necessarily me directly being in the meetings we have a team and the team are in multiple meetings with the client every week it's not micromanaging it's about communication so one of the most important parts of the agile manifesto is communication over documentation right so you have to keep talking you have to build that relationship and you have to create the space for conversations to happen because often everyone's very busy and unless you are sitting there with an agenda to run through this stuff it won't be deemed important until actually it's really critical so by no means should these be free these are part of running a successful project you charge for project management you charge for QA so you should charge for this meeting time as well so it's a valuable component of what you are delivering as an agency you're not if you're charging just to write lines of code that's a kind of a broken model if you're playing to offer strategy, advice, help plan as well as deliver the work then it's a far wider relationship Dori, okay cool hello, thank you for your presentation as I recall from your diagram the RNA data bucket was connected to the elastic search and the elastic search provides data to the WordPress to a react app that is embedded within WordPress to a react app that's embedded with WordPress yet so and please correct me if I'm wrong elastic search creates agnostic directories and lexicons so how did you manage to get the elastic search to concentrate on data and export data and values that can be used to show diagrams and data that it has a content value that people want to read so let me just check I think I'm answering the right question so the question is elastic search takes a data lake approach and you pull out what you want as I know from my knowledge elastic search is agnostic so the results, we don't know the results so how did you manage to filter out to prune the unnecessary data that's a pretty detailed technical question I was dying to ask you this sorry I'll be around all day so let's talk about that on a bit more detail thanks for coming everyone