 So my name is Jess Aindy Oreo. Thank you so much for coming. I head up product marketing at Aquia. And as we're sponsors of the show, we get to do this great presentation, but what we've learned from the audience is they never want to hear from us, and they just want to hear from people giving real world examples. So we've brought along a great customer today and the BBC Good Food team. I'll let you guys introduce yourself. Hi, I'm Duncan McKenzie. I'm the product manager for the BBC Good Food Top Gear as well. Hi, I'm Alan McKenzie. I'm the legal developer. I'm not related. One's MC, one's MAC. So I would lose my job if I didn't do just a couple of minutes on Aquia. Real quick, just to get a pulse of the audience, we typically have a split between those who are hands-on development and those who are more on the business side. Could you raise your hand if you're hands-on technical development? Okay, and then if you're more on the business side, business development, 50-50. All right, we will do our best to accommodate both audiences. And I'm gonna start with a really brief update of Aquia products, and then we're gonna turn it over to the BBC Good Food team to tell their story about the evolution of BBC Good Food digital experience. So how many people in this room are confused by what Aquia does? What we sell, what our products are? One brief, so I would guess pretty much everybody in this, I'm confused sometimes. It's my job to make it simple. So most people have a hard time understanding when we have different offerings, how it all comes together and how we help customers in a unified way. So we are releasing the Aquia platform, which is really just a unified way to learn about what we do in harmony with Drupal and the products and services we do that surround successful Drupal implementations. Agility is a great thing to talk about when we're talking about the position of Drupal and how that's gonna help organizations move faster in the digital age when digital disruption is making everything much more competitive and you have to have updates, social integration, et cetera. And you're gonna hear about how BBC Good Food has worked with the mobile access challenge. So agility is necessary. Resiliency is what we talk about when we talk about Aquia Cloud. Are there any Aquia Cloud users in the audience? Okay, great. When we talk about resiliency, people typically think of it as something that's just about the hardware side or the infrastructure side. They think of it as uptime, but we think about it a little bit differently. We provide a holistic set of tools that allow you to make sure the application quality is where it needs to be before you launch the site. So ideally the infrastructure is kind of an afterthought and it's all about getting the application ready to scale. So we have a host of tools that help ensure application quality, like Aquia Insight, which does a lot of testing on the application proactively, helps you understand what needs to be improved across performance, security, and general Drupal best practices. So for more information on this type of thing, it's not planned in the presentation today, but I'm happy to talk with anyone after, or you can go to the Aquia booth and we have amazing engineers and product team members ready to do great demos for you. The Aquia platform is about agility with Drupal, resiliency with Aquia Cloud, and then connectedness, which is both a Drupal value proposition because it's about the ease of integration with whatever technologies you've already implemented in your portfolio. And then it's also about how Aquia can help you get that next mile when you have a unique implementation or integration you need to work through with Drupal, and our team can help you through that. So we talked about the resilient cloud, it's scalable, reliable, and secure, and you're not gonna be able to see all of these because of the technical issues, but this is sort of a quick update. Was anyone at Drupalcon Prague? Okay, so we really just wanted to do a recap of what's new since then, and there's a tremendous amount of innovation that's going into the Aquia platform at any point in time, but a few really exciting releases are Aquia Insight I talked about as this testing tool. We released both an API and the ability to create your own custom tests, so it's not just us projecting tests upon your application that we think you should care about. You can dismiss them if they're not important and you can add your own tests to it. So that way the application becomes custom to your unique needs. So we think that's been really powerful for developers in particular. Teams and permissions is all about making sure the right people have access to do the right things within Aquia, and that was really critical from our partner community that we create that, and I think it's helpful for general development teams as well. Second version of Dev Desktop is as a release candidate right now, just as a couple of highlights. And then we came up with a bunch of platform monitoring tools, platform health and other ways to interrogate performance issues, whether it's application or, and we have our own that we've created for our own infrastructure to spot check there. The digital experience solutions, this is really just how we position Drupal. And we talk about modules, we talk about distributions, and we talk about many ways to get started. You can start from almost at the finished line with distributions, and this is really appealing to companies who haven't really gone all in on Drupal yet and they don't understand it and they need something that seems a little easier, seems a little customized for their use case. So we position that as digital experience solutions. And we have recently created a new commerce strategy, and this is really about bringing more visibility with all the marketing power that we can to Drupal being great for transactional sites. Some use cases that Drupal commerce is perfect for, and some use cases where there's a significant investment in the backend commerce platform. We've created a new use case we call content for commerce. So use Drupal for that better content experience which drives better engagement, give marketing more control, and keep your e-commerce platform as the plumbing and what makes the transactions go in the post sale support. So in terms of technical innovation for that, we're actually working through our connector strategy and we started with a company called Demandware which is a well-known commerce platform. And last thing I'm gonna give you an update on is the most exciting thing that we've done this year is launch Aquialift. Has anybody heard of Aquialift? Great. Well, should we have time towards the end of the presentation, Steve Smith, our VP of engineering for Aquialift is gonna give a quick demo, but it is a personalization solution that's tailor-made for Drupal. Many organizations struggle with what's called shadow CMS when they go to test. They have to create duplicate content in another CMS, test it and when they get the results then they come back to implement. But this is all embedded within Drupal. They've created great technical innovation that they've contributed back to the community and then we have our own service that we also sell. But it's embedded in your workflow. You don't have a redundant content system you have to go to. It drives relevance. We acquired a company called TrueCentric this last year which is also very exciting and that's how we acquired Steve and his excellent talent and they bring offers and recommendations to the table and this little logo you see here, smart, that is to be continued will be launching another product in November. So that's really about real-time data management and bringing in different profiles of known and unknown users. We're very excited about the digital engagement services. We feel like one of the things that we owe the Drupal community is to help enterprises think of Drupal as an enterprise solution. So we have to close gaps in point solutions that our competitors like Adobe and Sitecore offer they've acquired these technology areas. So we're doing everything we can to invest in making Drupal seem much more appropriate for the enterprise and providing the right level of integration with existing marketing systems and really trying to give Drupal a one-up with embedded solutions that help drive better engagement and at the end of the day impact the bottom line for whatever your business outcome is. So we launched AQUIA Lift in February. We were able to launch offers and recommendations in July which is when we acquired TrueCentric and then we're working towards integrated view of the customer which was what I was just teeing up there. So I would love to talk to anybody who is interested in learning more. Again, we'll do a quick demo of the personalization solution if we have time towards the end but unless there are any questions right now at this point I'm gonna turn it over to the BBC. All right, guys. Thank you. Thanks to everyone for coming. There's a huge number of people here which is slightly terrifying but also very exciting at the same time. Oh, hello, is that better? Great, yeah. Okay, I can do that. Multitask, click and speak in the microphone. Thanks. So we're gonna talk about BBC Good Food today. I just wanted to get a show of hands. Who'd heard of BBC Good Food apart from the people who worked on it before they saw it in the programme? Yeah, changing show of hands, okay. And how many people are from the UK? Yeah, it matches up. Okay, interesting. So as you can probably tell from that straw poll it's quite a UK-focused website but the brand itself is this sort of huge entity which actually turned 25 years old this month, came out of a big magazine business and just celebrating our birthday this month, as I said. And it's quite large scoping. It's not just about this website. We've got books, some of the most popular cooking books in the UK are from BBC Good Food. We've got print magazines from all over the world and we're hoping our digital business will eventually turn the same way. But just to give you a sense of what's really core to the product we have over 7,000 recipes that are triple tested on the website and that's really at the heart of it. So a couple of stats for people who don't know about BBC Good Food. Now I'm gonna hope this works. Amazing. We get about 11 million users a month and the majority are from the UK and we're growing about 50% year on year. We also get huge spikes on events such as Christmas Eve where last Christmas we got just over one million users coming to our site. And what percentage, and this is, I'm asking you loads of questions and it seems really harsh. What percentage of sessions do you think comes from two device groups? So say iPhone and iPad, how many sessions do you think have all our sessions come from those two devices? Anyone, shout out. 40%? 80? 40%. It's 44% just from those two devices and we have to pay so much attention just to those devices in terms of, there's an interesting talk earlier about the next steps for UX in the world and everyone sort of ignoring just a couple of devices and thinking about the broader devices. I think it does come back to business value and going, these are the two biggest devices. We should probably focus most of our effort on those. So yeah, two big devices. We also just rebranded literally last Thursday. I was up till, I can't remember what hour to make sure this went out on the dots in line with the print magazine, the book and everything else. Problem with not just the digital brands. So, what we're gonna talk about today. We're gonna cover really the good food rewrites and where we've come to now. We're gonna talk about the challenges and why we did the project, the approaches we took and the results that really came out of that. So first of all, the challenges. So there were really three core ones. One was the changing user behaviors, so the changing in the habits. The second was really internal. It was around the fact we didn't really have a proper build, measure, learn loop. That's not really an approach we had before. And the third was really around the editorial team is so core to our work. Getting content live was really difficult on the old system. And the stability wasn't great of the old stack. So first of all, the changing habits. Now the original website that was built, I think seven, eight years ago, was really focused around one core user journey. A user in this environment went to their desktop computer. They printed a recipe and then they went and cooked it unless they had a kitchen like this, which is probably quite rare. But we noticed there was quite a change in user behavior in the last few years, which I'm sure you'll notice across the board, but it really affected us quite a lot. The fact that people were starting to cook with tablets in their kitchen. And they were taking their mobile phones into the supermarkets with say an ingredients list in their hands or in their notes and searching for recipes even in store, looking for inspiration. We saw this in our numbers as well. This was before we relaunched. We were seeing growth, but actually we're seeing even bigger growth in mobile and tablet quite dramatically. The second challenge for us was the lack of the build, measure, learn loop. Does everyone know what that means? I mean, it's probably quite self-explanatory, but just the fact we weren't really putting things out there testing and then iterating on that, which is kind of standard practice now, but back then we weren't really doing it. So the problem with our loop was the fact that the build took too long. The multiple systems were too complicated. We were worried when we rolled something out that it was going to break the sites. We didn't really have a great analytics solution. We had GA, but it wasn't really well set up. We weren't really tracking everything that actually was business value for the product, or a huge amount of expertise around that tool. And there was just not really a culture of learning and building on that. We really focused on getting as many features out the door, which I'm sure the case for a lot of people at the time, rather than actually delivering any value. And our third thing, as I mentioned before, was about time to publish and stability. So just a bit of background, how our editorial team work. There's actually a kitchen in West London where we have a team of people who cook these recipes for us before we put them online. The great thing about that is they cook them three times and we get to eat all of them. I think it's quite an important perk to my job. Sort of up there on the list of benefits, eat food, money, no, eat food, that's fine. And every month, the editorial team upload about 70 new recipes, as well as a multitude of other content, videos, guides, how to eat healthy this month. Now the problems we had, where there were multiple systems, at least two or three, just to create a recipe, the systems really didn't talk to each other very well. That's probably a much better technical way of describing that, but I think lack of discussion is the best way. And it took between 20 to 30 minutes just to do one recipe, which is about three working days a month just uploading recipes that have already been cooked, already been published in a magazine, and we're waiting three days to get all that content live. Now we could have said this to the editorial team. I don't think they'd have been that happy about that. They'd have probably ended up looking like this. Another issue for us was around stability. So we have big peak traffic spikes at lunchtime. They're really obvious times. Lunchtime, you're hungry, you wanna look at recipes and see what to cook. And then just before you go home, you're really starving, and you've no idea what you're gonna eat tonight, and you wanna look at some recipes. So when you pop into Tesco's or Asda or any other good retailer, you've got something, you've got some inspiration, you know what to cook. We're also really driven, and that's again probably obvious, by massive peak events, pancake day, Christmas, as I showed before. Valentine's Day, on which the most popular recipe is onion rings. What's that? Valentine's Day. Oh yes, it is, it is as well. Great British, is that what it's called over here? It's a very important event for us as well. We do a lot of content around there. Now our problem in, I think, a descriptive human way was probably like this. Yeah, these hard-working Japanese train staff push literally cramming people onto the train, and that's really what we were doing. But the problem for us is we didn't get a full train, we tend to get a page like this, which wasn't really ideal for our users, and it isn't really what we're looking for. That probably happened twice a week when it was really bad. So the challenge we face was the user, their behaviors were changing, they always are. We didn't really have a build, measure, learn loop, and we weren't really quick at getting content live or having a stable site. So our approach to this. For the changing user behavior habits, something that sounds really obvious now, but we looked at responsive design. It was the first time BeapSea Worldwide really looked at it when we approached good food. For the build, measure, learn loop, the key for us was tools, process changes, and really mindset of the whole team to change. And for speed to publish and stability, I couldn't really go to DrupalCon and not say this. Drupal was key for us, and also a cloud infrastructure, which Acquira will like me saying, because it's theirs. So responsive design. In response to the changing user habits, we really focused around three core platforms. We looked at mobile, we looked at tablet, we looked at desktop, and we focused these around lifestyle choices. So a user in a supermarket with their mobile, a user in a kitchen with a tablet, and someone at work at their desk looking at recipes at lunchtime. So this is what we did for mobile. We really focused at minimizing the amount of content at the top and just having a lovely recipe image, which is what users care about the most we've discovered. For tablets, we made it a bit more of a browsing experience, you know, swipe across content, really enjoy the depth of it and the big images. And then for desktop, we still really had to have a great desktop experience. We've got quite an older audience than probably a lot of sites that the rest of the audience here work on. And it's core, it's key to us, is having a really good desktop website. So process and mindset changes. There were really three key things we did here. One was user testing, again, really obvious now, but we'd never done it before. The second was around changing the team dynamic and the ways that work. And Alan McKenzie here will talk for hours on length about how to best have the greatest team dynamic to get the best work done. And the final thing was having the right tools to be able to do the job. So we started to do things like this. We've actually got a user testing lab, we're really fortunate, I know not everyone has this, having a user testing lab in our building, literally a sort of a dual room setup where we can film people, while we watch them interact with our product sounds quite creepy at first, but when you get a stakeholder in a room watching it, it's amazing how many people you can convince just by someone clicking through on a button. We didn't think they'd click. We also did beta tests. We put a beta test live to a great group of people who are actually really real good food users to test our product. The second thing was around, as I said, ways of working that we changed. We changed our team setup so that the designer, the UX lead and the lead developer were actually involved in our whiteboarding process. It wasn't just let the designer and the product person go off, design something crazy that's impossible to implement and then come and present it to a developer who will then probably pull the hair out and go, this isn't possible. So we really did change that and just make it more dialogue-driven. And it's amazing, actually, when you bring that community of people together, how great it is, you get much better products out of it. And a key tool for us was around impact mapping. It's so important to understand, no matter what role you're in, why you're doing something and I think you'd be eternally frustrating if you don't know that. And it speeds up development so much and I've talked to Alan a lot about this. When he's building something, he knows why he's doing something. He'll take the logical approach based on that. We won't have to go back and forth 50 times discussing every minutiae approach. He'll know why he's doing something so he'll know the best approach. I think most importantly, it's not about software, it's about the business value that you deliver. And the final thing that we started to use and used on that approach was multivariant testing. Again, it was something quite new to me and it's an amazingly powerful tool. This is a really simple and quite boring example, so apologies, but we tested four different magazine subscription trials on our website. It's a really important thing to our business. It's still key to our revenue. Out of those four on the, it is the left for you, the four pink ones, who thinks the five for five pound won the test that we ran? Who think they save up to 40% on the test? We're looking for delivering business value. So the most conversions to subscriptions. Who think save 25%? Interesting. And who think save up to 74%? Interesting. So one thing if you have quite limited analytics on these sort of tests is you can get completely no results. So the click through rate on all of those was exactly the same. The important thing that we discovered was actually when they got to the subscription sites, the most value that was delivered was by that five for five pound offer. Like that out of anything delivered the most value. I think that's one thing that I've learned that's really important. It's not just about click through rate on a page. It's actually what's the end goal you're trying to get to. And I know that actually getting the analytics to do that is quite a pain a lot of the time, but doing the work to have these tests is only really worthwhile if you've got a true result out there. And the final thing in our approach was around Drupal and cloud infrastructure. It's not something I've really heard of sort of two and a half years ago, I'll be honest, Drupal. And I've learned a lot. And this is my first Drupal con. I'm sure there's a badge I should be wearing saying this is my first Drupal con. And Alan's probably best to talk about this, but this is what is really changed from our sort of culture internally point of view. We started to say things like someone's already built that, make it Drupal or Drupalize the problem. Which again, two years ago, I would have no idea what you're talking about and those second two points. But the benefits I've learned from that are we don't build a motor around our systems. They're not necessarily unique to us. A lot of people have built the same thing before. It's really quick to onboard people. You can pick up anyone and drop them in a team. And other than the team dynamic, actually understanding the system is quite quick. And it's less risky in terms of rolling out stuff. And it's a lot quicker to get to market. I've got an example later that I'll show you on that. And then we moved, and this is probably how you should celebrate cloud infrastructure from an internal stack to the clouds. Alan's got this quote that there's no secret source to our cloud infrastructure. It's just a standard Acquia cloud infrastructure. But the great thing about it that I've discovered is just the scalability of it. We don't have to have all these servers running the whole time. When we hit 1.2 million users at Christmas, we can scale up for that. We don't have to have that service always around. Do you want to speak any more on that, Alan? Better than that, cool. Silent as always. Okay, so then on to our results. So these are the things we were trying to change. We were trying to meet the change news behavior. We were trying to establish a build measure learn loop. And we were really trying to reduce editorial workflow and increase stability. So for the change in user behavior, responsive design was at the heart of that change for us. So prior to the relaunch, and apologies if you can't see this graph from the back, we had desktop, which was greater than mobile and tablets. You can see sort of the divide and it's pretty consistent. There was a sort of a curve coming, you can see towards the end of that graph, but it was pretty consistent throughout the year. Literally the week after we relaunched the responsive site, this is what happened. They literally flipped. So users were waiting for a responsive product or a product that worked well for them on their mobile and tablet rather than being responsive. They just wanted something that worked really good for them. And they started using those devices immediately as soon as we offered that feature for them. And that has only really continued to grow since then. So second really establishing this build measure learn loop that I keep hopping on about. And I think the best thing to do is just give a really quick example of one thing we've done recently that has benefited from this. So we wanted to add a new social sharing bar to the site. When we did the relaunch, we quickly threw one up as you do in the last days before you go live. You go, oh, we haven't put that on, have we? So you quickly throw something up. And this was the third iteration we had. It was from a third party. There was three core options. There was an expandable button at the end which had, when you expanded about 500 options for sharing, which I think is probably not necessary to most people in the world. And that's what the first iteration looked like. We whiteboarded, we user tested a load of different new designs across mobile tablets and desktop, and this is what we came up with. We added Pinterest, which sounds really obvious, but until we looked at the data we were like, of course it's a food website with lovely food images. Of course everyone uses Pinterest and shares good food on it, but we were prioritizing Facebook and Twitter like the rest of the world when we first did the first iteration. And also we put email up there, which is surprisingly important. The amount of people that email a recipe to their mobile phones so they can take it to the store is surprising. And the reason this happened was because of the dialogue between all of us, we all sort of inputted into this design and the output was a 500% increase in sharing overnight. And that's really continued. It wasn't sort of a one-off peak, it's just continuing to grow. And our final thing that we wanted to change was reduce the time to input the editors and really increase the stability of the site. So very simply, it took about 30 minutes before to put a recipe in and you ended up with quite a frustrated editor by the end of it. It now only takes 15 minutes for the biggest recipe, that is, sort of. So we've reduced that by half, which I think is great. So the editor can spend a lot more time just writing content and devising great strategy rather than focused on systems. And the stability of the site, I'm sorry I don't have any great numbers around up times 99.4% up. It basically doesn't go down like it used to anymore is what I've learned. We're not getting as many users emailing us complaining it's down, we're not seeing it go down when we hit a peak traffic at 4 p.m. on a Thursday and we all panic going, what do we do? It just doesn't happen anymore. And we also manage massive peaks so much better, so Christmas Eve last year, 1.2 million. I did traffic projections as anyone does, particularly when you're working with a cloud infrastructure, you're trying to forecast how many servers you're going to need and how much you need to pay for Christmas. And we just didn't know really, we sort of estimated but it was much bigger than we even expected. And that's what we've ended up with, half the three writes. A great responsive design product that just continues to grow. So thank you. Thank you guys. I'm gonna invite Steve up to get ready to see if we wanna do the demo, but at this point, we would love to take questions. Does anybody have any questions? So at the moment, it's a lot lower than we'd hoped in terms of it's not something we've really focused on, is what I'd say. When we rewrote the site, we had a binder functionality which basically allows users to save recipes to their binder. We haven't particularly expanded on that since the relaunch and that's something for our strategy. So the percentage is pretty small at the moment but we're hoping that will grow. I will pass this to Alan who will answer this far better than me. The gentleman is interested in whether or not we use a CDN. The answer is yes, we use Akamai and we cache pages as well as static assets. So the question is where do we get our traffic from? And I'll let Duncan answer that one. So the majority is organic search, a massive percentage is. So we're quite reliant on Google is how I'd say. And that was a big thing which I didn't really mention for the rewrite was reformatting all our metadata and actually surfacing a lot more of it. Any more questions? One at the back. It's a very good question. It's not a decision I made personally, it was sort of a decision we made as a group, a technology group overall. There was probably quite a lot of thought that went into that. I imagine there were some of the benefits but I'll probably let Alan explain probably why from a technical point of view. So cost, it's cheap. Time to market is excellent. It suits media companies extremely well because Drupal does content commerce and community very well. The IP is owned by the community rather than by a hostile company. So we do have to worry about I was getting price gauges in the future if we went with proprietary CMS. Cost, exactly yes, cost and flexibility. Any more for any more? So was that how's the maintenance going? Just so I could just repeat that. Since the relaunch, so that was July last year. Yeah, I will pass to Alan. So there's been occasional pieces of tuning various SQL queries but there's been continuous product development. There's not really been that much maintenance. It's been certainly an awful lot less than the old system. Yeah, we update core and modules when there's necessary security updates. There you go. The gentleman in green. So we continue to use Google Analytics. I would say one thing we started to do which we weren't doing say two years ago was as we were rolling out features, one we were thinking about what are the KPIs of these and at the same time we were rolling out the functionality to track that if it was GA events or something a bit smarter than that. But yeah, it's just again, it wasn't something that was thought about. We'd put things out and go, oh, how do we measure this? And then we'd implement the measurement afterwards and it was just bringing that into the process overall. In terms of performance monitoring, Alan will probably speak better than that but we're looking to expand on what we monitor in terms of app monitoring as well as just more traditional monitoring. Does that sort of answer your question? Correct. Any more? Person in the orange t-shirt. In terms of, there's not that. So the question was, well, it was more of an observation, I think. There's not that many fancy parts on the websites in terms of moving things. Is that fair enough? So I would say that actually probably came out the UX rather than us going we can't do that technically. Because of our user base, again, it needs to be simple and it needs to work. The content sort of at the heart of it. We've got amazing image areas for our recipes. We've got great recipes. And actually if we put a fancy UX over it that made it difficult for you to find that, it would be actually more challenging for our particular user base. I don't know if you want to say anything about gloss versus Drupal or... Drupal can certainly do gloss but we opted not to because our demo graphs, they just really don't like it. There was originally had a carousel which moved what I felt was fairly slowly but it's still too fast for some of our user base. We had to tune it down. The lady in red at the back? I'll let Duncan answer the community question first. So I sort of hinted it to her earlier but we've got a function called my good food which sounds very self-explanatory which at the minute is quite limited. It allows users to save recipes to their binder in effect so they're sort of their storage area. They can create collections of content as well but they can also comment on the site. One thing we're looking to do particularly on the roadmap for the next two years is really to drive that functionality up so offering gamification opportunities where we've got some really engaged users who are spending up to I think it's 10 to 12 minutes some of our users come and spend on the site and we're not really offering a huge amount to sort of benefit them or reward them and that's sort of the next step for me is to grow that and hopefully grow the community because of that. Do you want to talk about responsive? So we show all our content, we don't hide any of it. It just gets pushed further down the page. We keep it really simple. Quick question. When you said you took the time from 30 minutes down to 15 does that include uploading all their sorts of multimedia if you have some video attached to the recipes as well or is it just the publication of the written content? So it's images, text, everything that goes with that recipe, yeah. I'd better take the one just behind you because you went first with a great great jumper. That's a very good question. I wish we had real-time demographic information it would be a blessing and I know GA have started to introduce demo information which I think is still based on survey or usage of other sites. So it's still a bit finger in the air I think true demographics against your user base. Yes, we have done surveys across the whole brand. We did a huge one recently which we're actually talking about on the site at the moment but I don't think our demographic hugely changed. I think that audience was waiting for a responsive product to work well on their mobile and tablets. They're not the sort of people that would put up with something that doesn't work. They just go and use the device that it works for them on. We really have to bring it to them. That's what I said we learned. But to answer your question, I don't think our demographic has dramatically changed now. Sorry, in what sense do you mean? Do you want to speak to that, Adam? You said we did some slight customization. What you have to know is that the recipe form is actually quite complicated. So we split our ingredients into multiple different parts so we can feed that information to other systems. For instance, automatically fill a shopping cart on a Target supermarket website with all the ingredients in a recipe. So there was necessary complexity and we can't really do much about that. Now watch it. I would say there's actually been a change we made quite recently with good food where the kitchen that I described earlier is now a content hub for every platform rather than, in the past it was probably more focused on print mainly because we had recipes coming out in the magazine first. That's definitely starting to change that whole culture. I think one opportunity the web gives us that print doesn't is when there are trending themes that come up, Bake Off is something you can plan for, but when courgettes like two months ago suddenly became amazingly popular because there was a massive over-delivery of courgettes in the UK, supermarkets were flooded with them. Interesting fact, I know about food. That started trending on our site and that's not something you can react to in a mag that needs to be printed six weeks before it goes on shelves. So it just allows us to be a lot more reactive. So I say that's the sort of content we can really deliver on web. Yes, I agree. And something we're looking at in the future is potentially using Drupal as sort of the hub for all content publishing. That's not a place we're at yet, but it's certainly something, I guess you've got to get, the problem is you're working with a print team and you're used to working in Photoshop or in design and these sort of things and that's where they're putting the content in, but I agree. It's something to change for the future. And yes. Yeah, so we use Acre Solar Search with Search API. It's quite a large index with just over just under a million documents, I think. Yes, and then the orange T-shirt again. We don't, yeah, it's a difficult, we don't currently dynamically feed it from Drupal. Again, it's something we'd like to do in the future. It's building that infrastructure and rebuilding those apps in that manner. And it's certainly something we've talked about doing. One thing that's important for me with apps is working out what their product proposition is separate to a responsive site. It's a really tricky one and I think many people, none of our competitors have really cracked it either. Why do you have a native app? And for me, there's one core reason. It's a completely different marketing opportunity. You've got the power of Apple and Google, which is great. But you still need to have a reasoning why. Why is this product different to the one on the web? Why wouldn't someone just save it to their home screen on an iPhone? But there are different audience spaces and some audience spaces go into the app store and they live in the app store. They don't live in browser like a lot of us do. So yeah, I think it's just, to answer your question, we don't currently power them, but yes, we still are looking at separate product strategies for them. And in terms of the size, the good food, the main good food app we have is just over two million downloads. I'll ask one more question. So one of the coolest pieces of your experience from my perspective is that you can select a recipe and get that grocery list fired off to the retailer of your choice. How did you establish that? What do your partnerships look like? Do you get any benefit from that financially by grocery? Yeah, so I can talk to that probably not in a huge amount of detail, but to a certain degree. So yes, we work with a partner. In fact, we have a program at BBC Worldwide not to plug another program of BBC Worldwide called BBC Labs, BBC Worldwide Labs, a company in the first iteration of BBC Worldwide Labs called Foodity who are now changing to a name I think called Constant Commerce. We literally sort of sat sort of two desks from them. I went and chatted to John Agnes who's their chief exec now about what their product offering was. And what they offered was the ability for you to buy ingredients from a publishing website through a retailer. And it was kind of a no-brainer partnership. Initially, we just I think started with Astro Tescos and we're starting to expand that to more and more partners now. The functionality at the moment is limited to just being able to shop one recipe at a time which sounds really basic, but it was the first step to test if anyone actually even used it. And yes, we started to establish retailer relationships so we do benefit from that. But our future ambition that I probably hinted at is not just to be a recipe website, we want to be a place that people can come and shop for their weekly shop. We want to be able to offer you editorialized plans so if you want your four meals for the week plan, those four nights you're gonna be in, we want our editors and the system to be able to tell you what to eat. You click one, click on the button and then Tesco to deliver it to you. That's kind of the ambition for us. Great, we are gonna move to a very brief demo. Thank you guys so much. This was a wonderful presentation. I hope everybody enjoyed it. So I'll let Steve introduce himself but he's gonna demo the solution that was behind some of the testing that BBC Good Food has been working on. Hi everyone, I'm Steve Smith. I'm responsible for the Lyft product line at Acquia. And what I'm gonna show you today is a very quick demo of testing and targeting and native inside Drupal within the Drupal publishing experience. I'll show you how both of those things work. So this is what I've got up here is just one of the stock demo framework demos. This is the Nexus travel demo. And Lyft has been installed on it. So what that means is as people are engaging with content in real time, a profile is being built about that visitor if they're anonymous or identified and being stored in real time in the cloud and we build a high resolution profile about their interests and their behavior. So first I'm gonna show you how as I click on a few pieces of content here the experience is gonna personalize. I'm then gonna show you what the actual profile looks like and the segments that were defined and then how you target content within Drupal. And then I'll also show you a quick AB testing example. So I'm on this website here and you'll see that there's a bunch of a bunch of travel related content. If I start to look at things like Island Life and Puerto Rico. And maybe I start to look at trip packages here. So I'm looking for some kind of a tropical vacation. So maybe I look at this particular package that's available. So what's happening now is this content is tagged with certain attributes that suggests that I'm interested in travel content. That the kind of persona that I would fall into. And if I go back to the home page magic's supposed to happen, which it didn't might've been because I was sitting here for a second. Let me look at a few more pieces of tropical content here. So to get a look at San Juan tourism and let me go back and look at this Puerto Rico vacation again. All right, there we go. So what has happened is that profile's been built in real time and I've fallen into certain segments. And now this featured package is tropical related. This hero image is tropical related. So let me show you how that actually works in the backend, in the cloud. What does that profile look like and what were the segments that made this magic happen? So let me just switch over here to the lift web interface. The lift web interface and it timed me out here. All right, so these are all anonymous profiles that have been built about this user. This one here should be me. And so you'll see it's assigned a tracking ID, just a token and that's stored in a cookie. So if the user returns, even if they're anonymous it'll continue building that profile where they left off. They've been assigned to a getaway type persona based on the content that they were looking at. And then there's a whole history of that user's interests and behavior that's being built out here. So every visit, the geography associated with the visit, the marketing source, the kind of device that they're using, all of that information is being assembled in real time. And then all of the things that they actually did on the website, so every piece of content that they looked at, attributes about that content, any other kinds of events, these could be e-commerce events like Add to Cart, these could be shares, likes, any kind of event that you wish to instrument can be captured in this profile. And it gets aggregated so that you can pull out implicit signals that matter. So in this case, this particular user, the top keywords that they were interested in, spa, recreation, tropical, the top personas, getaway. This clearly was somebody else or maybe me in a previous session. But you'll see how you can pull out those signals that you care about, that your users are giving you through what they're doing on your website. As well, if they identify themselves, you can tie them together across their different devices and unify their profile. So seeing the percentage breakdown across the different platforms that they're using within that unified user profile. All right, so then what you can do is you can build segments. And these segments can then be shared with Drupal and you can define content variations that will reach users depending on whatever criteria you care about. The persona they fall into, their favorite keywords, their geography, any combination of things. So in this case, you'll see this adventure persona as a segment that's defined. And if a person has a persona of adventure or a persona of outdoors, and in this case, this was just based on taxonomy assignments to the content, then that user will fall into this segment. And then on the Drupal side, let me show you how it's set up. So within Drupal, here's all the different campaigns. So these are different kinds of targeting campaigns as well as testing campaigns. So if we look at this one called homepage banner, we can go in here. Here's all the different variations that are set for that user. And if I drill into that campaign, you can actually see the segments that have been used that were pulled from the cloud from the Lyft web profile. So here you'll see Aquila Lyft profile segment persona adventure for this particular variation or interested in tropical or persona getaway for this variation. So this is a way within Drupal, within the native Drupal authoring experience, you can use predefined content in combination with these real time segments to hit different people with different characteristics in your audience set with different experiences. So I'll show you very quickly. I've got one minute left here, what testing looks like. So if I want to do a test, it's just as simple as saying I want to create a new campaign and actually let me just make sure I've got the site loaded up here in the background. So let's say I want to do a new campaign. I want to do an AB test. I'm going to give it a creative name like example test. And then I can select any element on the page that I'm interested in creating a variation against. So I can click on bike the USA. I can say I want to change that to bike the USA right now. And now I've just defined a variation. I can then set up goals, which can be any sort of page or click or some kind of thing that the visitor has done on the website and start this test up within seconds and start collecting data and understand which one's my winner. Can you clarify quickly, you can use that microphone, what has been contributed back to the community so people can maybe get to know a little bit about what we've done there? Sure, is this turn up? All of the personalization modules have been given back to the community. And in fact, the Lyft web backend is only one option. You could use a different testing and targeting back end if you so chose. Or you can even use other data that you may have about a visitor and pumping in and use those modules. Any other questions? Yeah, it works with Varnish. It's all done through JavaScript. So in combination with the Varnish cache, there's JavaScript that has all the variations available. And that's kind of created dynamically on a page by page basis. And then the JavaScript selects which variations to go with. Yeah, there's a first party cookie that's stored. So if the visitor comes back, it continues to build a profile from exactly where they left off. We can also optionally use a third party cookie if you have multiple websites that you operate and you want to tie that visitor together across those sites. Notwithstanding any of your privacy policies and other considerations. There's a JavaScript API where you can actually define any event you want and ingest it into the platform. So you could create a listener that looks at the scroll distance and send an event in based on that behavior or any other behavior that you wish to specifically instrument. Yeah, because what you could do is in the profile, you can actually set up a bunch of different attributes. You could set up attributes that are dependent on the present session or the next session. Or you can even do kind of time-based criteria in the product so that with our offers capability, for example, you only see something once per 24 hours or once per week or every second page view. You can do all kinds of interesting things like that. One more question. Yeah, so we're very excited to be, I mean this isn't public public yet, but we're launching a product in the next month actually at our Engage conference that really is, it's an omni-channel offering and really what that means is it's a set of APIs. So Java, REST, file-based APIs to get data in and out and to get decisions. And then a data warehouse that you can connect your own BI tools like Toblo or any of the other big BI tools right into that visitor data warehouse. So you can go and discover new segments yourself without having to go through a vendor's own tools. Great, I think we're out of time. I wanna thank you all. You've been wonderful, we appreciate it. And thanks again to BBC and Steve.