I'm gonna talk today about how we've used data so, and I'll mention big data, hopefully I’ll only mention it once, it’s not a term I particularly like but as marketers, I'm sure you're aware of how you can use marketing to turn something that's quite old into something that sounds shiny and new, so we'll talk a bit about that, how we've focused on data and how we've used it with marketing. And then some of the things we’ve done with marketing. Our story’s quite interesting.

So effectively I think one of the earlier speakers was talking about this; about insights, you know, that is the only thing that matters, you can have people talking the amount of data they have, it's like a, it’s like a pissing competition - “how much data have you got?” “I’ve got loads of data, I've got more data that you, I’ve got gazillions of data”, I've met a guy from Nasa that's got more data than I will ever want to see! and basically none of it matters unless you know what questions you’re asking, what are the insights you want from it are, so one of the key things that I'm gonna talk about is building up from the foundation. What are the things that you can prove with data, what do you know? I’m gonna talk a bit about getting the right platform and getting organised, so when you think about how you use your data and how it comes together and then I’ll give you some examples.

So, it’s a marketing conference so what’s a CTO doing here? Well, that’s... a couple of people ask me why am I going to a marketing conference, as well at work, including the marketers, so the question is, - it’s sort of a question which I guess most people I would hope don't need to ask any more, but it’s still the difference between what a a CMO’s doing, and what a CMO’s doing is, where it doesn't work, it’s because people are not talking and thinking about explaining what they want and I think the most important thing as a CTO is to make sure that you’re building the right things and you’re getting the right platforms and you’re protecting the longer-term vision, because there’s lots of shiny stuff out there and lots of distractions on the way, I’ll talk a bit about that, but things that stop you from getting to a critical mass. So the importance of technology and the overwhelming power of using technology well in this space is something that’s very important. I think you know, this is a, from XKCD, so it just is a, it’s a data joke - if there is such a thing. But when you think about the people who understand this joke and the people who don’t understand this joke, one of the things that’s quite important to understand is the way advertisers talk about data and the way marketing people also need to talk about data and think about data. And in every way you can use data to your advantage, you know, the old adage about how can use to statistics to tell lies is still true, you can make data say whatever you want if you’re, if you’re prepared to, but I think it’s understanding how to get to the real truths behind it and what you can do with it and how you can use it to your advantage is the most important thing. So getting the balance right between building a platform and thinking about technology and caring about it is incredibly important. So what makes the geeks happy? I'm gonna talk about that and this guy will pop up occasionally. So we’ll talk a little story on disruption. So this is our problem, so going back to about 2010, so you can see the internet was becoming more important. We’ve been charging for content online since 2001 and we had the paywall, the metered paywall we pioneered in 2007 and the reason for that was fairly obvious; it was clear that print was in a slow decline and before it became a fast decline, we wanted to do something about it. And you, the idiocy of giving your content free online and selling for it on the newsstand makes no sense and of course everyone realises that now, but back in the day people thought the internet was just this little thing with it’s “w w w” and it was never gonna be a big thing, so didn't take it seriously enough.

So this is the problem we had, so this is your subscriptions going through print, so you can see that’s in decline, there is no let up, it will at some point print will stabilise because frankly, one of the great pleasures in life is on a Saturday morning is getting a paper or a magazine and sitting down with it in a quiet corner in a cafe, with a coffee, you know, there’s just some things that, there are people who will always like that and I like it and I still pick up the newspaper on the way into the building, you know, I can't help it, it’s a, it’s a great format but you can't deny that this is not something that’s gonna pick up. So if you, this is, you know, people talk about disruption, this is real disruption, this is disruption, not just ‘cos your business isn't doing so well, it's disruption because the platform that you sold on is going, print is becoming less of a mass-medium, people are moving to other formats, so that's real disruption. You're in the middle of it and what do you do? So what we did is we managed to do this. We turned it round and took our digital subscription business and grew it beyond our print subscription business which crossed over a little while ago - round 2013. And that was a key point for us, at that point we had more digital subscribers than we had print and now we have more digital subscribers than we've ever had. We’ve got more subscribers than we’ve ever had. We’re at about 677,000 now and that's, and what what the internet had done for us, what digital has done for us is allowed us to reach a more global audience. So we’ve turned it round - thought about what can it do for us, not stop worrying about the fact that the blues lines going down, what can we do to make the red line go up? And lets stop pretending that we can fix the blue line. So that's quite an important thing, you know, in an internet business, they call that pivoting, I suppose but it’s not really, it’s just thinking about your platforms and your outputs but just trying to find different ways of looking at your revenue and there’s other things we do, I'll talk a bit around the subscriber value, but events and other types of business we run.

So these are some of the stats. So 677,000 that’s 11% year on year growth - a 32% growth in digital subscribers, mobile readership is now the majority and it's over 50% of all traffic but it's depending on which day of the week you look at, it's up to 70 or 80% of traffic. The only thing surprising about that last statistic is that it's not surprising anymore - everyone wheels out their mobile stats now and it’s, everyones saying they've got more mobile traffic and it’s a big platform for them and it's bigger than it ever was and it's growing and that’s, that’s a general truth so that's one of the key takeaways from this year, when i was at mobile world congress earlier in the year was just the size and scale of the mobile industry is fascinating.

So how did we do it? So I think one of the most important things is thinking about how you do data and think about it from the top. So if your board doesn’t care and your CEO doesn't care and your heads of departments don't care, then you won't get the momentum you need. And by momentum I mean the investmentment. So we have thirty data scientists roughly. So for the size of the organisation that's quite a lot. And it's because they will generate revenue they will bring sales in, you know, they’re not just a cost centre, they are people who will generate acquisitions, they will generate understanding and advertising, they generate money. And it's getting to that understanding that’s key. So this quote from our CEO is a really key one, I don't know if everyone can read it, but it’s, when your CEO is able to say something like that and truly understand the value of data, then from the top down, it's gonna work. And it allows us to shape our strategy with data, it allows us to power our marketing, it measures our effectiveness. all these things. And they’re driven by data, it's underpinned by data. So that's a really big thing, you've got to sell it at the top. As I said it's not about big data, so I have used it twice now - there you go, I've failed! But the actionable insights, the most important thing, so even if you can find out certain insights and you can’t action on them they're useless to you. You've got to find things you can actually do and find the data that proves how you can move from A to B. So find the insight and find the things that will prove to you that you've actually moved from A to B. Those are really important things. So when you think about your outcomes: So our outcomes are growing over time, the number of things we like to look at. So we look at, obviously the number of, because we have a subscription and a registered payroll model, we are, we don't just think about traffic, we think about the number of articles you, as a reader, as a subscriber, have read. We think about how many times you visit the site, we think about your lifetime value, we think about things that are bigger than just how much traffic we get and that affects lots of things we do. We’re doing some interesting work with advertising at the moment around not selling by CPMs, selling by time, because with our barriers, with our paywall, we know more or less who you are when you log in, so we've now looked at the model between CPMs and when you're in CPM model you’re in a losing war against the trillion there’s something like a trillion ad impressions a day available now between and Google, Facebook, Twitter and the others have scooped up most of them so if you're just trying to sell on volume you're immediately in trouble. Most of those impressions half of them are probably bots, there’s loads of rubbish out there, so it’s not that valuable for an advertiser. One of the most important things about data that we were able to prove that if you advertise on the FT, we can sell your campaign by time and we can prove that someone saw if for up to five minutes for example if that's your metric. If you want CEOs to see your advert for five minutes we can do that and we can prove they saw it. That is incredibly valuable to advertisers and much more valuable than just loads of click-throughs, or even just CPMs. So thinking about outcomes and the data’s allowed us to define these outcomes and measure them.

Now the insights may not be obvious. So in the urban myths of data does everyone know the connection between beer and nappies? This is always my baseline. So back in the day in urban myth there was a supermarket who found there was a certain correlation between the sales of beer and nappies at certain times of day and they couldn't quite figure this out, it didn’t seem to make any sense, and what it actually was was dad being sent out at some time in the evening to just go and get some friggin’ nappies right now it's an emergency and while he was there he would think maybe I'll just pick up a six-pack on the way out. It’s an American supermarket so, so it’s an urban myth, there’s various people I’ve met who all claim to have been the data organisation that figured this out. In fact, most of the ones I talk to seem to have claimed the story but that doesn't matter, I think the most important thing is understanding that every supermarket you've been into since about 1990 is scientifically designed in the way everything is placed within the store, how everything is placed on the shelves, what’s the loss leader, what’s the thing that’s dragging you in. You know, if you ever need proof of the value of data, look at organisations that would not survive if they didn't use it. And I'll give you a, so when you look at what they connected - that's the real sign, that’s what the sign you should probably read to be honest - but it just proves that when you connect things together, it may not be obvious but there's an insight here that was important and the insight was certain things are valuable to put together. It may not be obvious, it wasn’t that you just need to put beer and nappies together, it was take that up a level and go there are certain things that are connected, figure out what's connected in the supermarket, what do people buy at the same time, and get those things either together or deliberately put them at different parts of the store so they have to go wandering through, you know, it's that sort of stuff which can generate sales and generate interest.

Another example, so UPS in the US don't turn left and they proved this by using software to look at the number of miles they drove and they figured out they were spending so much time at right turns waiting to turn, it was quicker to plan a route that only turned left and those are the stats that proved it. They saved 20 million miles delivered over a quarter of a million more packages and reduced the amount of pollution they made, saving into hundreds of millions. That’s the number is 2010 so God knows what it is now. But, you know, it's an astonishing statistic but it’s just by not turning left, so in the UK it would be not turning right so, you know, it's a really interesting thing to think about when you use data what’s the outcome you want. Do the most important thing is thinking about data. So people often go off and buy a big data product and some other product and start chucking data into it. Before you do anything get some alignment in your data because I guarantee you it's in a mess. And I'll give you some examples.

This is how computers talk to each other, so if we just turn the volume up but people think, you know, computers are really smart and they're able to communicate in complex ways - this is actually scientific research into what computers look like when they're talking to each other.

[Plays Video]

I’ve studied that video, I think they’re talking about a sock - one of them’s lost a sock, I don't know. So they are clearly speaking the same language, they are clearly just speaking the same word but they have no idea what they're saying to one another, it's completely unobvious what's going on there. And the key thing to think about is when you look around your organisation and you try and understand your outcomes and try and understand the data that you need in order to drive them, it's really important to get some alignment and understanding. You know, so some examples. So I'll give you an example: So how do you define a cancellation? If you were to say if somebody didn't pay, so when they came round to renewal and defined and you said they've not renewed, would you say that’s a cancellation? Just a rough show of hands, would people say that is something cancelled? Ah! You know I'm gonna trick you! So no, that's just a payment failure. So when you think about what might happen next is they might, we might go through and say well they have cancelled, they're not paying for the right reasons, it's not just payment failure but then we've called them up and that's what we call a save, that’s where customer services have kicked in and persuaded someone to stay. You then have, so imagine something’s gone wrong and you've not charged them properly or you've charged them the wrong amount - that we call a cock up, technically, a bit of technical terms in there, but that’s when things have gone wrong and that’s another different definition. The worst thing is you make a mistake and you have a whole load of campaigns ready to run which, then say, so the worst thing you can do is accidentally charge someone too much or accidentally cancel them and then send them an email saying “Oh my god, you've cancelled”, that’s like let’s do XY and Z to make you stay. As far as they know, they haven't cancelled at all, they have no idea as to what's going on.

So there's a guy in my team who lost a month of his life to getting everyone to agree what was a cancellation and that’s really important to know because when someone cancels like a completed sale or a not completed sale, you want to know have they cancelled. If they have cancelled, what are we gonna do about it without them knowing and what are we gonna do about it with them knowing. So if you can't understand those things, no amount of data is gonna help you, you’ve got different parts of the organisation are all gonna kick off doing things, running campaigns, calling them up, it’s the worst thing you want as a customer it will just convince them that they've done the right thing in cancelling. So really important to get organised, there's only a certain amount of data in the business that's incredibly important to get right, common understanding of important things, but if you get that right, you've got really overwhelming power in your technology to drive good marketing and drive good outcomes. So this is a real world example; it’s a marketing problem; trying to get two systems to talk to each other, it’s a bit wordy but this is verbatim from the person who sent it to me. So I'll read it out just in case:

“We’ve got a bit of an issue in Salesforce in that we have the correct data standards formats for countries there but some of these differ to those in Oracle.” They're trying to connect up customers and billing and try and understand what to do. “The list of countries shown below is the Salesforce ones and the ones next to it are the Oracle ones. How do I get these two thing to talk to one another?”

And basically this is the problem. So you have one saying the Bahamas, one saying The Bahamas. One saying Gambia, one saying The Gambia. One saying Myanmar, one saying Burma. How do you know what is going on in your business if you can't even understand where you've made the sale? If you think you have two countries, one called the United States and one called USA, you know, this stuff is the fundamentals of doing anything with data. You’re wasting your money on buying anything, building anything unless you get this sorted out, it’s incredibly important. So when people talk about technical debt I would say data debt is one of the worst ‘cos data debt is this invisible, lurking thing which everyone thinks if fine and no one cares about until it goes wrong and everyone assumes just magically works and it's always there and it's the most expensive thing to go and fix, it's an absolute nightmare to go and fix this stuff.

So another thing for, key thing around marketing and what we talk about with our marketing team is understanding the environment, understanding what’s going on. So this a view of FT usage. So the pink line is mobile, the blue line is desktop. And you can see during the week this is, [you] know this is already a massive change to what we used to see, you can see the commuter times when people are using their mobile devices, but what’s interesting is; in the evening, you know, it’s not just the commuter time, they are actually primarily using their mobile device anyway. What’s fascinating is what happens at the weekend. So the weekend, no ones using their desktop at all. Now it doesn't mean the desktop is dead, it just means, it’s not so much mobile first it's just desktop last, is the way I like to think about it. People go to the desktop when they need to, when there’s something they can do better on a desktop. So this changes your marketing and the interesting thing about this, was when you look at insights. I've talked a bit about the speed of of data earlier and we notice some interesting things about how people use the FT and then what we should do around our marketing and how we should focus. Obviously there are things we need to do differently at the weekend. We have to think about how we measure things, so we’ve got all these different devices and outputs, we've gotta think differently about how people use stuff on those platforms and we also have to think about how we make stuff. So the traditional - this isn't where we are now but some time ago, we looked at when are we writing writing stories and when are people using the content and you can see the disparity on the newspaper workflow which is very much aimed at “let’s get the paper out and then we’ll come in late, get that done, stay late, finish off and print the paper”. That doesn't fit with a digital workflow that, where people are possibly most engaged in the morning, or possibly late in the evening, but you need to be thinking about what's current at those times. So all these things change your environment and in understanding your environment you get to understand the puzzle so you start to understand the pieces and you start to get focused on the right things at the right time. It gives you an insight into the right questions to ask. When we think about being organised, I think the most important thing is to harass your IT team, you know, if they're not doing what you want, don't go round them, go and force them, embarrass them, shame them into to doing what you actually need them to do, because when you see a team that is well organised, you know, who have marketing and technology completely aligned, you have Google or Facebook, you have an organisation that frankly is absolutely terrifying to the opposition. You know, someone who is well organised in this space, can do things that look close to magic and all they're doing is they’re just talking about the obvious things to do, they’re using the data and they’re connecting it up with the technology and they’re making decisions based on that and getting organised is really important.

It's important as well for marketing people to market within their own organisation, so think about this is a dashboard, it's deliberately a bit fuzzy ‘cos theres some sensitive data on it but basically this is a dashboard we put into the newsroom and it tells them how often their stories are being read, on which platform, on social media, on various things and it was really important for them to understand that so that when the marketing team come down and say “we'd like you to do this, and we'd like you to promote the story and we'd like you to use social media” that they can show “and this is why” - because we’re getting traffic from this route. It's not about changing the editorial, it's about promoting your story, it’s about using social media, so using tools like this is really powerful and it, you know, as marketers - make people care about marketing within the organisation. Another thing is also to champion things that other people care about so obviously, latency is an issue. So when you look at websites, you might be doing some fantastic work on your front end, running some great campaigns. If your website’s taking too long to load, nobody's seen it, nobody cares, it's not happening and you'll be following all the wrong leads down to make the wrong decisions because you don't understand that your page is taking ten seconds to load so this is just a breakdown of one our page loads in Sydney and there was an issue with that which we fixed and it was really important for everyone to understand why that's important because if things are too slow, you don't make money. Now there's an absolute co-relation between effectively your shopping basket, in whatever form it is and the number of clicks and the amount of time it takes to get to the completion and fulfilment of that order. Every click, you lose 50% of your audience, you know, those sort of metrics are scary but largely true. People drop off really fast. So you might be optimising your campaigns and optimising your site and doing some great A/B testing and think it’s all taking you in one direction but it might just be no one's seen it and they actually wanted something else. So you have to understand and care about these things. If you care about these things, the people that care about these things will also care about what you care about. So all those things are true unfortunately, bigger’s not better, faster to end, it's good.

Latency can undermine a lot of things, you know, it's distraction in the data. It's bad. Bad information is being provided. If you don't hone in on these things it's really very distracting. So I’ll tell a little story on tracking - so people obsess about tracking things. There was one agency we were working with who wanted to put a tracking script on the front page that tracked every single interaction and was every, 250 times a second it wanted to track every single user. Can you image the number of people on the FT homepage with this script running for all of these people? And we had a much better way of doing it, but this this thing appeared, the homepage ground to a halt, we had to remove it immediately, and it was just figuring well “how did this get live? Whose idea was it?” And we sort of traced it back. I think when you look at what you're trying to measure, it was, it was sort of an air of, there’s an air of...i wouldn’t say desperation around it but what they actually wanted to do was run a survey. So if you wanted to run a survey you didn't need a tracking cookie that was tracking everything people were doing, you just needed a pop-up the survey at the right point - maybe a certain time after the page has loaded, so this was the only tool they had so they used that tool. So I think the conversation was really important to think about how you decide what it is you're gonna do and that's where the different people around you can help. The technology people and the advertising people, they've all got tools in their box that can maybe help, but you can very much focus on trying to track everything and it’s a really good example of just having more data than you can possibly use or possibly understand and generate all these tracking points and then what are you going do with it? You only really wanted to run a survey. So it's quite important to think about these things.

We do do a lot of A/B testing and it’s, you know, it's gold standard stuff, you should do this. We can make a lot of clear points around how effective it is but it is very important to track it and have the right tools so you know the outcomes, the right test groups, the right control groups. We like that.

So consistency is a key thing, so when we're looking at different platforms we need to be able to track people across, we need to be able to count the number of articles you've read, serve barriers if you've gone over your limit and we've got to do that effectively across different devices. So consistency is really important. And as I said, latency is your enemy and I'll give you another story about latency. So the power of having a platform is being able to do things quickly. We’re running the beginnings of data at the FT, when we really started to crank this up was when we realised for particular campaign they were running it was a RFE-based campaign and basically we were looking at promoting based on people who has a propensity to subscribe, based on some model we built and we were sending emails out after a certain period and part of the problem was it took a long time to get the data out so it was taking seventy two hours to follow up on looking at what people had done. So these were registered users who wanted to convert to subscribers, so we were looking at what they they'd done, trying to find out what we should do, were they in the model and then push that data to make an email and send it and that was taking quite a long time. Various things were fiddled with on the campaign that made a very, very marginal difference. We then decreased, we put some more power into the system and thought more about the platform, made it go faster and reduced that to forty eight hours from seventy two. There was over 100 times increase in number of people who responded and followed through and surprised so we now got that down further and further, it's down to twelve hours and it's one of our most effective funnels now. And it's one of the things that’s important, That’s because we focused our efforts and have lots of people with spreadsheets running around, we started to focus down and think right, how do we get people to subscribe, how do we do it fast enough, get good data, good models and make it work?

We collect data from lots of places. So we push all this into the funnel - let’s put that up a bit - so there’s stuff from website, mobile, tablets, desktop, we take stuff from social media, we look in all sorts of places and we pull that data in and we have a - just done something weird to the slide there but anyway - we take registered users, I'll go through the different models we have. So anonymous users are people who can see bits of the site that can't really see the content, that can see video. And we have market data - some of market data out - you can't use the portfolio tools. Registered users can see eight articles a month and subscribers get unlimited access and then we have other subscription models as well. So that's our, those are our packages and our funnel is bringing people in. So we try and bring people in, try and get them to the site, then try and get them to register, then try and get to them to subscribe, so clear view of where we want people at each stage, what we’re trying to do at each stage, different techniques, different things to think about. And when we think about the RPU, so the average revenue per user and extending that to lifetime value - this isn't the exact numbers but it’s just to give you a feel of understanding the importance of different things. So we look at things that are fairly low - average revenue per user, registered users, then you move those to newspaper subs, standard subs and then there’s premium subs on the, in here and also how we look at events in other part of our business. There’s lots of other bits to the model but it's just in understanding this and understanding where users are on that curve, we have a model so we know what it is we're trying to do.

Always in all technology slides you should have some slides in that the people at the back can't really see and are very difficult to understand, so we've got a few of those but effectively this is a view of two things. This is how an element of our, and we’ll talk a little bit about our platform and what we use as a third party. So were using DSPs to identify and serve basically, serve advertising, serve incitements to, incentives to register and subscribe, we're also, we use a third party to help with pulling some of that data in and how we promote it, so it’s, marketing’s almost become like advertising in this space, we’re looking at data to figure out where to push and then build models around that. We then connect that up with our data which has the grand view so that we can understand the outcomes from those systems because otherwise the third party would start to have more data than we would, we wouldn't really know what is going on. So what we do in this bit, which is what the third part does, which is pulling together some of that short-term data and what we do here, where we look at the long term value and look at our important decisions, those are things that we have to be very careful about what we do in each box. So knowing, it’s not that you shouldn't work with third parties, sometimes they have really good products to offer and they help and we do as well, but it’s knowing everyone you meet who has a data product will want to be your data warehouse. They all want to be your data warehouse, every single product you buy now will want to sell you a data warehouse and then you’ll end up with ten data warehouses which sort of defeats the object so be careful about what you get them to do.

This is our data warehouse, there’s lots of pieces but basically the simple part is we have various bits of data that come in from different sources. We use Amazon to host it which is another whole presentation, but we we found that was really successful for us. It allowed us to scale really quickly and the power was far beyond what we had hosted before. If anyone wants to talk about data warehouses, I will bore you to tears about data warehouses. But the thing that is interesting thing is we've got data coming in, we thought about how we were gonna get it in, how quickly we needed to get it in and then we looked at how we want data out and the most important thing there was moving away from reports to real-time data. So the most important thing for us was not being able to make decisions on a weekly basis or a daily basis but being to make decisions in real-time so we can very quickly decide what we're gonna do based on any particular scenario we see and that creates some really powerful, interesting changes.

The most interesting thing about that was finance and marketing and other teams, when they saw the report, they all said well we want the reports back - they used to give us reports, we want the same reports out of this system, and then we gave them some cubes and some real-time interfaces to data and they just, we just didn't build the reports because we knew they wouldn't ask for them after a while and they didn’t. They love being able to go into the data in real-time and seeing what's actually going on from a single point of truth, you know, that's so much more valuable.

So, slightly wordy slide again, sorry - but another quick thing just on security and it's important to think about this, how you control data access, so who can see what, who’s able to make decisions on what, you don't want everyone to see all your data and one of the things we do with that is we use single sign on so the important thing about this is being able to save time so we can get data quickly, they don't have to log into another system. If they’re logged into the network, they can see the data that they are allowed to see and again that's very valuable. I mean, these are quotes from people who use the data but it's incredibly valuable for them. And one of the things that's important about that is it makes it easier to stay in the system. So this is one of the quotes but very quickly; so the data science team had a problem that would have taken three hundred hours to solve. In the old days they would have downloaded all the data, fired up a spreadsheet, installed some stuff, tried to process the data trying to make an insight. Then they might have deleted the data, they might have forgotten about it, they might have put it on a USB stick, they might have just lost it somewhere. That's not very secure, it’s also not very effective. So they key thing is to build the right solution, the right system - it was actually easier for them to do this in the data warehouse. They just literally had to press a button and it got the data for them, set up the environment, they were able to process it, look at what report they were trying to create and get the insights and move on, all managed. And that’s one of the most important things, that's the, you know, the investment in data is key. We like that.

So what sort of things so we do with all of this - how am I doing for time actually, how much? Ok cool.

So we support our goals with the data, so if we look at, we’ve looked at registrations, we've looked at which people were managing data, how they were understanding the registration barriers, -what people were doing when they got to the registration barrier and what failures they had and just by making some simple changes, so maybe testing and some removal of some steps, we were able to make 100% increase in the number of people who registered and 100% increase in the number of people who subscribed. Those are amazing stats, you know, you’re doubling your funnel just by making some changes to the way the interface works and understanding what were the right things to remove, what were the right things to fiddle with. So really key stuff. We measuring engagement so a bit washed out, but basically this is around recommended reads, this is using semantic technologies and looking at contextual relationships between things, pushing the right content to people who are interested in it. Payment platforms, if you’ve ever dealt with payment platforms they are all a pain in the neck., it's incredible how hard it is to get these things to work properly on the internet after years of ecommerce. When you’re doing something simple it's all fine, you start getting into card numbers and subscriptions and direct debits and so on, it all starts to get really hard. So things like understanding the right people to work with around managing payment options, account data that sort of stuff, allows you to keep when credit cards expire which is a huge problem in the US. Hundreds of millions of credit cards were hacked and were lost last year so they had to be renewed when people changed their credit card, you need to find out what the new card number is, again that's not a cancellation that's a payment failure. Empowering your customer services team, so giving them a model so they know how important customers are so they have some sort of metrics, they understand what to do when someone calls up, so they get context. This is a model we use and it allows them to do things like when they made some calls and looked at how people use products, which were the high value products and they were able to make increases by calling people up and driving some deeper engagement, so talking people through things, you know, we’re quite a personable organisation, we will call you up, we’re not shy to do that so we also look at propensity models, so building models and scenarios so we can see who’s likely to subscribe, what they're looking at now, how we push the right barriers towards them and what sort of incentives we can offer, what content they've consumed, etc. This is a big funnel for us. This generates just on it's own 15% of new business. And being smarter in acquisitions; so this is around using models with DSPs, I won’t sort of focus too much on that, butwe look for models for potential subscribers outside the FT, we calculate the propensity and then use DSPs to serve advertising and target them. Also don't forget offline. We just did a, you may be familliar with “How to spend it”, so we had when we released the, rebranded the newspaper, well not rebranded but we redesigned the newspaper we had a “how to fold it”, which was very popular, it's powerful. It goes on social media, people love this stuff, they share things. Any shareable thing is a really powerful thing so don't forget the power of offline and posters and video and so on.

And don't get distracted as marketers, don’t get distracted by the marketing. People slap big data on a product, it doesn’t mean it's gonna solve your problems. It’s very interesting to think about terms that come out, so things like responsive web design, so thinking about cross-platform. When you think about responsive web design, there are some issues around what that means for you, so how you design things you can make a website slower by trying to make it responsive and that's not the goal, so you have to be clear about what you're pushing for. Let’s go through this, it’s going a bit slow sorry - so thinking in a universal way - so we think about universal publishing, we think about all the platforms we want to be on, so we’re not thinking about making websites smaller or bigger on every device but thinking about what's the right content on the right platform and this is really important for your decisions around how you push and advertise things ‘cos these are all the platforms we’re on, there’s loads more, some of them are syndicated so things like Facebook we don't even control the presentation, we’re just sending them content, they control how it’s presented and just when you think there’s enough platforms there'll be more, there’s always another one round the corner. We deal with this by doing, so universal publishing means using API’s and using powerful ways to get content out in a consistent way, so worry less about the visuals, worry more about the message and to give an example this is something we've just launched which is a smartwatch version of the FT. So the question around smart watches and wearables is “How on earth are you ever gonna read anything on a smartwatch?” Well the insight was to use a speed reading tool, so basically this pink bit here, sorry - this bit here is speed reading the article to you. If you've ever used speed reading, if you’ve ever used speed reading, after you get up to a certain speed you think it's gonna fry your brain but actually it is actually a very effective way of reading stuff on a watch, so it was the right technology married up with the right solution, as you can see it has an interesting space for advertising and other things as well so how you use these platforms will often only come through experimentation. You’ll find the right option that way.

And just one final thing - so when you think about building stuff, it's often very tempting to just push every button and try everything and that will leave you with a certain amount of debt. It will leave you with some baggage that you'll have to clear up at some point so try and be more focused and find some quick wins, things that will make real differences. So for us things like the subscription barrier and the propensity models work really well. So focus on something you know you can get an outcome from and the investment in data, getting thirty data scientists doesn't come from just trying lots of things. It came from a really coordinated effort to prove the value of data and how it was gonna help advertising, help marketing, help the business so think about the platform, just think longer term about what you want to get out of this in a year, two years, three years. So that's it basically. Get a central view, try and organise around it, put some solid foundations in around how you use data and how you link it and understand you’re talking the same language and get people to care about this stuff. Because they do care about marketing and they care about your goals, so they will also be building the right things so when they build products, they build monitoring in, they build a new website, they build the right tracking in - it's not something you have to go out and ask them for. It’s there, they’re doing it because they know it's something you’re gonna ask for. So that’s, that's me done. Sorry I may have gone slightly over there, sorry but thanks for listening.