WEBVTT 00:00:14.649 --> 00:00:19.709 The Grand Dumb of Journalism.co.uk and News Rewired. How are you, Marcella? 00:00:20.949 --> 00:00:24.169 Grand. It's Journalism.uk. Good. 00:00:24.749 --> 00:00:30.569 Very good. Yeah. So tell us, right, you've taken over the site, changed a few things. 00:00:30.789 --> 00:00:34.969 So let's go back to me and tell people what Journalism.uk was. 00:00:35.249 --> 00:00:38.769 It was Journalism.co.uk. You've slightly rebranded it. 00:00:38.909 --> 00:00:44.709 Tell us what the backstory is. right and hi Charlotte and I'm delighted to be 00:00:44.709 --> 00:00:46.649 here on the podcast I'm really thrilled to have you, 00:00:47.649 --> 00:00:55.849 so we are a trade publication focusing on the journalism world and we've been founded in 1999. 00:00:57.149 --> 00:00:59.989 When journalism started to appear on the 00:00:59.989 --> 00:01:02.849 thing called internet that was 00:01:02.849 --> 00:01:05.909 not very well understood even less embraced 00:01:05.909 --> 00:01:09.729 by newsrooms and the founder his 00:01:09.729 --> 00:01:12.709 name was john thompson started as a as a 00:01:12.709 --> 00:01:16.049 blog pretty much in a jobs board and just exploring 00:01:16.049 --> 00:01:19.609 what this digital publishing thing is all about and 00:01:19.609 --> 00:01:26.289 it grew over the 26 years into pretty holistic service for journalists so we 00:01:26.289 --> 00:01:31.649 do still publish editorial and have a job support and that we also have a conference 00:01:31.649 --> 00:01:36.729 called news rewired It takes place twice a year in London and a number of other 00:01:36.729 --> 00:01:38.709 services, freelance directory. 00:01:39.009 --> 00:01:43.869 We provide training courses and generally are a resource for people who are 00:01:43.869 --> 00:01:46.009 looking at innovation in the media industry. 00:01:46.269 --> 00:01:50.249 We want to provide them with information to give them the skills and the confidence 00:01:50.249 --> 00:01:53.169 to just make decisions and keep journalism relevant. 00:01:54.196 --> 00:01:58.576 Yeah, I want to come to the conference in a bit because I've been a media partner 00:01:58.576 --> 00:02:01.036 for it. I've been a few times, always enjoy my day there. 00:02:01.156 --> 00:02:07.736 But let's talk more about the site because you took over kind of the start of this year, really. 00:02:08.796 --> 00:02:13.136 Right. So the start of the year is in January. Yeah. Rewitness. 00:02:14.436 --> 00:02:19.096 Well, no, because it's been going on. You took it over. May last year, yeah. 00:02:19.596 --> 00:02:26.196 Exactly. And it's kind of been going on since then. And there's been some transformations 00:02:26.196 --> 00:02:29.736 that we're particularly seeing now. So what have you changed? 00:02:29.996 --> 00:02:37.096 Right. So I took over the website in May. I bought the brand from John and we are a team of three. 00:02:37.436 --> 00:02:40.916 So really as little people as possible to run an operation of this size. 00:02:41.916 --> 00:02:46.296 So there's only three of you. Indeed. Yeah. And, you know, the website. 00:02:46.656 --> 00:02:48.356 Right. So what didn't work? 00:02:50.276 --> 00:02:55.716 Tell us the dirty secrets. Listen, what we were doing was great. 00:02:55.976 --> 00:02:59.556 The platform we were doing on was causing a number of issues. 00:03:00.436 --> 00:03:04.076 First of all, the first thing was design. And you might think it's, 00:03:04.156 --> 00:03:05.396 yeah, you know, whatever, design. 00:03:05.696 --> 00:03:08.516 I mean, that website hasn't changed since 2011. 00:03:08.836 --> 00:03:11.976 Someone said, you know, just wait for a few more years, it'll be vintage. 00:03:12.236 --> 00:03:16.096 And people will literally just come and see it out of curiosity. 00:03:16.096 --> 00:03:18.576 Like, how can something like this still exist on the internet? 00:03:19.236 --> 00:03:21.996 And, you know, it wasn't just the question of aesthetics. 00:03:23.076 --> 00:03:29.016 You know, we were, and we still are, providing intelligence on innovation in digital publishing. 00:03:29.316 --> 00:03:34.116 And what we looked like was, I mean, it couldn't have been more at odds with 00:03:34.116 --> 00:03:35.196 what we were talking about. 00:03:35.756 --> 00:03:40.796 And at some point, you know, people who were our loyal readers and users kind 00:03:40.796 --> 00:03:42.616 of got used to it, slash got over it. 00:03:42.736 --> 00:03:46.196 But every time we met someone new, I mean, I had people on the call who were 00:03:46.196 --> 00:03:48.176 like, oh, yeah, so what do you do? that this is our website. 00:03:48.576 --> 00:03:51.216 And people would open it on their browser, and I would just see them go. 00:03:52.459 --> 00:03:55.539 You know you can't have people laughing at 00:03:55.539 --> 00:03:59.239 your website it was it was bad especially when 00:03:59.239 --> 00:04:02.299 you're working in the space you're working in right exactly so 00:04:02.299 --> 00:04:05.619 you know that i mean you know it was it was really causing reputational damage 00:04:05.619 --> 00:04:09.339 i mean and it was noisy it was extremely hard 00:04:09.339 --> 00:04:12.899 to navigate lots of money making features were were 00:04:12.899 --> 00:04:15.899 hidden uh in in the noise 00:04:15.899 --> 00:04:18.979 and unless you truly knew that website and 00:04:18.979 --> 00:04:21.699 knew how to navigate it you just 00:04:21.699 --> 00:04:24.619 couldn't get your head around it and i think loads and loads of 00:04:24.619 --> 00:04:27.599 and we see the loads and loads of search inquiries have 00:04:27.599 --> 00:04:31.519 probably just been lost to people who just abandoned you can see people sort 00:04:31.519 --> 00:04:38.379 of bouncing off so what did we do so in a space of you know i bought it in beginning 00:04:38.379 --> 00:04:45.879 of may then uh two weeks later was news award conference The timing really was crazy. 00:04:46.279 --> 00:04:49.859 So once we kind of got over it... Can I just say, when I saw you at News Rewired, 00:04:50.119 --> 00:04:53.799 because I was a media partner for the latest one, you were tired, Marcella. 00:04:54.579 --> 00:04:57.559 Let's put it that way. I mean, you know, the main News Rewired, 00:04:57.739 --> 00:05:03.399 we literally just completed the company buyout and it's been a long and emotionally 00:05:03.399 --> 00:05:06.039 and financially draining process. Sure, these things are never straightforward. 00:05:06.739 --> 00:05:10.739 And for the November News Rewired, we literally just completed the rebrand and 00:05:10.739 --> 00:05:15.479 relaunch. So it went on the 2nd of November and then the 26th of November was the conference. 00:05:15.699 --> 00:05:20.059 So we just need to change the timing of the way we take on big projects. 00:05:20.079 --> 00:05:20.719 Yeah, you do things the easy way. 00:05:20.999 --> 00:05:25.559 And so the first thing was once we kind of got our act together after News & 00:05:25.559 --> 00:05:30.739 Wired, we started to look into how to modernize the platform. 00:05:30.919 --> 00:05:34.459 So we needed, it was absolutely obvious we need to get a new CMS, 00:05:34.719 --> 00:05:35.939 the backend and frontend. 00:05:36.099 --> 00:05:41.339 I mean, everything needed to change. what we were using was a bespoke CMS that 00:05:41.339 --> 00:05:44.779 was created by this one guy in the 90s and no one in the whole entire world 00:05:44.779 --> 00:05:46.939 knew how that thing worked apart from this guy. 00:05:48.179 --> 00:05:52.779 So, and it couldn't be changed, you know, the structured, the design, 00:05:52.959 --> 00:05:54.039 nothing really could be changed. 00:05:55.196 --> 00:05:58.476 There were a few frontrunners. I mean, we obviously worked in this space, 00:05:58.656 --> 00:06:02.936 so we have insights, you know, just by doing our job, we knew about lots of tools. 00:06:03.816 --> 00:06:05.496 And there were loads of CMSs. 00:06:06.616 --> 00:06:10.356 We had a short list of CMSs that were great. I was like, yeah, this would really work. 00:06:10.956 --> 00:06:15.196 And then the quotes started to come in, and we were looking at it, 00:06:15.276 --> 00:06:16.316 and I thought, you know what? 00:06:17.476 --> 00:06:21.256 And this kind of goes to your point about sort of, hey, how our publishing is going. 00:06:22.416 --> 00:06:27.216 I was really looking at these quotes. and I understand what you're paying for. 00:06:27.396 --> 00:06:34.916 But I thought with where internet is going in terms of search and agentic search 00:06:34.916 --> 00:06:36.376 and sort of zero click future. 00:06:37.696 --> 00:06:45.156 I don't believe that if I spend £70,000 of a new CMS, it will work £70,000 better. 00:06:45.796 --> 00:06:49.116 If we spend this money, we'll never make this money back. 00:06:49.556 --> 00:06:55.316 Not with where publishing is now. Were you seriously getting quotes in that kind of region? 00:06:56.096 --> 00:07:00.036 Yeah. I mean, if you take everything into account, so migration, 00:07:00.436 --> 00:07:04.516 development, obviously bespoke features, and the time it takes and the testing 00:07:04.516 --> 00:07:10.816 and prototyping and maintenance and the rest of it. 00:07:11.136 --> 00:07:13.336 That's about what it takes. I mean, for the website of our size, 00:07:13.476 --> 00:07:17.956 I mean, it was 26. You know, we had around 1 million articles in our archives. 00:07:18.996 --> 00:07:22.956 1 million? 1 million, yeah, 1 million. Wow. So, I mean, you know, 00:07:22.996 --> 00:07:25.476 but articles, it was everything. It was editorial articles. 00:07:25.676 --> 00:07:29.736 It was job posts, training posts, announcements, you know, whatever it can be. 00:07:29.836 --> 00:07:32.676 So it was a massive job. I mean, it truly was a massive job. 00:07:33.416 --> 00:07:38.436 And so kind of by elimination, we said, well, what do we really want? What do we really need? 00:07:38.736 --> 00:07:44.856 And so instead of looking like what CMS can do what I need, we flipped the process. 00:07:45.596 --> 00:07:48.756 Where do we need to get? What do we want to be able to do at the end? 00:07:48.976 --> 00:07:50.496 And then we kind of worked backwards. 00:07:51.136 --> 00:07:55.896 And we decided to re-platform to Ghost. And there were a number of reasons for 00:07:55.896 --> 00:08:01.916 that. One was very much, I see the publishing going more in the direction of creator economy. 00:08:02.876 --> 00:08:08.776 And we wanted a tool that will have this sort of, you know, solid newsletter feature baked in it, 00:08:08.976 --> 00:08:12.696 social media posting and, you know, podcast and the rest of it, 00:08:12.796 --> 00:08:20.896 where we see how creating content has to be much easier than going through a gigantic CMS. 00:08:21.276 --> 00:08:23.416 Yeah, and hoping Google sends people your way. 00:08:24.949 --> 00:08:30.529 The second reason was obviously cost. You don't have to purchase it, 00:08:30.589 --> 00:08:31.609 but it's still open source. 00:08:32.309 --> 00:08:38.769 So unlike some other competitors of Ghost, it doesn't lock you in, keep your data. 00:08:38.949 --> 00:08:44.789 And if you want to replatform, take everything or can impose costs on you, 00:08:44.849 --> 00:08:47.969 like we saw with Substack, when they sold you in Apple Store, 00:08:48.129 --> 00:08:49.549 suddenly subscription, they didn't 00:08:49.549 --> 00:08:53.869 increase by 30%. You can't do anything about it. So at least for now, 00:08:54.953 --> 00:08:58.413 hasn't got the structure of it is different and 00:08:58.413 --> 00:09:01.453 worst comes to worst if we need it to leave it's open source 00:09:01.453 --> 00:09:04.833 so we can just take everything every platform you that's 00:09:04.833 --> 00:09:08.053 there's so many interesting points from this and this is quite a nerdy conversation 00:09:08.053 --> 00:09:11.893 but that's fine this is a nerdy this is a nerdy show it's fine you're gonna 00:09:11.893 --> 00:09:15.333 say you're a nerd i'm a nerd you're in a safe space marcella it's fine hello 00:09:15.333 --> 00:09:19.273 to all the nerds yeah exactly you're in a safe space here and but the point 00:09:19.273 --> 00:09:22.173 is lots of publishers are thinking about this stuff. 00:09:23.493 --> 00:09:27.793 And having to have the same conversations that you were having and the same 00:09:27.793 --> 00:09:30.013 thought processes that you were having. 00:09:30.253 --> 00:09:33.993 Now, I don't think we're going to ever end up in a world, or particularly want 00:09:33.993 --> 00:09:37.133 to end up in a world, where everyone is publishing either on Substack or Ghost. 00:09:38.373 --> 00:09:42.853 But people have to think in the way you thought about how everything connects, 00:09:43.073 --> 00:09:45.013 that everything looks clean and simple. 00:09:45.233 --> 00:09:49.033 And most importantly, what I'm getting from what you're saying is 00:09:49.033 --> 00:09:51.993 the most important part of it was that you could talk and 00:09:51.993 --> 00:09:54.853 connect directly with your readers you weren't 00:09:54.853 --> 00:09:57.593 relying on google to send people to you as they 00:09:57.593 --> 00:10:00.753 once did that definitely 00:10:00.753 --> 00:10:03.573 i mean you know in the end of the day what 00:10:03.573 --> 00:10:06.713 do you really need you need to be able to create this direct relationship 00:10:06.713 --> 00:10:09.353 with your audience and i mean 00:10:09.353 --> 00:10:12.753 you know the seo went up the traffic went up discoverability went 00:10:12.753 --> 00:10:16.013 up even with ghost and looking into 00:10:16.013 --> 00:10:18.933 it but you it was from where we started the bar was so 00:10:18.933 --> 00:10:21.893 low even just like loading speed or 00:10:21.893 --> 00:10:26.853 the complete basic things right the theme etc then actually having an on-the-shelf 00:10:26.853 --> 00:10:31.713 solution suited us i mean there were downsides i will get to those but you know 00:10:31.713 --> 00:10:37.613 by um by sort of elimination that really was the best choice we could have done 00:10:37.613 --> 00:10:41.313 and i think what was decisive why we went with ghost was, 00:10:41.990 --> 00:10:46.790 is that we just bought a company. I mean, I didn't have a penny left to invest. 00:10:47.090 --> 00:10:48.550 I don't want on the purchase. 00:10:49.250 --> 00:10:53.590 Going with Go's just allowed us to actually do it rather than spending time 00:10:53.590 --> 00:10:59.430 trying to raise investment, potentially sell equity. We are still a 100% shareholder. 00:10:59.670 --> 00:11:00.910 We are 100% independent. 00:11:01.450 --> 00:11:07.710 We have no investors to answer to. We answer only to our readers and we wanted to keep it that way. 00:11:07.990 --> 00:11:13.230 So I said, you know, in the end of the day, the independence and the speed. 00:11:14.550 --> 00:11:20.270 Are more important than the spoke feature or, you know, this thing or that thing 00:11:20.270 --> 00:11:24.930 that could have been maybe done better by publishing CMS, but it would not allow 00:11:24.930 --> 00:11:29.290 us to move as quickly and decisively and independently as we did. 00:11:29.710 --> 00:11:34.370 Yeah. And sometimes you don't want to let the best be the enemy of the good. 00:11:34.530 --> 00:11:39.010 Did you ever think about moving to Substack and just launching on Substack? No. 00:11:39.650 --> 00:11:44.970 The main reason was exactly the kind of being locked in terms of, 00:11:45.050 --> 00:11:52.750 I mean, the other conversation is monetization. 00:11:53.110 --> 00:11:59.170 We do not want to charge for content and still want to operate on the free content 00:11:59.170 --> 00:12:05.110 basis. So on that, Substack was a bit less interesting for us. 00:12:05.210 --> 00:12:09.010 I think it's more relevant to creators who look to monetize their content. 00:12:09.730 --> 00:12:11.990 And that's not our business model. 00:12:13.201 --> 00:12:16.521 Okay, so let's break down the business model now you've led us there. 00:12:16.841 --> 00:12:18.341 There's a bit of a red wall. 00:12:18.701 --> 00:12:22.761 There's a registration wall when I go to certain articles, but there's no paywall. I've never asked. 00:12:23.641 --> 00:12:26.321 I've read lots of articles on your website. I've never sent you a penny. 00:12:26.421 --> 00:12:30.741 I'm very sorry. But what is the monetization then? 00:12:31.261 --> 00:12:36.061 Are we talking adverts? Are we talking, you know, obviously, 00:12:36.081 --> 00:12:38.421 I know you do events and conferences and training. 00:12:38.741 --> 00:12:43.401 Is that the key to the monetization then? Yeah, so all of those, as you said. 00:12:43.521 --> 00:12:48.021 So we want to keep on providing free content because we believe that it's absolutely 00:12:48.021 --> 00:12:52.241 necessary for people who not only don't want us, but who can't send us a penny 00:12:52.241 --> 00:12:56.501 because they are the ones who need to innovate and build the trustworthy journalism. 00:12:56.501 --> 00:13:01.281 But, you know, maybe they don't have spare cash to send us. So the main monetization 00:13:01.281 --> 00:13:03.961 is events, as you alluded, so it's the NewsRoyed conference. 00:13:04.821 --> 00:13:08.841 There's both ticket sales and sponsorships and advertising. 00:13:09.241 --> 00:13:13.681 In terms of advertising, we have JobSport that has gone a little bit quieter, 00:13:13.901 --> 00:13:16.641 but it's picking up, which we are extremely happy about. 00:13:16.821 --> 00:13:21.401 There is training courses, so we both provide both open training courses that 00:13:21.401 --> 00:13:25.621 you can sign up for, and we also provide bespoke training courses for companies 00:13:25.621 --> 00:13:26.861 who want to train their teams. 00:13:28.381 --> 00:13:32.221 That's going kind of behind the scenes. 00:13:32.461 --> 00:13:35.621 And we have a freelance directory as well where you can advertise your... 00:13:35.621 --> 00:13:38.881 Yeah, I've just fired that up on my other screen. It's a very clever idea. 00:13:39.861 --> 00:13:43.721 And we are building more directories. And essentially what we try to do is to 00:13:43.721 --> 00:13:50.081 help people discover more people and more products to make the better decisions 00:13:50.081 --> 00:13:51.581 about how to run their newsrooms. 00:13:52.910 --> 00:13:56.430 And so where does the editorial fit into all of that? 00:13:56.850 --> 00:13:59.790 So editorial, really, it's core of our mission. 00:14:00.070 --> 00:14:03.990 We really want to inform people, we want to empower them to make the right decisions 00:14:03.990 --> 00:14:07.950 or get an edge on their competitors, just get intel. 00:14:08.450 --> 00:14:12.190 Because journalists, the journalists are clever. I've never met anything like a stupid journalist. 00:14:12.610 --> 00:14:18.110 I mean, they're ambitious, well, maybe in other areas. 00:14:18.370 --> 00:14:23.790 Are you sure about that? You know, by and large, let's say by and large. Okay. 00:14:24.750 --> 00:14:28.350 Clever, ambitious people, they want to have an impact. They want to build their 00:14:28.350 --> 00:14:31.010 operation. They want to keep on producing good content. 00:14:31.270 --> 00:14:34.950 But they often, you know, you don't really have these conversations in the newsroom. 00:14:35.130 --> 00:14:39.010 You will talk about stories, but you're not going to talk about business model 00:14:39.010 --> 00:14:41.330 and discoverability and distribution. 00:14:42.030 --> 00:14:47.450 They just know the things. And we want this conversation to take place in newsrooms 00:14:47.450 --> 00:14:53.790 because if you don't have sustainable operations you don't have sustainable journalism and you know, 00:14:54.900 --> 00:14:57.840 You know, we sort of recently put our vision and mission out. 00:14:58.960 --> 00:15:03.860 We don't want, you know, we don't accept the future without impactful journalism 00:15:03.860 --> 00:15:06.560 because our industry refused to evolve. 00:15:07.480 --> 00:15:13.300 And this is what I see, this resistance to change and resistance to evolution. 00:15:13.540 --> 00:15:18.380 And, you know, stories are one of the ways to help drive that impact. 00:15:18.940 --> 00:15:22.460 We do want people to have this conversation. We want them to try new things. 00:15:22.460 --> 00:15:28.160 We want them to think about impact and distribution and why what they do matters 00:15:28.160 --> 00:15:30.280 and why they're doing what they're doing. 00:15:30.900 --> 00:15:34.100 So editorial sits squarely in the middle of our mission. 00:15:34.320 --> 00:15:40.680 Now, in terms of monetization, for now, we can provide this content for free. 00:15:40.820 --> 00:15:43.240 As you say, it's a registration wall as of today, actually. 00:15:43.860 --> 00:15:48.620 Right, yes. As of today. So we do want to have registered users. 00:15:48.620 --> 00:15:52.300 As I said, we are 100% independent, so we do not share data. 00:15:52.460 --> 00:15:54.940 With anyone, no advertisers, no external company. 00:15:55.480 --> 00:15:59.780 Really, we just want to know who reads us, how we can connect with people, 00:15:59.960 --> 00:16:01.260 how we can better serve their needs. 00:16:02.282 --> 00:16:05.942 And the reason why we try to increase the number of, or why we try, 00:16:06.282 --> 00:16:11.182 why our number one KPI is registered users is that we want to have the direct 00:16:11.182 --> 00:16:12.562 relationships with people. 00:16:12.682 --> 00:16:18.822 So we want to bypass all of the platforms and truly talk to people as them being. 00:16:19.002 --> 00:16:20.102 Forgive me for interrupting. 00:16:20.442 --> 00:16:24.302 And I think this is the kind of point where we need to take a step back. 00:16:24.302 --> 00:16:27.122 Catching you're the perfect person to discuss it with because not only 00:16:27.122 --> 00:16:30.462 do you want an editorial product you talk 00:16:30.462 --> 00:16:33.282 to lots of other people that run editorial products you want 00:16:33.282 --> 00:16:37.362 a conference where people talk about all these editorial products everyone 00:16:37.362 --> 00:16:40.882 says now look you have to have this direct relationship with 00:16:40.882 --> 00:16:43.702 readers you know i do that please sign 00:16:43.702 --> 00:16:47.402 up to the newsletter if you haven't already newsletter.the edition.net but 00:16:47.402 --> 00:16:50.522 we all have to set that back to that do you 00:16:50.522 --> 00:16:53.322 think that really is the only solution now or do you think people 00:16:53.322 --> 00:16:56.162 are actually getting fed up with always being asked to send over their email 00:16:56.162 --> 00:17:00.622 address probably both i think that's 00:17:00.622 --> 00:17:05.542 right you know i mean it's it's not like it's not some alien concept like you 00:17:05.542 --> 00:17:09.502 know there's always people were always buying you know when people talk about 00:17:09.502 --> 00:17:13.802 paywall it's like oh pay what is this new concept isn't there has always been 00:17:13.802 --> 00:17:17.102 a paywall it was called newsstand it was called going to the shop Yeah. 00:17:17.762 --> 00:17:21.642 And you were literally, you know, giving your money to the person and the new 00:17:21.642 --> 00:17:22.882 skills can get in your newspaper. 00:17:23.542 --> 00:17:29.002 Nothing, none of this is new. And, you know, you always had newspaper would 00:17:29.002 --> 00:17:34.042 talk to readers by the very fact the readers had a paper in their hands. 00:17:34.502 --> 00:17:39.002 And there was an advert, you had call to action, you had advertising classifieds, 00:17:39.262 --> 00:17:45.382 you know, questions, editors letter, whatever, you know, there was always some 00:17:45.382 --> 00:17:46.622 degree of reader engagement. 00:17:46.622 --> 00:17:50.702 I mean, although there need to be, obviously, should have been much more. 00:17:50.842 --> 00:17:53.462 That is probably a different conversation about how we made journalism. 00:17:54.702 --> 00:17:58.182 I think I've said this on shows before, and you're a really good person to ask about this. 00:17:58.342 --> 00:18:04.802 I think as journalists, we're really bad at asking for money for our work. 00:18:04.922 --> 00:18:10.402 And we sort of actually feel bad about it. I think maybe the sort of business 00:18:10.402 --> 00:18:13.762 people in the media are okay at doing it. 00:18:13.902 --> 00:18:19.282 But us, the people that write the stuff, are really bad at asking for money 00:18:19.282 --> 00:18:21.582 and are a bit embarrassed when we do it. 00:18:21.682 --> 00:18:25.782 It's uncomfortable. And I think absolutely everyone is uncomfortable with this, 00:18:25.902 --> 00:18:30.002 you know, discussing money as anyone who goes to a job interview. Right. 00:18:30.802 --> 00:18:35.242 Yeah, the bit they hate is when they ask for, yeah. No one likes to talk about, 00:18:35.302 --> 00:18:38.702 but, you know, it's a valid question and obviously you can't. 00:18:39.957 --> 00:18:44.217 I think it depends what you ask for. If you say, you know, I can't do this unless 00:18:44.217 --> 00:18:46.637 I get paid, I think people get it. 00:18:46.757 --> 00:18:50.337 And from what I know, people are happy to pay for things they value. 00:18:50.597 --> 00:18:54.657 And if anything, once you pay for something, you care more because you feel 00:18:54.657 --> 00:18:57.517 invested, you spend. I do agree with that. 00:18:57.997 --> 00:19:03.657 So, you know, there is no reason why journalists shouldn't be asking to pay for their content. 00:19:03.957 --> 00:19:07.537 And again, it depends on the kind of business model you have. 00:19:07.597 --> 00:19:12.857 But, you know, we do run all these other revenue streams so we can. 00:19:13.537 --> 00:19:17.897 If you want, it's almost like our content is a shopping window for all the other 00:19:17.897 --> 00:19:22.977 things we do, because you can see the expertise we have and access to experts that we have. 00:19:23.157 --> 00:19:26.397 So then if I tell you, welcome to a conference, because this is the quality 00:19:26.397 --> 00:19:32.217 of that times 10, it's much easier to say, okay, well, I can understand why I need to come. Yeah. 00:19:32.597 --> 00:19:35.437 For you, it's that terrible expression, top of the funnel, isn't it? 00:19:36.337 --> 00:19:40.797 So, yeah. I'm sorry to say, use that expression. Right, you've got away with 00:19:40.797 --> 00:19:44.037 it for 20 minutes. Let's talk about AI. 00:19:45.617 --> 00:19:50.437 Because, I mean, when I've been to News Rewired, it's the dominant topic. 00:19:50.777 --> 00:19:54.157 You've had some fantastic speakers on it. I mean, I have to tell you, 00:19:54.257 --> 00:19:59.397 one of the most inspiring talks I've been to over the last couple of years was at News Rewired. 00:19:59.617 --> 00:20:04.257 And I forget the gentleman's name who gave it, but he was very blunt and basically 00:20:04.257 --> 00:20:09.837 told us all to forget about AI and go back to doing some proper journalism was 00:20:09.837 --> 00:20:13.657 actually the most uplifting thing was just like, no, AI is not going to kill you. 00:20:13.737 --> 00:20:16.897 And I left that really inspired. 00:20:16.957 --> 00:20:20.317 But also, of course, we know as journalists that there are, and I do it, 00:20:20.397 --> 00:20:24.717 there are tools I use for all sorts of things that are really helpful that are AI tools. 00:20:24.877 --> 00:20:29.277 One thing I've noticed with you at the New Journalism UK site is that a few 00:20:29.277 --> 00:20:32.137 of your articles, most of the articles now end with the line, 00:20:32.557 --> 00:20:37.957 this article was drafted by an AI assistant before it was edited by a human. Explain that to me. 00:20:39.210 --> 00:20:43.170 So we have two main tools. 00:20:43.430 --> 00:20:51.090 One is called No One, and it's an AI tool that is trained on only our own proprietary data. 00:20:51.270 --> 00:20:54.730 So again, we kind of come back to this sort of independence and a closed loop. 00:20:54.730 --> 00:20:57.590 So it doesn't get all the crap from the internet. 00:20:57.790 --> 00:21:03.790 It's trained only on us. So you can build the kind of knowledge about your company, about what you do. 00:21:03.990 --> 00:21:07.610 It's the kind of unique source of truth. and then when you ask it things, 00:21:07.750 --> 00:21:11.290 it only consider it from the point of view of what you put in it. 00:21:11.950 --> 00:21:15.670 And that for us is extremely important because unlike, you know, 00:21:16.210 --> 00:21:20.870 ChatGPT or Claude, clever as they are, once you start to kind of pull everything 00:21:20.870 --> 00:21:24.930 from the internet and obviously there's ethical questions about how these models 00:21:24.930 --> 00:21:28.210 came to be, et cetera, et cetera. And the issue of hallucination. 00:21:28.510 --> 00:21:32.050 And the issue of hallucination and the rest of it. So the tool that we use is 00:21:32.050 --> 00:21:38.810 trained on our data and it only uses our own writing. It's been trained on our articles, et cetera. 00:21:39.530 --> 00:21:43.630 So we do the interviews. So for example, I would interview you. 00:21:43.810 --> 00:21:46.790 I would break it all down. I take notes. So the way I do an interview, 00:21:46.970 --> 00:21:49.650 I take notes. I don't like to record unless I have to. 00:21:50.650 --> 00:21:53.510 And then I feed notes to no one. 00:21:54.490 --> 00:21:57.510 And it kind of writes it in full sentences. 00:21:57.710 --> 00:22:00.650 I mean, you know what the notes look like. It essentially makes sense, 00:22:00.850 --> 00:22:04.890 kind of cleans up the notes. You want to see the notes I make from interviews is scribbled. 00:22:05.210 --> 00:22:09.190 But then I train in on my style of writing. Right. 00:22:09.810 --> 00:22:13.150 With some expression I use, the type of sentence structure I use, 00:22:13.610 --> 00:22:16.170 just plain language, nor jargon, and the rest of it. 00:22:16.650 --> 00:22:21.430 And then I go through it. I say, okay, so now I have four sentences that make some degree of sense. 00:22:22.270 --> 00:22:24.630 You know, it's still, I mean, machines still write like machines. 00:22:25.190 --> 00:22:29.310 You know, you go through the sentences and change it. But once you have the 00:22:29.310 --> 00:22:34.550 first draft, it's so much faster to fix it rather than try to rewrite the whole 00:22:34.550 --> 00:22:35.670 article from, you know... 00:22:37.004 --> 00:22:43.144 That's really interesting. So I was wondering about this in terms of, 00:22:43.364 --> 00:22:47.044 one, the tight speed, and two, the quality, because those are the two things 00:22:47.044 --> 00:22:49.524 that are the big debating points around AI. 00:22:50.844 --> 00:22:54.024 So first of all, you seem quite comfortable with the quality, 00:22:54.264 --> 00:22:58.304 because you can go back and make it sound human. 00:22:59.244 --> 00:23:03.644 Do you get more articles out? Do you publish more because you're doing stuff 00:23:03.644 --> 00:23:06.384 this way? So the answer is no. 00:23:08.026 --> 00:23:16.446 And I can tell you, AI or no AI, I've never heard a single person crying for more content. 00:23:16.766 --> 00:23:19.666 I've never heard a single person. If only there were more articles published, 00:23:19.766 --> 00:23:24.706 I would just love it. Like, why don't I get 10 newsletters instead of 75? 00:23:25.006 --> 00:23:27.606 I can think of some major publishers that don't agree with you, 00:23:27.726 --> 00:23:30.566 who seem to be churning stuff out. Right. 00:23:31.106 --> 00:23:36.746 No, no, no one needs it. And the point has never been to churn out more content. 00:23:37.006 --> 00:23:43.306 Okay. There can be a case for some articles we would otherwise perhaps not be 00:23:43.306 --> 00:23:46.286 doing because it would be taking too much time and now we can do it. 00:23:46.346 --> 00:23:52.506 For example, the humans in journalism, you know, it's not often is not the most 00:23:52.506 --> 00:23:56.746 engaged content, but people who read this section about sort of culture and 00:23:56.746 --> 00:23:59.206 people in journalism really appreciated content. 00:23:59.906 --> 00:24:03.606 So there is not maybe the crucial give me the edge, help me, 00:24:03.846 --> 00:24:08.106 educate me type content where people primarily come to our website to find, 00:24:08.306 --> 00:24:10.126 but it's a feel-good content. 00:24:10.266 --> 00:24:15.446 And I think we need to celebrate as well successes on the industry in history. 00:24:16.086 --> 00:24:21.226 For sure. How many times have I moaned to you about the sort of acceptance of managed decline? 00:24:21.386 --> 00:24:26.666 I think if you look through our WhatsApp chat, it's me moaning the whole time. 00:24:26.806 --> 00:24:32.426 Why is everyone so miserable? So, you know, we can certainly do more variety 00:24:32.426 --> 00:24:35.286 of articles, even if it's like, okay, maybe this is not a core, 00:24:35.366 --> 00:24:37.706 core thing, but it's important. It's a good content. 00:24:38.086 --> 00:24:41.926 But the aim has never been to produce more content. 00:24:42.146 --> 00:24:46.646 Rather, it was to produce it faster so we can actually do lots of other work. 00:24:46.886 --> 00:24:48.326 You know, we are three people. 00:24:48.766 --> 00:24:52.626 Like before the pandemic, we were eight. Wow. 00:24:53.857 --> 00:24:57.137 And nothing really substantially stopped. 00:24:57.737 --> 00:25:03.397 Wow. Let's put it that way. I mean, some things stopped or at least got slowed 00:25:03.397 --> 00:25:05.917 down or got reduced a little bit. It's one or two things. 00:25:06.297 --> 00:25:09.877 But for three people to run two big conferences, 00:25:10.637 --> 00:25:14.477 training, various directory, big spoke training as well, 00:25:15.237 --> 00:25:19.697 sports, sales, all the audience engagement, all the audience research we are 00:25:19.697 --> 00:25:23.897 doing, going to events, We speak in podcasts, go to conferences, 00:25:24.177 --> 00:25:27.297 be involved in our readers' lives, provide mentoring. 00:25:27.977 --> 00:25:34.557 The amount of work we do is absolutely insane. And I step up and I need to do 00:25:34.557 --> 00:25:39.257 financial planning and non-journalistic things. Jacob is community editor. 00:25:39.617 --> 00:25:42.637 And yeah, I mean, I have a three-year-old as well. 00:25:42.837 --> 00:25:49.617 So running this operation without some help would be doable. 00:25:50.337 --> 00:25:54.637 So AI helps us to do work, but a point has never been to, you know, 00:25:54.717 --> 00:25:59.297 if we wanted to, we could probably turn out more. I don't see the demand. 00:26:00.397 --> 00:26:05.677 That's really interesting. And are you positive or negative on the idea that 00:26:05.677 --> 00:26:08.477 more sites will operate in this way? 00:26:08.857 --> 00:26:11.357 There'll be lots of journalists listening to this show going, 00:26:11.897 --> 00:26:14.877 my God, Marcella, what are you doing? You're going to cost journalists jobs. 00:26:15.137 --> 00:26:18.397 Why are you showing that we need human writing? 00:26:19.117 --> 00:26:22.757 Are you why are you positive about doing it this way hmm. 00:26:24.177 --> 00:26:28.057 You know, I'm not really sure I have a very intelligent answer to this question. 00:26:28.397 --> 00:26:30.577 I don't know where the industry is going. 00:26:31.037 --> 00:26:37.717 I can tell you this year with how the finances work and with the replatforming 00:26:37.717 --> 00:26:42.477 and rebranding and the rest of it, we couldn't have hired an author. 00:26:42.757 --> 00:26:47.577 Just it either, you know, the question was not hiring or not hiring a human. 00:26:47.817 --> 00:26:50.777 The problem was do it this way or not do it at all. 00:26:51.737 --> 00:26:56.277 But you understand my wider question do 00:26:56.277 --> 00:26:59.257 you think more news i mean look i've done stories about 00:26:59.257 --> 00:27:02.597 how reach use ai other newsrooms 00:27:02.597 --> 00:27:05.897 are using it for different things but do you understand the 00:27:05.897 --> 00:27:09.317 sort of core of my question and what people will be worrying about is 00:27:09.317 --> 00:27:12.117 the industry shrinking essentially this is the question or is 00:27:12.117 --> 00:27:15.277 going to shrink to a certain i mean yeah ultimately that 00:27:15.277 --> 00:27:18.437 is what it comes down to but should we 00:27:18.437 --> 00:27:21.637 or is it productive for 00:27:21.637 --> 00:27:26.997 journalists or positive for journalists to be creating articles on the basis 00:27:26.997 --> 00:27:33.197 of ai and you know i know that you're spending time making it human making sure 00:27:33.197 --> 00:27:38.217 it's accurate that's fine do you think this is a positive way forward for journalism 00:27:38.217 --> 00:27:41.277 to produce articles in this way hmm. 00:27:42.361 --> 00:27:46.641 It's a philosophical question. Yeah, it is a bit, yeah. It's the kind of thing 00:27:46.641 --> 00:27:48.581 we discuss at News Rewind, right? 00:27:49.161 --> 00:27:53.561 One thing that AI can't do is actually go and source these stories. 00:27:53.961 --> 00:27:58.121 Like AI doesn't understand things. It's not human in that case. 00:27:58.561 --> 00:28:03.381 So journalist's job is to go and source these stories and understand them and 00:28:03.381 --> 00:28:07.981 have enough knowledge to call people up on their bullshit. You know when stories are emotional. 00:28:08.261 --> 00:28:11.481 You know when you need to be hard-hitting. You know which quote makes sense. 00:28:11.481 --> 00:28:13.241 I cannot make these decisions. 00:28:14.041 --> 00:28:17.061 And I think journalism jobs, like probably all the jobs really, 00:28:17.361 --> 00:28:22.201 are now more about judgment and less about tasks. 00:28:22.601 --> 00:28:29.221 I think if people understand journalism job as task of sitting down and transcribing 00:28:29.221 --> 00:28:31.881 an interview, most journalists started like this. 00:28:31.981 --> 00:28:35.601 We went to newsrooms. We've been given an interview by some senior journalists 00:28:35.601 --> 00:28:39.641 and spent hours and hours typing. Now we put it into Otter or a similar thing. 00:28:39.741 --> 00:28:43.901 You know, and I've never seen anyone saying, oh, but we should be all still 00:28:43.901 --> 00:28:46.541 typing our interviews because AI is doing it. 00:28:46.781 --> 00:28:50.721 Well, guess what? You get those notes and you can actually do the two and the same. 00:28:50.801 --> 00:28:57.101 I think if anything, you can do much better stories because you get these tasks out of the way. 00:28:57.221 --> 00:29:00.261 There is certainly a case, and I've made it to a certain extent, 00:29:00.441 --> 00:29:05.121 that some AI tools free you up to do the bigger investigative things. 00:29:05.121 --> 00:29:07.961 I think there's a really strong case for that 00:29:07.961 --> 00:29:10.701 and what I really think the 00:29:10.701 --> 00:29:14.401 truth is what I really think is it's about finding the balance right so getting 00:29:14.401 --> 00:29:17.961 an AI to transcribe your interview 00:29:17.961 --> 00:29:21.761 which basically every journalist does now is a very positive thing because it 00:29:21.761 --> 00:29:24.901 speeds time you're always going to listen back to check the quote you use is 00:29:24.901 --> 00:29:29.081 accurate I think that's an excellent people will worry I think about whether 00:29:29.081 --> 00:29:34.721 starting a blank page based on AI is positive or not, 00:29:35.941 --> 00:29:41.821 But I totally accept that you as a human can then draw out the bits of the story you want. 00:29:43.172 --> 00:29:47.392 I mean, you know, AI is a tool. It's not a master. 00:29:47.812 --> 00:29:52.152 Correct. And I think it becomes problematic when people start to put a bit too much faith in it. 00:29:52.492 --> 00:29:57.712 You know, I would never ask a tool, create me a story on X. 00:29:58.432 --> 00:30:03.132 You know, you use a tool for something completely outlandish. 00:30:03.332 --> 00:30:06.552 Like that is not what it's created for. It doesn't think, you know, 00:30:06.592 --> 00:30:10.412 it's a very clever way to predict words, how they will come together. 00:30:10.732 --> 00:30:15.892 That is all there is. so it can't think and you know every so often when there 00:30:15.892 --> 00:30:18.652 was loads of interview notes and i say now it needs to be shortened i was like 00:30:18.652 --> 00:30:21.952 well wait a minute you completely you know this is that bit is important you 00:30:21.952 --> 00:30:23.952 can't just like get rid of it no no no but. 00:30:25.372 --> 00:30:29.652 Human journalistic instinct obviously and you know we i mean we are a trade 00:30:29.652 --> 00:30:35.452 publication as well so you know the kind of stories we cover uh i'm not deep 00:30:35.452 --> 00:30:38.192 you know six months investigation into and do something. 00:30:38.372 --> 00:30:45.572 But when you are a journalist, those stories are, no AI will ever do them. 00:30:45.772 --> 00:30:48.592 No person will talk to an AI over you. 00:30:49.932 --> 00:30:53.792 No, and I feel really strong about that. And we're going to end on a positive 00:30:53.792 --> 00:30:57.952 because you make sure we do it on News Rewind. That's what I'm going to do on this show. 00:30:58.172 --> 00:31:02.632 Because I really strongly believe that. I think we're going to learn the human 00:31:02.632 --> 00:31:05.992 value or what the value that humans bring to journalism. 00:31:06.212 --> 00:31:11.212 I really do believe that. So yes, using Gemini to do a brief bio of your interview 00:31:11.212 --> 00:31:15.452 subject or I use Riverside for this podcast and it will edit the podcast for 00:31:15.452 --> 00:31:18.272 me and that saves me time. All of that is really positive. 00:31:18.532 --> 00:31:24.312 I do think that is different to an AI drafting an article, but I also accept 00:31:24.312 --> 00:31:28.772 that you would never just hit publish on what the AI has produced. 00:31:30.245 --> 00:31:34.145 Aren't. Because you can't. Why would you? And you know, it's still, 00:31:34.305 --> 00:31:41.845 if you read through it, it's 85% my note writing just in full sentences. Yeah. Yeah. 00:31:42.345 --> 00:31:45.285 It can't create something that I haven't written in it. 00:31:45.665 --> 00:31:50.305 Yeah. It's a really interesting development. So my final question really is, 00:31:50.525 --> 00:31:55.405 you said journalism.uk started in 2009. Let's go 2029. 00:31:55.725 --> 00:31:58.345 Do you think the site you now run will look like? 00:31:59.205 --> 00:32:05.325 So a number of things we are doing. One is a feature called Ask J-UK. 00:32:05.685 --> 00:32:11.225 So we are going to use a large language model for search in our archives. 00:32:11.445 --> 00:32:15.965 So instead of trying to go through, we exported about 100,000 articles in the 00:32:15.965 --> 00:32:18.385 end. And that's a lot of knowledge about journalism. 00:32:18.765 --> 00:32:22.485 It's still in the... So sorry, of the 1 million, you exported 100,000? 00:32:23.145 --> 00:32:25.905 Right. that's right and i mean and that covers i mean 00:32:25.905 --> 00:32:29.065 not just editorial but that was about what we're looking 00:32:29.065 --> 00:32:31.805 at so this you know we are not going 00:32:31.805 --> 00:32:36.065 to sit there and wait for chat gpt or whatever to crawl our websites and then 00:32:36.065 --> 00:32:39.565 you know and then bitch about oh but people go to ask chat gpt instead of asking 00:32:39.565 --> 00:32:44.265 us well you know guess what we want to become the destination site too and you 00:32:44.265 --> 00:32:50.865 know and and naturally a zero click search is clever and it's useful i'd much rather ask. 00:32:51.787 --> 00:32:56.787 Something to give me an answer rather than spending a lot of valuable time trying 00:32:56.787 --> 00:32:59.867 to look for it and then skim the article and try to get the answer out. 00:33:00.207 --> 00:33:04.607 So this model, again, is trained on our closed loop data. 00:33:04.687 --> 00:33:09.587 So it's only our articles that you can search. You can not search external web. 00:33:09.807 --> 00:33:14.847 Then you come to our website and if you wonder about how to use AI in journalism, 00:33:14.847 --> 00:33:17.347 is it ethical? What's the revenue model? 00:33:17.707 --> 00:33:22.847 Is FTE implementing or whatever that can be? You can just ask it and that gives 00:33:22.847 --> 00:33:25.207 you an answer from our reporting. 00:33:25.607 --> 00:33:30.247 And we truly want people to have the opportunity to access this knowledge in 00:33:30.247 --> 00:33:34.987 a way that makes sense to them and just unearth that important information that 00:33:34.987 --> 00:33:36.887 they need in that very moment. 00:33:37.087 --> 00:33:44.267 I will go as far as saying, I'm not sure the future of B2B publishing is in articles. 00:33:45.627 --> 00:33:54.107 Okay. I see a future of producing artifacts that are digestible by something 00:33:54.107 --> 00:33:56.287 like Latin Language Model Search. 00:33:56.747 --> 00:34:01.247 You're coming back on the show in a bit. We'll discuss that next time you're on the show. a nerd. 00:34:01.547 --> 00:34:06.367 But you know, we can, we definitely want to embrace this behavior, 00:34:06.567 --> 00:34:10.147 but just give people quality content, ethically sourced content. 00:34:10.327 --> 00:34:13.907 You know that, you know, people who are doing this content are paid and nobody 00:34:13.907 --> 00:34:15.427 doing their editorial standards. 00:34:15.987 --> 00:34:20.667 This is not, it's instructed not to hallucinate. So if it doesn't know the answer, 00:34:20.747 --> 00:34:23.447 it just says, sorry, I don't have enough information to answer your question. 00:34:23.687 --> 00:34:29.067 And conversely, we can see these inquiries and these for us will be the story 00:34:29.067 --> 00:34:32.847 leads, really, because if I see that you are wondering about X and there is 00:34:32.847 --> 00:34:35.327 not enough content on that, that is our queue. 00:34:35.467 --> 00:34:39.447 We need to get more information on that. So we really want to become that destination 00:34:39.447 --> 00:34:43.107 that will give people, there was a source of information of what people we need. 00:34:43.627 --> 00:34:47.367 And we are doubling up on directories as well. So we want to help people discover. 00:34:48.664 --> 00:34:52.384 And especially tech and consultants, because it's just too much. 00:34:52.564 --> 00:34:54.364 It's too much noise, it's too much products. 00:34:54.584 --> 00:34:57.144 How do you choose? How do you even see through anything? 00:34:57.504 --> 00:35:00.144 And, you know, we just really want to be there to help people, 00:35:00.764 --> 00:35:04.384 people, journalists, newsroom creators make better decisions. 00:35:05.364 --> 00:35:08.724 Fascinating. I look forward to following all the developments and seeing you 00:35:08.724 --> 00:35:10.784 at News Rewired in the not too distant future. 00:35:13.084 --> 00:35:16.344 14th of May. Breathe. It's going to be okay. 00:35:16.684 --> 00:35:20.704 So the 40th May is the next Juicy Rewired. Obviously, you're at journalism.co.uk 00:35:20.704 --> 00:35:23.304 is where we can read all the work that you and your colleagues do. 00:35:23.424 --> 00:35:24.964 Where else could people keep up with you, Marcella? 00:35:25.564 --> 00:35:30.024 Lincoln is probably the best way. I post loads of things that I do not post 00:35:30.024 --> 00:35:33.484 on our platform and subscribe to our newsletter. 00:35:33.824 --> 00:35:37.864 That's where the most action really takes place. I mean, our website is just articles. 00:35:38.424 --> 00:35:43.784 Newsletter is where the real juicy bits are published and they're not on the website and it's free. 00:35:44.564 --> 00:35:49.244 Wonderful. I'm Charlotte Henry and I'm at Charlotte A. Henry across social media. 00:35:49.464 --> 00:35:54.484 You can head to theedition.net, which is A-double-D, to sign up to my newsletter, 00:35:55.004 --> 00:36:00.224 read different articles, watch some videos, all sorts of good stuff going on there. 00:36:00.444 --> 00:36:05.024 My latest book, Streaming Wars, is out now and you can buy that wherever you buy books. 00:36:05.704 --> 00:36:09.104 Thank you again, Marcella, for joining the show and I'll see you next week. 00:36:12.004 --> 00:36:12.084 Yeah.