Claudiu Lodromanean joins Joshua Goode to discuss how Newspack has revolutionized WordPress hosting for news organizations. Starting with grant funding from Google News Initiative and other foundations, Newspack has grown to serve over 330 independent publishers worldwide, providing not just hosting but a complete technical partnership that lets newsrooms focus on journalism instead of technology.
The technical challenges of news publishing create unique hosting requirements that standard WordPress solutions can’t address. With databases containing hundreds of thousands of articles, complex subscription systems, and the need to handle sudden traffic surges during breaking news, Newspack has developed sophisticated caching strategies and performance optimizations. Claudiu explains how they use Elasticsearch, strategic cache timing, and custom bot management to keep news sites running smoothly.
This episode provides valuable insights for anyone involved in WordPress hosting, particularly those dealing with content-heavy sites or unpredictable traffic patterns. Learn about the balance between performance and content freshness, the growing impact of AI crawlers on hosting infrastructure, and how open-source principles can coexist with managed hosting services in the modern web ecosystem.
Links:
- Newspack
- Digital Trends
- WordPress VIP
- Google News Initiative
- Knight Foundation
- Bangor Daily News
- Elasticsearch
- Jetpack
Transcript
## 124.1 – Claudiu Lodromanean (Newspack)
#### Introduction and WordPress Journey
Claudiu Lodromanean: These days it’s pretty standard. Almost all news organizations use WordPress because it’s so easy to customize and to find staff that already know how to work within WordPress.
Joshua Goode: You’re solving a lot for the technical hurdles. So to me, Newspack is the hosting, the service and everything that you provide. But you all do open source most of the stuff, right? A lot of the at least how you manage some of the hurdles there that organizations have, the technical challenges that we’ve briefly mentioned.
Claudiu Lodromanean: We’re always rolling out new features for Newspack and improving the software. We are providing really good support. You know, if someone’s like, hey, how do I accomplish this thing? Here’s how. We have a Slack community that all of the publishers are in.
Joshua Goode: Welcome to Impressive Hosting with Jesse Friedman. I am not Jesse Friedman. I’m Joshua Goode. And at the time of this recording, Jesse is on a sabbatical and graciously asked me to take the reins of the show while he’s away. So for the next however many episodes, I’ll be attempting to fill some impressive shoes. If this is your first time listening to the show, we’re dedicated to uncovering the core tenets of what makes great, and dare we say, impressive WordPress hosting as well as the ecosystem around it. Disclaimer: Jesse and I both work on WP Cloud. That’s Automattic’s WordPress-first cloud hosting platform. I personally serve as a technical solutions engineer, and I’m embedded in our systems engineering and operations team. I act as a bridge between our partnership development team and our partners. But don’t think of this as a big ad for WP Cloud. We explore many different topics and talk with people across the hosting ecosystem, whether they’re using WP Cloud or not. Before we dive in, be sure to check out impressive host. That’s where you can find and comment on previous episodes. Ask follow-up questions, submit questions for future guests, and submit questions, links to like, follow and subscribe wherever you listen. My guest is Claudiu Lodromanean. He’s going to fill us in on what’s going on over at Newspack, the platform, and let me say the word one more time, an impressive platform for news empowering news orgs by giving them the tools and service to thrive. Claudiu has been with Newspack since the beginning and has helped architect, develop, launch, and scale the platform, so he should be the perfect person to fill us in on how they deliver impressive hosting. Claudiu, welcome. Thank you for joining us. It’s awesome to have you. I’ll give a moment and then I’ll finally stop talking for a moment. I have worked alongside you. We’ve seen each other in chat. We’ve looked at issues together, but I’ll be honest, we haven’t been able to sit down and have a conversation like this, so it’s very great that you could join me today.
Claudiu Lodromanean: Hey, Josh. Yes, thanks for having me. I’m really excited to be here. Good to finally speak to you in person after, you know, years of Slack.
Joshua Goode: Yes, it’s been a lot. So I always like to start things out a bit on just learning. First and foremost, big reason why we’re here. We all do different things with it, but I like to know people’s WordPress journey. So I know I think you joined Automattic in 2017, but I have no idea about any of your kind of WordPress background and stuff before that. So why don’t you fill us in a little bit?
Claudiu Lodromanean: Absolutely. Yeah. So I’ve been at Automattic almost nine years now. Before this I was at Digital Trends, which is a large tech news company. They have a few different properties. Digital Trends is the main one. Kind of like a CNET or a Verge. Before that, I think the first time I used WordPress would’ve been maybe the late 2000s, when I was actually working at an upholstery shop and their site was on WordPress and I got stuck with the task of managing their web presence. And you know, it was super easy to use. Much better than, you know, how I had been doing it way back when. You know, like FrontPage or writing manual HTML. Big step up.
Joshua Goode: It was a big shift for me too, going from static site creation and stuff and just having a pretty decent at the time, even the early days, I would say it’s a great CMS, but you know, even at the time, we didn’t have a lot of good options and having something that’s open source that the community could work on that was also delivering something, you know, quite powerful to democratize publishing. It was a pretty cool little flip there. So Digital Trends, that’s, so you did some of the engineering on their publishing platform?
Claudiu Lodromanean: Yeah, exactly. It was WordPress at the time. It still is WordPress right now. It’s on WordPress VIP.
Joshua Goode: Nice. So yeah, so that’s a pretty cool tie.
Claudiu Lodromanean: Yeah, at the time it was sort of before WordPress, nowadays, it’s kind of the standard for online publishing. So yeah, Digital Trends was one of the biggest WordPress sites at the time. These days it’s pretty standard. Almost all news organizations use WordPress because it’s so easy to customize and to find staff that already know how to work within WordPress. But at the time it was much more niche.
#### What is Newspack
Joshua Goode: For sure. So great. That really kind of gets you to an earlier, you know, experience with some early WordPress as a publishing platform. So I guess it’s perfect transition to kind of take us over to what is Newspack. I have a little bit of a personal experience and knowledge of it, but for the greater audience out there, why don’t you start breaking down just what is Newspack.
Claudiu Lodromanean: So Newspack is essentially a complete news site. It’s a total package. You get all of the special software and features that news sites need. Workflow stuff, layout, advertising, calls to action like subscriptions, memberships. Basically anything that you see on a news site as you’re out reading the news online is included in Newspack. We have support. We have really strong technical support. Most of our publishers don’t have any sort of technical support on staff or developers on staff. So we are their technical partner and the idea is we handle all the technical heavy lifting, all the software, updates, hosting, everything, and they can focus on the news stories.
Joshua Goode: Yeah. Yeah. So I guess that kind of, I was thinking what were the problems that you were trying to solve when you were first conceiving, the team was first conceiving Newspack. So I guess that’s a big chunk is just a lot of the technical burden that comes with trying to operate something like that. And obviously there’s many other challenges, and I’m sure we’ll get into some of those. But are there any other notable things, notable issues, problems that you were trying to solve that first said, okay, these news organizations really need something different than just standard WordPress hosting.
Claudiu Lodromanean: Yeah, so one of the big impetuses way back in the day when we were first doing Newspack was these organizations like Google News Initiative, the Knight Foundation, that help journalistic organizations. They were finding that all of their, basically their grant money, or their help was not as effective because everyone had a different tech stack. So the first thing you know everyone has to do is get onto a modern tech stack that can do the best practices. So our goal was to provide that tech stack for organizations to really help them succeed, be able to implement the best practices of the news industry.
Joshua Goode: Gotcha. So, I guess right now, how many news organizations are you working with?
Claudiu Lodromanean: There’s more than 330 at this point.
Joshua Goode: Oh wow. Great. Great. And these are just independent news outlets across the world? Or are you kind of focused in certain regions?
Claudiu Lodromanean: Yeah, yeah. No, it’s independent news outlets. It’s not the sort of hedge fund owned giant conglomerates. They’re all independent. They range from hundred year old ones like Bangor Daily News to fresh startups. Digital, print, probably about 75% in the US, 25% around the world. About half of them are non-profit. So it is a sort of huge spread of organizations.
#### Open Source vs Service Model
Joshua Goode: Nice. So obviously, you know to me and I’d love to dig into these a little bit more. You know, you’re bringing a lot, you’re solving a lot for the technical hurdles. So to me, Newspack is the hosting, the service and everything that you provide. But you all do open source most of the stuff, right? A lot of the, at least the how you manage some of the hurdles there that organizations have, the technical challenges that we’ve briefly mentioned. I’d love to dig into some more of those here in a moment. But to me, when I hear Newspack and what you’ve described, that’s, you know, the kind of the software stack that you’ve offered, the specific usage of WordPress plus the service, the hosting, the delivery that you’re offering, combining, but I see that, you know, Newspack actually has a lot, most of its stuff open source. So you know what would drive a news organization to, you know, go with the full Newspack service rather than going, you know, taking what you’ve built that is open source and trying to use that somewhere else.
Claudiu Lodromanean: So, for context, our open source roots go way back to our initial grant funding. So Newspack was started partly with a grant from the Google News Foundation and a collection of other sort of grant funders looking to help the news industry. And as part of this, they wanted to make sure that whatever they were funding, we couldn’t just lock them into some terrible service and jack up the rates on the organizations and really milk them. So the open source kind of keeps us honest in a sense that, you know, at any point. Newspack can run on any WordPress site. So if a publisher feels that they’re not getting good value from what we deliver, or we’re charging much more than we’re worth, they can, they’re free to leave. They can take it and host it on any WordPress.
Claudiu Lodromanean: So for context, our open source roots go way back to our initial grant funding. NewsPack was started partly with a grant from the Google News Foundation and collection of other grant funders looking to help the news industry. And as part of this, they wanted to make sure that whatever they were funding, we couldn’t just lock them into some terrible service and jack up the rates on the organizations and really milk them. So the open source kind of keeps us honest in a sense that at any point NewsPack can run on any WordPress site. So if a publisher feels that they’re not getting good value from what we deliver, or we’re charging much more than we’re worth, they can leave. They can take it and host it on any WordPress.
Joshua Goode: Got it. So I guess it’s a good point to transition into what are some really technical challenges that you’re able to solve with your specific hosted version of this platform and the services that you provide. Can you just give me a brief walkthrough? We can do a very high level and then we can start getting into details of just like, what all are you doing and solving for by offering that fully hosted version of NewsPack.
Claudiu Lodromanean: Alright, so the hosting itself is impressive. Like that in itself is a really good value on what we charge publishers. We manage it all for them. We make sure that their usage is good. They can hold up to crawlers. The site is performant. We manage all their plan updates. We’re always rolling out new features for NewsPack and improving the software. We are providing really good support. You know, if someone’s like, “Hey, how do I accomplish this thing?” Here’s how. We have a Slack community that all of the publishers are in. So it’s a Slack community of 2000 independent news organizations that collaborate amongst each other.
Joshua Goode: So you have this kind of full-stack solution. Now let’s dig into something a little bit more. So as part of, now I will give disclaimer on here, that NewsPack does leverage WP Cloud for some of those offerings, but they have done, they’ve taken great efforts to really make their own hosting. So while we provide a nice base layer, they’ve taken it off. And this can demonstrate the ability and the power to really customize and take it beyond what we just provide at the most basic level in WP Cloud. So let’s go into some specific scenarios. What happens behind the scenes or what, let’s start from the preparation, what are you doing to help a local news site be able to take a surge in traffic? You know, let’s say a tragic event like a wildfire or just some celebrity marriage announcement or something and their article goes viral. What are some steps that you’re taking preemptively to help address that or help be able to withstand that? And then afterwards, we can dig into like, what are you doing to monitor that? How are you addressing it? If there are some improvements that could be made.
Claudiu Lodromanean: Yeah, so news sites are a little trickier than the average blog or marketing site, for a number of reasons. Primarily it’s the amount of content on there. So some of these publications have hundreds of thousands of articles, and a lot of dynamic features. So like step one, obviously, make sure everything is performant, make sure everything is scalable. When you have a database that has 300,000 articles in it, you can’t on every page load analyze the database to find posts in these three categories with this tag, but not in this tag. Because the SQL query that generates is gonna be so gnarly. Your page is gonna take like 10 seconds to load. So in the background, we offload some of that to Elasticsearch via Jetpack. We cache some of these gnarly queries, different parts of the page. So like as a baseline, if the site is gonna receive any traffic, it needs to be scalable. It needs to be performant, otherwise it’s going to fall over.
Joshua Goode: Yeah, so that must be an interesting challenge. Because you’re trying to provide these sites with the ability to publish news as it comes out. So keeping everything fresh, and obviously they’re retaining their backlog of years and years of news. I see that you also have subscription membership stuff. So a lot of dynamic content going on and you have logged in users, or at least they’re making uncachable requests. So I imagine that caching is probably major focus and it’s gotta be a very careful balance of how you are able to both handle all these dynamic things that need to happen while using cache as much as possible. Is there anything unique or any unique challenges, or maybe your unique approach to how you solve for caching in those situations?
Claudiu Lodromanean: There definitely is. I have spent so much time working on caching, so much. Hard to overstate how much time I’ve spent looking at caching-related stuff. So you know, normal article that goes viral, a million people visit it. Caching works great. We can serve it, especially with WP Cloud and the edge caching and everything. It won’t even hit the server if a million people visit an article and it’s a cachable article. But you start getting into weird parts around, for example, scrapers and crawlers, where if someone is crawling the site. They’re not visiting one article. They’re visiting a million articles from four years ago, which are not gonna be cached. So we’ve tuned the caching in such a way to, like, on these old ones, we can cache it for a year. If an article hasn’t been updated, like why not? There’s no downside to it. On recent stuff, you know, cache for five minutes or even during these wildfire-type or election-type scenarios. There’s another balance there you have to find between, like, it needs to stand up to the traffic, but also people want to get timely information as soon as possible. So you can’t serve them a 15-minute cache because the information will remain cached. They’re not gonna get that timely stuff they’re looking for. So we do like two, three-minute cache on election day.
Joshua Goode: Yeah. So what I’m hearing is a lot of optimizations when and wherever possible, cache what you can. So that way you’re ready for any of these more dynamic or really heavy on database requests that you’re getting. Because ultimately you can’t, if someone wants to subscribe or join and get a premium membership or whatever it may be from one of these news outlets, that’s gonna require some things that are simply not cachable and are going to take up some resources there. So I’m hearing optimize, you know, maybe add some database indexes, optimize the queries, and just make sure caching works wherever you can. So great point on old traffic and keeping old pages and posts cached. I will say we are working on some preloading trying to keep things a little bit more, have a better opportunity for old stuff to be cached. It’s kind of in the works. It’s with our edge operations team that’s dedicated to our entire edge network here at Automattic, but especially WP Cloud. And we are taking some different approaches of what’s too much. Because these preloads do require some visits, just like a scraper going through hitting the origin server at times. So, and especially on old content. So we are tweaking and testing that right now. And I’m hoping that here in the next few months we can get something rolled out. It will mostly be beneficial for medium to new-aged content. Because we’re taking a few different approaches. We’re not simply just taking the site map and going down through all of that. We’re kind of starting, it’s like if we see that something has been updated. We kind of make sure that that cache gets flushed out but it’s immediately edge cached when and wherever possible. So I’ll, but we can talk about that on another day. So, but we see that challenge at the WP Cloud level and definitely in terms of scrapers and traffic spikes. And we do take some steps to try to limit scrapers and bad bots. At the WP Cloud level, we have to play a careful balance. Because on NewsPack you may see very specific scrapers that have a very specific signature that you want to have limited, and yet another client with another set of sites may actually want those to come through. So we have to take a bit of an approach, not too much of a heavy hammer and heavy-handed approach to it, but at the same time, we do have to protect our partners like NewsPack and your partners and your sites. So, I’m looking forward to talking about edge caching a little bit more. We are going to eventually have some domain-specific language opportunities where you’re going to be able to craft the whole set of rules. And you can even do it down to specific sites so that way you, if you’re really seeing some bogus traffic or you really want to do some limitations there, you don’t necessarily have to consult with us as much or suggest things to us. These are gonna be opportunities for you to be able to really shape and control that traffic right at the edge level and either cross your client or at a per-site basis.
Claudiu Lodromanean: Nice. That’s exciting. I’m looking forward to learning more about that when that’s ready. Yeah, I mean, out of the box, the bot blocking that the platform provides is 90 some percent of the time. We do have a couple extra things we’ve built on top of that for news stuff specifically because especially these days, the crawlers are out of control. Like I saw a stat on one of the sites yesterday, it was like 70% of all requests are bot traffic.
Joshua Goode: Yeah. There’s a lot out there and the signatures that we use can vary. So again, we always want to allow all legitimate traffic when possible. And so we have to take some very careful steps of making sure that some legitimate traffic is not for some reason using the same, whether it be user agents and all kinds of other signatures that we take into account. And sometimes that happens. So we are slowly chipping away at it, giving a little bit of inside source knowledge of different ways that we can allow News Pack and anyone else on WP Cloud, you know, balance that themselves based on their users’ needs. That way, we’re not blocking a bunch of stuff that is legitimate too. So, yeah.
Claudiu Lodromanean: Yeah. One of the cool things, if I could jump in, that we rolled out probably a couple of weeks ago is the robots text file. So we rolled out a feature where it will actually block the stuff in robots text. Like if you say, I don’t want ChatGPT crawler to visit the site. It will analyze and block that dynamically by parsing your robots text. And that’s been really cool for providing control around it because different publishers have different thoughts about getting their content scraped by AI. So this really puts the control in their hands.
Joshua Goode: Yeah. Was that the main thing that kind of brought that up? Like AI crawlers?
Claudiu Lodromanean: Yes, AI crawlers, for sure.
Joshua Goode: Yeah.
Claudiu Lodromanean: It’s a very nuanced topic. You know, some of the non-profit people are like, our goal is to get our message out as far as possible. You know, if AI crawlers want to take this and spread it through their LLM, let’s do it. But then other people are like, we, you know, our value is our writing. We need our registration walls and things like that in order to keep these crawlers out.
Joshua Goode: Gotcha. So, I think we have so much we can dig into even more. I don’t want to wrap things up here because we had just touched the surface on things. I think what we’re going to do is do two episodes. We’ll split this off into a two-parter, so for the next, I guess, is there anything else in terms of foundationally, like what is News Pack, who you are, what you do that we want to cover right now? Because I’ve got so much, we just got into the caching and stuff, so there’s so much that is now hitting my brain that I want us to get into. So maybe let’s, this can be, we can tell people to listen to this first and then we can get into some more technical stuff in the next episode. So anything else that you kind of want to cover or close on like right now before we wrap up this episode?
Claudiu Lodromanean: Yeah, I mean, I think to sort of reiterate, you know, we are the publisher’s technical partner. So we manage their site, we keep it online. You know, one of the publishers the other day was saying that like, one of the big things they get from us is that they know that the site is fine, they can focus on the news. They don’t have to worry about like, are there plugin updates I need to apply? Or something like that. Like it just, our goal is to manage the site and provide them the tools. So, you know, we want to make sure that they have the best in class experience, the most optimized calls to actions, abilities to segment readers and all these ways that, you know, previously would only have been available to like the New York Times or the Washington Post. So we take those and we make them available to everyone.
Joshua Goode: Perfect. Perfect. So yeah, fully managed and dare we say, impressive WordPress hosting specifically for news organizations. As we kind of get closer to closing this episode out, I think that’s a good introduction for just a base introduction for what News Pack is, and let’s definitely continue this conversation a bit more in the next episode because like I said, there’s so much I feel that we can really dig into and dissect in terms of what you do specifically. Because I know all kinds of stuff about what we do on WP Cloud, but I’m always interested in seeing what the users of WP Cloud are doing to take it and make themselves unique, serve their audience and their partners, and take it to the next level. So I guess on the next episode. Just to give people a little teaser here. I think we’re going to maybe delve into some of the prep for the 2024 elections. I think the team was preparing many months for that. And then we may get into a little bit of alerting, some monitoring, go back to how they handle traffic spikes in a few different ways. And of course, we may cover a little bit of security. So as one little final question, we’ll probably get your opinion on the same one afterwards, as of today, what is one of the hardest technical problems that you’re working on for News Pack? You don’t have to give me the complete answer now. We can do half it in half. You can just give me that and answer it in the next episode, but what is the hardest technical problem you’re still working on?
Claudiu Lodromanean: These days, I would say the problems are less technical and more product focused. So it’s providing a good reader experience, building relationships with readers, determining how to how news organizations should be using AI or thinking about AI in their workflow and in reader facing stuff.
Joshua Goode: Perfect. So yeah, so you’ve kind of settled. You know, you’re hosting the platform. A lot of your base tooling is all kind of there, and now you’re really able to expand, iterate on both your product and your relationships.
Claudiu Lodromanean: Yeah. Absolutely. If you do need something super, super technical, debugging performance issues and plugins, it’s so hard to find like what hook is making this performance problem.
Joshua Goode: Well, let’s definitely dig into that a little bit more. Well thanks everyone that’s tuned in to this episode of Impressive Hosting with Jesse Friedman. Again, I am not Jesse Friedman. I’m Joshua Goode. Claudiu, thanks again for joining us. We’re going to carry on this conversation in a follow up episode where we can really dig into some of these technical challenges a little bit more, get into some of the debugging and troubleshooting, and then you brought up a good point on one of your biggest challenges is iterating on the product and relationships, so I think that’s important to get into too, so we can take a moment to expand on that. Thanks for joining us. Be sure to check us out on Impressive Host. Again, that’s where you can find all of our links. I will save any other closing statements for Jesse regarding the podcast, but if you want to find me out on the web, for one, you can find me on wp.cloud. But, go dot P-R-O-G-O-O-D-E, dot PRO. How about yourself, Claudiu? Where can people find you?
Claudiu Lodromanean: LinkedIn is the main place. LinkedIn. Just search for my name, Claudiu Lodromanean. There’s no other ones in the world. You’ll find me.
Joshua Goode: Awesome. Well, thanks again for coming on, telling us a little bit about News Pack, and I look forward to digging into that even further.
Claudiu Lodromanean: Thanks for having me.
Joshua Goode: Perfect.





Leave a Reply