Skipping the blame game and working across teams to fix newsroom Ad Tech.
Session facilitator(s): Aram Zucker-Scharff, Jarrod Dicker
Day & Time: Thursday, 12-1pm
Room: Classroom 305
Aram: Hello, everyone, you are in skipping the game blame working across teams to fix newsroom ad tech. Hopefully you’re in the right place. We’re going to quickly go through who we are and some of the challenges that we are facing as newsrooms that need to continue making money so that we can all stay in business. And then we’re going to try to break into some sort of back and forth, some conversations about challenges you face, metrics that you love but can’t seem to turn into money and also opportunities for projects that may be you see or that we can help think of together.
So I am Aram Zucker-Scharff, I work with the Salon media group as a full developer.
Jarrod: I’m Jarrod Dicker, I run all technology operations at the Washington Post.
I’m going to quickly go into the mentality of how we successfully—so just back on. I worked in newsrooms, and I’ve run in technology at the Huffington post, a couple of startups and now Washington Post. It’s never easy. Especially when you’re trying to work across different teams, and you have the mentality that ads suck, they slow the site down, everything’s terrible, and you continue on that route.
So putting this new mentality, we’ve been successful in doing that, especially at the Washington Post. And what I’m going to basically go through is how we change our mind-set when it comes to creating new ad prospects and technology.
So I start with this Jeff Bezos quote often when I talk about things, he owns the Washington Post, which I’m sure everyone in here is familiar with, and this is completely out of context from how I look at it. But what he’s basically saying is everyone is focusing on what’s going to happen and how to get there. But no one really focusing on what’s not going to change and what hasn’t changed. And then that realm, it’s really been ad technology and product in this space. You look back on how we deliver news and content to our consumers, on desktop, or mobile, it’s flexible, it’s responsive, we still hav 350 for as long as anyone can remember. And there’s no building in that realm. Even today when you see what Google’s done with amp where they want it to be fast, they want to bridge fast to the publishers, yet they don’t address the whole side of double-click and how we distribute throughout all of our publisher site. So no one says, hey, let’s fix that foundational layer
while also simultaneously bringing us into the new wave. It was, no, let’s ignore it and keep ignoring it and keep pushing this way.
So the main construct when I’m talking to my newsroom and across the entire word is main we can be the thought generator of pushing into this direction.
What’s been amazing in my field was the introduction of ad blockers, viewability, and these issues that have forced everyone to actually deal with the problem, you know? For years and years and years, we’ve known that these have been key issues that all of our ad servers and the creative and working with clients was a way that we made money, but it was a sacrifice, it slowed the sites down, we were where we needed to get to go, but we had to do it. But it got with the point with ad blocking where if we’re speaking to clients in particular, they don’t need to spend money with us. It was always, hey, how do you get 5% or 10% of this media buy? Well, now it’s if your ad blocking ratings are up, we’re not going to pay for it or spend money with you.
So the whole entire organization, speaking to the Washington Post directly and Aram can speak along that too, now the whole company is, like, okay. Let’s fix this problem.
And we’ve been very lucky, you know? Personally speaking at the post, we’ve actually had a dedicated team that we call red if anyone’s interested on learning more into the postdocs that only focuses on ad technology, working across the newsroom, recognizing the trends, recognizing things that our editors are doing well and how can we package them and deliver them? We’re not working anymore on driving ads and viewers to our site, we’re actually selling technology. 95% of the ad tech is shared, I’m sure there are a lot of publishers in here but the New York Times and Washington journal and Washington Post and buzz feed, we all call our products something different but they’re all powered by the same. So there’s no differentiation. And if we’re in this world, we have to have organic. And that’s what has gotten our teams excited in this space and kind of how we’ve led change throughout the industry as well as working with publishers and with clients alike.
PARTICIPANT: Yeah. So that’s what we’re going to be talking about today. That because the advertising industry is changing, because the Web is in such bad shape, we have an opportunity. Everything is changing, and we can lead that change from the newsroom side. The question that sort of brought this idea forward is that the economy of our media organizations is something that exists around editorial. But editorial is what our companies produce. That’s what we’re about. Making journalism. So we need to ask ourselves, how can advertising value editorial, when it does not share the values of our editorial, and this is what we need to start building for.
So the first question that I’m hoping we can talk with you all about is what editorial values and metrics are not reflected in the advertising that surrounds it?
In preparation for this session, one of the things that we talked about to start us off in this discussion is authors as a value. If you’ve ever had to implement an advertising unit, you know that author may be a tag that gets send out to the admin. But nobody’s building on that. There’s no standardized way accepted across all of these news organizations, across all websites, across all social that will designate the author.
So what would it look like if we built an ad product around authorship? We could understand audiences very differently if we built that as our center of metric. And that’s just one example. I would like to take a moment. This is hopefully going to be a discussion. A lot of us build stuff for newsrooms obviously. And a big piece of that is just building out traffic so that’s our reporters, our editors, our journalists knows what’s being successful, knows what’s going on.
So are there metrics that you here have built or enjoyed looking at that you don’t pass off to the sales team? That the sales team doesn’t know about or doesn’t value? Or that you don’t understand to add value to the sales team because we’re here to maybe change that.
So let’s go around the room. What is your favorite newsroom metric? Let’s start with that. Anyone at this table? We’re going to come back. Keep thinking. Newsroom metric. I’ve given you authorship. There’s a great plugin run by fusion that also includes who edits the story and who is the photographer and allows you to build dashboards off of that. Is there any favored metrics? And I can keep talking all day but—there we go.
PARTICIPANT: The metric is engagement, engaging with the story, how far they’re reading down, how much time they’re spending on it. And I know preppie has done a lot of work around that. But still rarely most common metric on the actual data.
PARTICIPANT: Yeah. So engagement or viewability are big ones. Viewability for those of you are not familiar is a metric in the ad space right now that is 50% of an ad should be in frame for at least one second.
That’s a nice idea. But in practice, it doesn’t work very often. And in newsrooms, if 50% of our content is in view for one second, that’s a failure; right? So how do we understand engagement and viewability from editorial perspective? And then sell it? There’s opportunity of look at how many paragraphs people are reading, what is the length of article people are reading. All of these are great metrics of great examples that once we know, we can see that these are articles that have increased engagement. And then you can reach across the aisle and say to the sales team, maybe we can build a product that says sell ads on our most engaging content.
So that’s a great example. Is there any other stuff? We’ll keep going.
PARTICIPANT: Track facts.
PARTICIPANT: Track facts. Yeah, that’s a good one. People who are linking back to you. The flip side is there’s all of this social media stuff that’s giving us information. So people who are looking back to us has a lot of value. There’s things about interaccepting of where sources of traffic are coming from to better understand what your audience is and also better understand by advertising. That’s a great opportunity as well to reach out to your sales team and say, hey, we know that ought of people that are linking to—we know politico are linking to us frequently. Those people who are coming to us are very fast movers, so we need to drop our interspecial that we’re running. And maybe there’s an ad that would better serve that particular audience. Maybe there’s a geoaspect; right? A lot of the people might be coming in from DC. That’s an opportunity to look at those metrics and pass them along and to build new products around them.
So this is a big part of it. We’re talking about taking those metrics, taking those things that you find to be making successful journalism and creating successful journalism and reaching out across the floor to the sales room. This is pretty important for a lot of reasons. Like I said this means that we can build a sales funnel, a sales room, and a sales practice that reflects the values of editorial. We don’t have a situation where the Church of Scientology is running a native ad on our site; right? That happens, then the sales team doesn’t know or share our values.
We got authorship, we’ve got reading time, and word count was another thing we mentioned. So those are major metrics. So we’re going to talk about some of the things that have been done in practice.
PARTICIPANT: Yeah. So I’m sure everyone in here is somewhat involved, you know, in building. But does anyone actually work across or connecting dots between ad teams, edit teams, core engineering teams? Oh, cool. Awesome. Awesome. I thought I was alone. Great.
So something that we’ve really practiced and that we found to be very valuable is that everyone likes to be tied to revenue, you know? Personally I’ve always tied product and engineering, and I do both product and engineering, and I like to stay closer to revenue so that I keep my job because where the money is I usually get to stay. But what’s been really interesting is that I found a lot of people really tie themselves to the values to be able to keep that company going.
And the main reason there’s been such disconnect is products haveose. So basically what you mentioned, when you build product about consumer editorial, it’s about engagement, it’s about what’s being shared, who’s going to see it. All of these sunshine-type metrics. When it’s ads it’s, like, how do we collect data? How do we retarget? How do we reach the user? I’ll talk about a product later where we basically opened it up to all of our users and asked what they thought about the ads and it’s I already bought those sheets and that’s because, well, shit the technology we use will follow and follow you and make you buy them twice.
So what’s been a major shift and what works is really finding the right people to work across and to bring those values together, you know? We learned this with native advertising, which I think many people are familiar with. But basically the content that advertisers created and gets to users on our site, it has a gray box that says don’t touch that. And the main benefit of what that has been is not so much, hey, we can offer clients editorial tools or we can offer them a way to reach our users the way that, you know, our editorial team reaches them. Which is of course a benefit, but that’s the obvious one.
The thing that’s most amazing is that build products like that that leverage newsroom intelligence and value has opened up a conversation with advertisers now. So it’s not, like, hey, here’s our ad, run it. Or here’s a video, run it. It’s, hey, this is the message we want to get across, you know your users, how do we best leverage these products to reach your users in a very authentic way?
And that started in native, but that’s because products haven’t built to solve all of these existing products. People ad block and display is dead, well, it’s a 30 billion-dollar business in this industry. From the client’s point of view, they’re spending millions of dollars in the TV commercial, they spent 1% of that to reach the entire population across the United States across desktop display. So it’s not going anywhere, but we can make it better, or make video better or products better.
And that comes to the effect when the entire organization is working together.
So I’m going to talk about a couple of examples where we leverage editorial tools to build new product. Again, on the post talks I’ll link tolts tools. Many of them are open source, so we can kind of collaborate together. The goal of this is if you’re in this room, you’re actually interested in this, and there’s not that many people interested in it yet. They will be once they realize that their businesses are more and more at risk. But all of us together could help make this change. And at least at the Washington Post, we’ve seen a ton of success in doing this.
So the first one quickly I’ve been talking ab Bandito and blind testing. Beta testing has always been crucial in how we deliver to our user on the site. And that should be same for native as well. And testing always works. You want to double down on what works best, in the advertising world, it was always all right. Well, there’s two different creative and whatever creative clicks on most, let’s double down on that to go from .07 or .09. We can do better if we allocate to it.
What happened in Bandito is one of the best methods to leverage our newsroom to get our users to reach them and click them in areas that they want to be reached. And Bandito, which basically stands for the name was taken from bandit, like a one-armed bandit in a slot machine meaning when you pull that slot machine and hit that slot machine and hit gold, you’re going to win.
But you need to take a bunch of different risks and factors in order to find out what that rotation is to win.
So Bandito was built within our content management system and the idea is that when editors create content, they have the option to create, like, a ton of different headlines and different images—sorry. And be able to double down on what’s working best without them having to worry about it.
So in the newsroom as much as in the ad product side, there’s been a ton of difficulty to get our editors to think about technology. They’re journalists, they just want to create content, they’re not as focused as much on it. So we have to build the tools to do this. So what we did was we saw a huge success in the editorial of our content.
So we said, hey, let’s leverage this Bandito-type algorithm when it comes to not just our content but our display ads as well. And what we do is work with clients and take their measures. And what we’ve done is a lot of them have 300 by 250 ad tags, very standard placements, but they want to be content or consuming our content and reach our users. If you put it simply, a lot of our—or all of our consumers coming to all of our publishing because the want to read our content. Advertisers are advertising with us because they want to read our content.
So this really brings in two different levels. One is with all standard display-type units, we leverage the Bandito editorial news tool to make a 300 by 250 banner and a 300 by 600 pulling in relative content based on what that user is interested in on the page.
And what we’ve seen is our users are actually engaging more with our ads on our site because what used to be a 300 by 250 ad, which now is bringing in relative content beneath it for the user, the consumer is associated that with something, like, you know, the most popular content or related content. So they’re bringing a value and a different utility. And that comes through what we’re learning here.
So bridging these gaps, we’re finding that leveraging Bandito and AD testing on headlines and ads are both bringing higher engagement rates to our content as well as engagement to our clients as well.
Oh, and then I’ll go quickly again. Did you want to say anything?
PARTICIPANT: I mean the idea as we’re going through these is just let’s—these are sort of to seat us for the next half of the session, which will be to pick up some of these ideas to implement on our own. You take this one, I’ll take the next.
PARTICIPANT: So I think another huge issue besides working across our newsrooms and really being able to say okay. How do we work together to build better experiences on the revenue and moderation side is having to build on top of all of these old stacks. You know, one being double-click. Is anyone from double-click or Google in here? Okay. So we can talk freely. We can stay on the record. It’s fine.
So building on top of legacy tools. We’re in a space right now where our clients are very comfortable with what they’re getting. You know, they want viewability, you know? They want high engagement rates, they want to reach certain users. They have their creative. So anyone who has worked with the client knows maybe not so much in the native world but on the owner side, they’ll deliver us creative that has 50 different ad qualities all verifying all of the sudden scripts that are measuring all the things that are valuable to them. But with that becomes very heavy ads, poor user experience, but it’s hard for us to go back and say, hey, if you change this or optimize this, that would be great, and they would say, no, we’re running somewhere else. And then you’re at risk.
So we had to work within the confines of what’s already existing out there to help build and optimize tools that are even outside our control.
So one example that we’ve done, and I encourage everyone to check it out. Is we built a progressive Web app through Google. So we went all in on Google when it first came out, which allows anyone in search to search our content and same with Facebook.
But what we recognized early on, and this was, again, in collaboration with the newsroom and the advertising side. Is that no one fixed it for ads. So I mentioned it earlier. Amp was great. For the two months that amp was out, you could read a whole article without seeing one ad. So we weren’t getting paid, there were issues on the advertising side, but it wasn’t addressed. And I’ve had frequent conversations where with other publishers where it’s, like, well, this is an issue, but it’s Google’s to fix, you know? Or Facebook’s to fix. But at that point we’re losing money. And if we’re becoming a distributed world more and more, and you want to on the opt in because we want to be where the users are and we’re not fixing these problems injunction that, in conjunction with. It’s our dot-com, we launched progressive Web app, so go to Washingtonpost.com/PWA, and it’s live right now if you want to check it out. And we worked editorially with Google to build our progressive Web
app and what does that really mean? It’s the taking of amp and also PWA, which is a new tool to be able to build a new Washington Post. And you have to be secure to do it, and it was a beta program. If you go to our site, our site loads in 80 milliseconds, and amp loads at 400 milliseconds. So that’s great and the editorial team gets excited, but we can’t move everything over there unless we can make money off of it. And ads load in four, eight, sometimes 15 seconds, and you make content faster, that’s going to feel like 40, 50 seconds. So we were building on top of legacy tools.
And what we did within the PWA environment, we one pushed Google. So the one amazing thing you can do is fight the monetization methods, if your team is progressive, you can go to Google and all the partners that you’re working with on both sides of the coin and say, look, we’re moving this way, and we need you to push and go this way, otherwise we’re not going to be able to work with you. There’s enough competition out there where we’re all feared.
But what we’ve done as well is we’ve worked with Google, and we basically built on top of double-click to satisfy the needs of our company and does is going to be open source as well.
But we built a development within the ad server which basically allows us to load a script, load the initial ad script immediately as the page loads so that two to three seconds that used to go through the hierarchy, we’re able to eliminate that within what we call our PWA server, but it’s on top of double-click, so it’s not new, not proprietary. With Google, they’ll load ads and lazy loading and all of the benefits of that but also other things where you can predict when an ad’s going to come into view, and you can load it and so on.
And what we’ve been able to do very early on is cut our ad load times in half, we’re going to Google and building ad around it and Google wants to bring that to everyone else.
So what’s interesting is we found when you dedicate time to do these things, you have time to be flexible, this space and our advertisers are excited about it, and we can have those conversations. And there was a great conversation last week when the product came out because they’re basically saying our clients need to be creative a certain way. There needs to be a certain set of scripts on that creative and so on.
So I wouldn’t—I guess the main thing I want to get just out of this example is one we built on top of it, and I’ll share with you. But two you shouldn’t be kind of slowing down or feeling discouraged because of what’s already out there. There’s companies that are already competing with these, and there is the opportunity to push and build on top of it. You may be able to celebrate for a minute before Google says, oh, that’s cool and take it. But then at least we’re all helping each other, you know? The one great thing about that and everyone has their own feelings is that before publishers used to compete and work and keep our tech within our own and not want to share it, now what I think has happened is we’ve had conversations like happened here but also happening around, like, back at home where I’m sitting down with the New York Times now and other publishers in the space and say, well, how can we make everything better because this is our news source and so on.
PARTICIPANT: I’m going to quickly as well—so these are some things that have already happened as we’re going into the second half, I want you to think about what sort of things we can build. If you go to the session page, we do have already a document up with a whole bunch of links and resources and some nice, blank pages.
So one of the ideas that we wanted to talk about if you are lacking in ideas to com complement and get to this session is exclusions and tags. When you tag something for SEO reasons, within your media, within your journalism, those may be great for helping the readers find your stuff. But sometimes you may end up with the wrong ads next to it.
So ideas that will allow us to send specific tags to advertisements as opposed to just general ones. And sometimes tabs that may not be readers but for advertisements.
That is a way to coordinate and think about our taxonomies alongside the sales department.
Another thing we can look at is how do we reach out to our users? They are having bad experiences. If you’ve ever been on one of your sites mailboxes that deal with users sending feedback, you’ve probably heard about probably in detail. So is there an opportunity to build a tool that makes reporting bad ads easier? All of us are probably running on very similar advertising platforms. So if we can build a tool that can do this, it helps us all.
And there’s opportunities as well to talk about what ads use that we don’t want them to use.
So a great example of that is how many of you in your current sites are using the JavaScript document.write. We’ve got one. Two?
PARTICIPANT: Probably.
PARTICIPANT: Probably. Well, don’t. But it’s understandable. That said, after you’ve used it, you don’t want advertisements to use it. One of the major problems in advertisements is they will hijack gauges using document.write. So can we override that function on our pages with a very simple line of JavaScript? We totally can.
So this is an opportunity to talk about how ads create bad experiences technically and finally to intercept those options and disable.
And if we come up with a common set of scripts there, that’s another place all of us can use it. I know that I personally have been—had my browser hijacked on a many number of sites.
So the last thing is talk about the questions, the problems that we’re solving in building this technology.
When you break it down, your average user who comes to your average news website is earning for your site an ad revenue $6 a month. This is not a huge number. It’s not an amazingly large obstacle to overcome. If we’re building new technology that comes out of the newsroom, that’s the value that we solve for to break even on coming away from the old stuff.
So as we’re thinking about this, sometimes you look at the advertisement situation on your site, if you’ve ever been to Salon.com, you know we have a lot of advertising, and it seems overwhelming. But it doesn’t have to be. What we’re solving for, those numbers per user, the value that we have to ask of our readers in order to keep ourselves in business is not that large when it comes right down to it.
So those are the things that we are aiming to solve for.
Before we break out into hopefully some ideas on programs we develop in code, I know a lot of you are working in newsrooms, working with journalists. I work a lot with ad tech, you work in ad ops. This is your opportunity if you have any questions about ad tech to ask people who work with it pretty much every day.
So is there any questions before we go into the next session? The next part of this session rather. Go for it.
PARTICIPANT: Can you talk a little bit about how you monitor and track, like, ad load times? It’s one of the things my team is having a bit of trouble with is having reliable, consistent data.
PARTICIPANT: Yeah. Yeah, so I can speak to that directly. We’ve been working a lot on that.
So one is I think that a big struggle for really measuring the ads and their speed and latency and we say should we use this header? Should we use this integration and you pull it back, you lose 50 grand in one day and your boss freaks out. And you’re, like, I don’t want to take that risk.
The things that you can control and you can’t control. So the things that we can control are initial load time and building products on a better experience post click. So, for example, we launch something called fuse, which we dubbed as the Facebook instant for ad advertising. So once that ad loads, if someone spends at a certain level, when they click the level, we already have the post cache loaded, it doesn’t load a creative URL, which was a huge problem.
We can control with Google’s help of course because we’re on double-click, but double-click is obviously the most common one, we can control initial load a bit by working through the hierarchy of our load times but also building those elements on talk, and I’ll talk to you and show you after what we did so that you can take that back.
But there’s things we can’t control. So you can measure initial load, we do use internal monitors, we also use Ph.D. tests and a few others to kind of see that issue. But the thing we can’t control, which is the biggest problem with the advertising industry at large is what the clients are providing to us.
So I was talking to someone earlier where okay. That’s great that Google amp has ad requirements so tha antesser wants to advertise, they’re not putting all of that on double-click. So when your clients are buying on your sites, we’re supposed to be measuring and regulating it. But even those regulations are high, and we can’t control what’s going through that, special problematic and so on.
So I would say focus on what you can work on control. We found success on that and cut down on 50% of our ad load times. It’s good to have a Dev environment and testing environment to do that. It’s hard to do that on a live site. I would measure that so that function those preinitial load times because our 200 by 350 is loading at 1.4 second with header bidders in there. Without header bidders, it would be 400 milliseconds. You can see who the violator and speak to them about that. But if you do have a test, that’s the main indicator because it would be hard to see which variables are affecting it on your current site.
PARTICIPANT: And AV testing in particular so that you can test a small section of your audience and see their response. You can look at using tools like Chromes, developer tools to black box particular scripts and see how that affects your site.
When we know about particular scripts that we’re running in our partner advertising firms that we’re concerned they are the cause of a delay, we can use tools to disable a particular script and see how that changes our site. And of course there’s the ability to ad some JavaScript hooks around various files loading and metrics being sent out while you know what they are.
Of course you don’t always know what they are. And there’s opportunities DFP has a sandbox mode if you’re using DFP. And so that can also allow you to test some safer versions of the ads and perhaps provide some push back.
More questions going around the room. We’ve got one.
PARTICIPANT: So I’m in a smallish newsroom, and we don’t have much expertise right now, we’re looking to hire what sort of skills do we look for in ad ops developer, engineer as compared to editorial or platform person? And how does it overlap with performance and security experts?
PARTICIPANT: Yeah, so there’s a couple of things. I mean one I think he’s perfect to ask because he doesn’t even work in ad tech, but he’s been forced to because they’re small; right? So that’s a great example. You need him. So if he’s looking. But, yeah, I mean I think it really depends. I think the question before raises something that’s extremely important that during the breakout session, we should focus on which the way it works now, it’s wrong—it’s not working; right? And we’re identifying all the ways current advertising and ad logistics do not work, but we’re still trying to work within the confines of what it is because that’s how our clients get there.
So the main thing we have to do is come up with new ideas and Ren rich the ideas to drive that revenue, but we need to do that together. Publishers didn’t always work together with this. At least when we’re looking for talent within the ad ops space, you want people who are fluent in a lot of products, a lot of ad ops products are proprietary, so you want someone who’s fluent in double-click, you want someone who’s fluent in other exchanges and program-type bidders that if you’re smaller is how you’re going to make a lot of your money working with the ad Xs and then others out there. There’s great tools like prebit.JS that helps you with those load times and stuff like that. But that’s a big company like Washington Post what we look with someone in those expertise of the relationships someone working with those. Now, you –
PARTICIPANT: I mean I would say also you want someone who’s a partner who understands the front-end well enough to optimize for what your needs are. And just, like, a quick short cut if your hire can explain header bidding to you, you’ve probably got a solid person because that is not a large pool.
Any other questions?
PARTICIPANT: I’m just curious. So I might have missed this. But what are you doing differently on your PWA site?
PARTICIPANT: Yeah. So basically when we were building amp, you know, we found a lot of the caching functionalities that we were able to use, we could take off our site. So the Washington Post when Bezos bought the company three years ago, he wanted the site to load in one second. I’ve only been there ten months. But for the past three years, we have a meeting with him, and we tell him how much it loads, and it’s always over one second, and he’s just dissatisfied. So the whole goal has always been how do we make things faster? And advertising has been the biggest restriction. And what PWA really came out of was we saw the value in amp, and we started asking the question, well, why can’t we just do that on our own? And to go back to what I said earlier, it’s hard to pull back. Our ad is pretty fast and has tons of celebrations with the video players, so there’s a ton of variables within there.
But we knew that we weren’t able to make that faster than we’ve already made that. So PWA is—and I encourage anyone who is interested to talk to your Google partner about it. But what it enabled us to do is one we were a secure site. So I don’t know how many of you are HTTPS. Great. So we went secure last year, and that opened a ton of opportunities for us, besides the obvious ones of not being hijacked. And what that did is through the app integration there is Google had a testing thing called PWA. So what it means is progressive Web app, it’s on the URL, but it behaves like an app. So the beauty of is it’s pulling in all the content, the hierarchy it will pull the text first and then the titles, and then the images. So if we lose Internet or you don’t have service, you can still read content because ed. So it behaves like an app, but it’s on the Web I’m going to do that add something. I think there are a number of smaller publications here that may not
have as good of a relationship with Google. Also along those lines, like, the stuff Google does for PWA and amp involves a lot of progressive loading techniques. And that’s an option that you can do without having to go to Google. React is a library that’s particularly good at it. The idea is to understand your page and we’re talking about the metrics before. To understand what are the most important parts of your page to the readers, and what is the most important parts of the page to continue to stay in business, and figure out a way to optimize how to serve those first. And then you can use JavaScript in the case of something like React to essentially hold on loading pieces of your page.
Amp is open source. I’m not as big of amp as they are at Washington Post, but that said, open code, it’s out there, it’s useful, maybe it’s useful to you. There are a number of others out there if you Google search progressive Web design out there as well. And we lost the thing.
I don’t want to cut anyone off, but I do want to give us some time to think about projects. And what we can start building.
So this is a link if you’re not already there to the ether pad for this session where we can type in projects. In addition, everyone should have paper and pen and some little sticky note-type things on their desks.
Basically what I’m hoping we can talk about is there’s some examples in there, and we’ve gone over some stuff. Is if you are going to come out of this session with something beyond this ether pad and the wonderful transcript that’s going on here, the thing we want to take back is some projects you can bring back to your newsroom, and you can start development on them, and you can reach out to your sales team with and start a collaboration. And not only that, but also to start a collaboration with each other.
Everybody is on the ether pad now hopefully. I would encourage you to at the end of it add Twitter information, your GitHub, any of that. But also as we come out of this, to think about whatever the tool that you’re going to be building if you’re going to be building a tool or workflow or a way that you can keep those communication channels open between you and your sales team to put that in the open. At this point, the—what is it? The percent of ads buy that goes to Google and Facebook? 75%. 75% of ad buys go to Google and Facebook. We are all in this together as news organizations. Perhaps not by choice, but definitely because we need to be.
If as news organizations, we need to survive, we need to think about new ways of doing things, which is this is about, and we need to cooperate, which is what SRCCON is about but also hopefully what comes out of this.
So what I would like you to do is maybe join the smaller tables together if you’re, like, two people and start thinking about some ideas. You don’t have to start writing code right now. But please write, like, maybe an outline of what you think it should do. We’ll bring it up here at the end and have some discussions, and let’s just break out.
I will be here and walking, and you will be as well. If you have any questions, or would like some feedback on a idea. The slide show, did we put it in the thing? The ether pad? We’ll put the slide show in the ether pad.
PARTICIPANT: Yes. It’s in there.
PARTICIPANT: It’s in there? So that the slide ideas are there. And then let’s talk. Let’s think about some ideas that we can break out together. Metrics for your newsrooms. Ways to talk to sales.
PARTICIPANT: Just a little relating plug. I work at night foundation, and I love this idea of coming up with collaborative ideas because it’s looking to something we’re also wanting to support. And Friday we’re having a lunchtime discussion around newsroom technology. Exactly this type of thing. So I invite you all to come to Friday lunch. And if you’re interested, let me know, and I would love to stay in touch with you.
PARTICIPANT: Yeah. I think what we need actually is basically a corral project for ad tech. We need to ing on bringing newsroom values to sales and using those to make more profit.
So, yeah, we’ve got the time. Let’s talk to each other in groups hopefully.
[Group discussions]
PARTICIPANT: Okay. There are great discussions going on, and I hate to interrupt them, but we’ve got five minutes left, and I would like to hear from your tables what sort of stuff you’ve come up with. I highly encourage you to continue these conversations afterwards. We are going to lunch, which means you’ve got some time. You don’t have to rush to the next session. So please feel free to stay here and continue this discussion. But first, let’s just go to each table and get some—some stuff that was on top of mind. So let’s start with this table here.
PARTICIPANT: One thing –
PARTICIPANT: If everyone could just pause for a moment. I know it’s very exciting for safety journalism. But I invite you all to hear from each other.
PARTICIPANT: So one thing we talked about is there’s been some issues with advertisers backing out after they have some content of some sort. So at the times we have something called tragedy that’s super simple next to content or. It’s part of essentially open data that goes through all of the systems, works on the Web, works everywhere. And so we have not had those issues. The only issue in fighting between advertising is what if something were a tragedy were to happen? But usually it happens well.
PARTICIPANT: Yeah. I would say on top of that, we—when I was at the Huffington post, like, four or five years ago, we would do the exclusion, inclusion tagging. And what was amazing is that the editors were okay with our project managers going into the CMS and these tags that were at tags, and they weren’t on the front-end at all. So that sort of collaboration worked. We also tried to, like, build our own server within one of our applications where that was actually really hard to build. So that was one good thing with the double-click relationship is that they aren’t thinking about those things, even though there’s a ways to go there being able to connect with those and having communication is key. That’s a good one.
PARTICIPANT: Actually brings us—is that—good? Okay. Moving on. I know one of the ideas over discussed at this table. Would you also like to talk about things you talked about? Shuttering has we talked about two main ideas, one of which was the fact there’s no lightweight sort of publisher focus ad server, whatever; right? A lot of tools are way overkill a small, niche organization know who you’re selling to effective approximately sop what are the tools out there that would help with that? Or should something exist in order to implement that into CMS or whatever your workflow is.
And the other thing we talked about is essentially ad validators for—so that you can—if you’ve got work with or you can say, hey, does this ad actually meet the criteria? Is it secure? How fast is it? Like, why are we having to check this on the back end rather than when the ads are actually being ingested into the system or whatever.
PARTICIPANT: Yeah. So smart. I mean when we went through the HTTPS migration, the amount of work that had to be focused on just resetting off our end equals and working with our clients was manual labor. So those kinds of technologies are key.
PARTICIPANT: Some good ideas. Next table.
PARTICIPANT: Okay. So we talked about a lot of things. But I think last thing that we came down to was there’s trying to balance for user experience that keeps people coming back to the site instead of going somewhere else with better user experience. We talked about trying to come up with a tool that lets the particular ad. With the effect that had on the page so that when the CEO is saying you need to get the $7 per click auto play video to go back. And say, well, we’ll run these ads, we’ll lose half of our audience.
PARTICIPANT: Right. More from that table? Okay. We’re good? Awesome.
PARTICIPANT: So on our end, our conversation sort of centered on client sophistication. How, you know, when you’re working with big corporations or, you know, fancy agencies or whatever, that’s one thing. But in my case at least, art sales teams are working with pretty unsophisticated clients who do most of their ad design in-house. And sometimes more often than you would think, they’re not very good at it. And so that took off what we see from display ads. We get a lot of terrible display ad period and a lot of narrative period. And that’s a problem that I know sort of stretches across the industry that even, like, “The New York Times” has studio specifically to create content for their clients because their clients can’t figure out how to do it for themselves.
So if those types of clients are having problems, then our types of clients in local, regional markets definitely are going to have those problems. So trying to come up with ways of, you know, building tools that might be able to mediate that problem, you know, where clients can come in and potentially, you know, put in the key words, put in the information that they want to convey to their, you know, to the visitor and, you know, suggesting, like, this is what a good ad might look like; right? These are the key words that you have to focus on and stuff like that.
PARTICIPANT: Yeah. And I think that’s a great thing also it’s something that can be sold as an addition. Like a design fee for an ad. An optimization fee. And local is an opportunity in that there’s an entirely different set of requirements that can be coiled into building native ads for specific local use cases that come around like social media, campaigns, we talked about classified some of these tables, all those types of things are aspects that can be pulled in specifically for local advertisers. A different use case than national.
PARTICIPANT: Absolutely. . I think there’s a lot of opportunity there.
PARTICIPANT: Yeah. And I think something big that we’re all kind of speaking about right now that’s huge to recognize is that prior—like in the past ten years, we never really sat together and came up with these ideas to solve these problems, again, with suitability we’re now forced to on all different levels, and I think you kind of look back at product innovation that we all kind of laugh at because we’re, like, this content doesn’t make sense. But because publishers make it together to solve that problem, someone else came in and now we all pay for it. And while we laugh at it, we’re getting a ton—at least the Washington Post at a larger scale, but others are getting revenue from that. And ad tech, we can wait for others to solve the problem or come up with ideas and make them scalable and not have to wait for someone else.
On the flip side, there’s all of this talk about how Google and Facebook are going to house our content but they don’t know how our editors work or how to sell advertising. So these tools that are built can be placed on Facebook and Google and other programs. Because as much as it seems they’re taking everything from us, they don’t know exactly everything that we know and everything that we’ve gone through. So by working together and building these sort of things, we can then come to market and not just ourselves but the industry and also help partners that are hopefully helping us with the distribution and stuff like that as well.
PARTICIPANT: I think an important financial aspect of that is that every third party platform piece of software, tracking code that’s between the ad being served and your site is biting into money that will be potentially argue. The more stuff that we’re building ourselves, the more the pie that we can grab and the more profitable we will be.
When an ad is coming to you with 30 different trackers on it, not only do those trackers slow down your site and make for a terrible user experience, but each of the trackers are getting a little tiny piece of money that could have gone to your budget and your journal and your technology.
So we are out of time. I think there was a lot of great ideas. We already saw that within this room there were things that were so important there was a little bit of overlap. There are ideas in the ether pad, hopefully you put into the ether pad that SRCCON has provided. To remind you, that is on the session paid. Hopefully you all will roll call no matter if you feel comfortable doing so and give us your information so that we can collaborate.
And in addition, it would be great to hear some of the platforms or ideas come reality, prototypes, all of that sort of stuff. And if that happens, you should bring it out there, talk about it, and Twitter promote it to this group, promote it to GitHub, promote it to all the wonderful things that are happening out there that are being used to talk about the great things we’re doing to make good stories.
But not as often talking about the things that are happening that are great to keep us in business.
So this has been a great session. I hope a lot of questions were answered. Our contact information is of course on Twitter, on ether pad, all that stuff. If you have additional questions, I’m here for the entire time. You can ask me. Now, you’re only here until the end of the day, but we’ll be glad to sort of delay our lunches to talk more. And please keep in touch and start building out these things. Take the lessons from the session, bring them back to your newsrooms, talk to your sales rooms, start breaking down some silos. We all need to work together because we’re all in it together.
And that is it. Thank you for coming.
[Applause]