ย Cloudflare TV

๐ŸŽ‚ Customer Spotlight: Nativo

Presented by Derek Yee, Oded Cohen
Originally aired onย 

Nativo, the leading native content technology platform, recently launched its Edge solution on Cloudflare Workers, which is 8X faster than current market solutions, increases revenue per ad impression and offers a better user experience for end users.

English
Birthday Week
Customer Spotlight

Transcript (Beta)

Hey, welcome everyone. My name is Derek Yee. I am a part of the Cloudflare product marketing team and today's customer spotlight is on Nativo.

And I'm joined by Oded Cohen, who's the CTO at Nativo.

And Oded is joining us today from LA. So, Oded, welcome.

Nice to be here. Excellent. Great. So, Oded, why don't you tell us a little bit about yourself, your role, Nativo, and actually what does Nativo do?

So I, you know, I'm a tech guy.

I'm the CTO of Nativo, been in tech for maybe 20 years now.

Originally from Israel, moved to the U.S. about 10 years ago. And at Nativo the last seven years, been in the ad tech business for the last 15 years.

And that's part of what Nativo does.

That's kind of like a brief, my background.

Nativo, I would say in the core, we're a technology company that helps publishers monetize their sites, keep the sites open and free, kind of maintain the open web mentality.

And we do that, we're focusing on doing it without, with as minimal interruption as possible for the experience on the site, both in terms of performance, user experience.

So when we started, basically, the whole idea was, you know, if you look at how sites are making money, if they're not selling anything, they're using ads.

And, you know, eight, nine years ago, mostly this was banners.

Okay. Banners, maybe some video. And the idea behind banners is that the advertiser is trying to grab your attention from the site and pushes you towards their site, right?

So this is not great, not great for the user. So they're kind of trying to, you know, catch your eyes and kind of get you to move to a different site.

It's not good for the publisher, right? You came to the publisher, to the site to consume content, to read articles, and somebody is like pulling you away from that site.

And the idea was, what if we can actually give a much better user experience to consumers, to the customers that are coming to those sites, and still have monetization capabilities around that.

And so the idea was, what if we basically mimic the site behavior, instead of trying to attract your attention, let's just embed some of the ads within the site in a more native way.

Okay. But so, for example, if you're looking for articles, we're going to show you sponsor articles, articles that advertisers are willing to pay money to show up on the site.

And if you find this article interesting, if you find the headline, same as you find a, you know, editorial article interesting, and you click on it, instead of pulling you away from your experience and send you to a different site, we're going to show you the article, the sponsor article within the site.

So the idea behind it is that if I'm, you know, if I have a new product, if I'm Ford, I have the new Bronco, or whatever, an electric car, we were talking about a bit about it.

And I want to educate the consumer about it, even early on before it's even in market, I can take an article, I can write an article about the concept of electric car where it's good, and maybe about how Bronco is helping that.

And I can then distribute that article where consumer are consuming that type of content, maybe on automotive sites, maybe on news sites.

And so it's what we're seeing, what we saw is that it's a lot more engaging, the users reacting to it a lot in a lot better way.

It's good for the advertiser, because of that, for the publisher, it's a better user experience, it's more natural to the site.

And for the user, again, you're giving them information, you're not trying to attract them away from a site, you're just giving them the right content that they came to consume, just that it's sponsored by some publisher, some advertiser.

Got it. So it sounds almost like what you're doing, you know, it's funny you bring it.

So when I think of ad tech, we think of the traditional, you go to a website, there's a banner.

And the whole purpose of these banners, in terms of conversion is to click through to actually get them to go to somewhere else.

And what you're finding now is like, hey, if I can find I know information that's going to be really relevant for this visitor that's on this page, why go somewhere else?

Why not present that in a space that they can consume it right there and then and I know it's going to be relevant for them.

And there's no sort of detracting from the experience when someone goes there, and they're still getting the impression, you're probably tracking the engagement in terms of how much time they're actually engaging with that content.

So it's almost like it's moving away, it's evolving away from the typical banner ads that we know as far as traditional advertising.

Yeah, so it's exactly that. It's highly focused on contextual relevancy.

So we have tools around making sure that we are presenting the right content to the right people in the right context.

So you don't go to a family and parents and get it like an article about some, I don't know, sports event, because it's not what you came to do.

So we're trying to kind of match what your context right now, what you're reading about right now.

And yeah, you're right.

Now, it's not just about the clicking on the banner. Now, when you're consuming the content, when you're reading the article, we can track a lot more metrics about how do you engage with the content.

And all the studies that we've done showed remarkable results in terms of the impact on potential customers, the brand perception.

It's a lot better experience and more reliable to some degree.

Or if you actually read the article and reliable sources on good sites, you kind of get some of that branding into your content as well, even though it says it's sponsored by Ford or by whatever company, you kind of get that overall experience and it feels a lot better.

And that's why the reaction from consumers is better.

Great. So Nativo facilitates this. And is this fairly prevalent in the industry in terms of ad tech today?

I mean, or are there other plays that does the same kind of thing that you guys offer or is it sort of a unique sort of space?

So eight years ago, it was very unique. It was the term native advertisers didn't even exist.

Over time, the industry started adopting that. And you had a lot of that coming also for social media, where you have all the sponsored content embedded within the social media.

What the industry was not able to do is what we're doing, where you can actually consume the content within the site.

So when people say native advertising, what they mean is just the unit that promotes the article or the content.

But once you click on it, you do go to a different site.

So that's what most people refer to as native advertising.

It's kind of like a fancy banner to some degree. It's better, but it's not that different than a banner.

I don't think anybody is able to do in the scale that we're doing.

Just take a piece of content and distribute it across hundreds of different sites and reaching the audience where they're there.

Now, over the years, we've evolved.

We introduced more formats of more type of content.

We can do video. We actually can do stories. So as you look at your site via mobile, you can have a piece of content that is a stories format.

You can do a lot more things through this tech.

And the idea here is that the tech is generic enough that we can create any type of experience that we want.

It wasn't built as banners, which is a bit more strict and less flexible.

It was built as a more flexible platform so we can introduce more formats and more capabilities.

But we definitely know there's competition around native advertising in general.

Yeah. Great. So talk to me about, so your role, you're the CTO there.

You've been there for five plus years. You've been there for quite a while.

And you think about your responsibilities is all about maintaining the infrastructure that's needed to be able to provide this platform for advertisers.

Give me a little, if I were to guess, a little peek behind the curtain as far as what is your sort of setup like?

And what are some of the things that are, I guess, most difficult in you guys being able to do what you do?

And so what are the key issues that are always top of mind for you?

So obviously scale is a key thing.

Scale, reliability of the platform. Remember, the publisher relies on us on this tech to be able to monetize their site.

If our system is down, they lose money.

And so all that reliability is a crucial point, always on a hundred percent of the time, kind of like lab flare, right?

Second part is scalability. We're running on thousands of different sites.

We haven't touched that, but the system started as what we called a publisher ad server, meaning publisher can run their own campaign.

We also introduced buying platforms so that advertisers could go into the platform and kind of run their own campaigns as well and have the publisher choose what demand they want to allow to flow into their site.

Over time, we also introduced programmatic integration, meaning that not all buyers will go into our platform to buy.

Some of them have their own buying platform, what we call DSP, demand side platform.

And they would want to bid in real time and choose whether they want to win this impression and at one price.

So our system now auctions in real time.

So every time we get a call from a page, we then send on that page, there may be maybe two, three, four different locations.

Each of those locations, we will send requests to maybe 40, 50 different bidders in real time, process it through the auction, and then return the ad to the page.

All of that, obviously at a high scale is very challenging.

And so there's a lot of things around that that put a lot of emphasis.

We're collecting a lot of data. So every interactions with our servers, every bid request that we send, every response we get, every price that we get, every interaction the user did, all of that is being collected and tracked and being surfaced back to the app close to real time within a few minutes.

And so there's a lot of challenges around scale that we have to deal with.

Got it. So you guys are obviously hosting this all in cloud.

I would expect nothing else. But so can you talk a little bit about how your architecture has evolved?

And maybe, I know you guys have been with Cloudflare within the last year.

I think you're a fairly new customer to us. I'd love to hear a little bit about what was existing before and maybe what are some of the challenges that actually prompted you even to consider looking at Cloudflare?

Because I know you're actually using a variety of CDN products.

So we're not the only provider in that respect.

But I'd love to kind of hear about that journey and hear what brought you to Cloudflare and what are some of the key issues that you were looking to solve?

So we'll probably have to go back almost two years. I'll start by saying, if you remember, I said it was critical for us that we're doing what we're doing for the publisher with minimal impact on their site and with the best user experience we can have.

What we found is no matter how good we are and no matter how good the publisher is with their own CMS, with their own pages, what we found is that because there's other tags, other ad platforms running on, ad tech running on their site, what we end up seeing is that publisher would use, let's say they would use Cloudflare as their CDN to get the page to start rendering very quickly in the browser.

So within less than a second, maybe a few hundred milliseconds. But then when the browser starts rendering the page, there's so many JavaScript running on this page, calling different servers, loading other JavaScript, calling more servers, and eventually injecting ads that it takes.

First, the ads will show up four or five, sometimes 10 seconds in delay.

In that timeframe, because the browser engine is occupied executing JavaScript, it's really hard to scroll and navigate the site.

It's stuck because of all the JavaScript running. And so we realized that the problem, there's so much we can do if we're staying on the browser side.

And there's so much the publisher can do when they're optimizing their CMS. And what we realized is that the best way to solve this is to move everything to the server side.

Instead of doing all this ad injection that kind of restructure the page, as you start scrolling, things are, you know, you scroll and then after four seconds, you come to click on an article and suddenly because an ad showed up on the top, everything drops down.

All those things can be solved if we just do everything on the server side.

And so we figured, okay, two approaches. One is go to the CMS, try to do it on the CMS.

Try to inject the ad on the CMS side. The problem is CMSs are not scalable.

They actually use CDNs so that they don't have to deal with every request.

Also, there's endless amount of CMSs. Some are proprietary for the publisher.

So that was not a good solution. So what is the best solution?

We need to sit on the edge. We need to be able to do what we do. So we found that we have all the information and all the capabilities that we need to do server side injection of ads and reduce the clutter of JavaScript on the page that the browser has to render.

And so we figured, okay, a couple years ago, okay, let's look at the CDNs.

Let's look, you know, top CDNs, what do they have to offer?

The good news was that the notion of, we knew that we need compute capability.

This was not about some configuration, you know, changing a header.

We need actual compute capability code running on the edge to do that. Two years ago, the good news is that it started.

It was already moving forward, almost in all the CDNs, but it was very early on.

Like almost nobody had the basic capability that we need.

And practically we couldn't really solve it on the CDN, not even on Cloudflare, even though Cloudflare was ahead of the competition at that point.

So we ended up building it ourselves. We had to build a layer outside of the CDN that sits between the CDN and the publisher CMS and kind of do that, facilitate that.

Let me stitch the page and the ads and deliver that through the CDN to the user.

And we launched it last year. You know, we got 10 sites on it. We learned a lot.

There's a lot of challenges around how do you monetize in that situation?

And there's challenges that we've learned, but the result were amazing.

The results were really, really good, significant increase of potential impressions.

Because remember, if the ad takes five seconds to load, the user might have already moved to a different page.

You lost the money for that. And so we see a lot more impressions.

We see the page loads a lot faster, what we call the time to interactivity increased or improved, decreased, improved.

And so we saw a great result.

But one of the things that we realized is that us sitting outside of the CDN is still a problem because publishers do not want to add another tech layer into this mix, right?

I mean, they trust their CDN. We knew that the solution for the long run is being able to take this code and run it on the edge.

And so end of last year, we did another round. Everybody made good progress.

Everybody were close, but only Cloudflare actually had what we needed. And so we partnered with your team.

There was a few missing things that we needed, like small things.

So the team really helped us a lot to get us exactly what we needed.

And we launched it again. We basically moved all this layer into the Cloudflare edge compute on the workers and went live a few months ago.

I mean, we kind of transition all the existing customers to this solution and then double the amount of sites, more than double the amount of sites, still in a beta phase.

Again, seeing really, really great results, really, really great feedback from the publishers.

And now we're actually ready to take it live. And again, the notion, what we like about it is that before we had to tell, we had to explain a lot to the publishers, oh, you have to change your CDN routing to route to us and all those things.

Now we just tell them, here's a piece of worker code, right? Just install it.

You don't have to do anything beyond that. Now all the ads are server -side injected.

So you don't have to literally do anything except install on your CDN, the worker, and suddenly you get an increase in revenue.

And we're talking about significant increase in revenue, double digit percentages, the revenue per page that they get with all the functionality that we added through this could go beyond 50% increase in revenue.

That's huge, right? And so this is how we end up partnering with Cloudflare.

We're very excited about it. We're happy that we can open it to the public, to all the publishers, specifically at GA coming soon.

And yeah, it's exciting. So the big upside to this is that you've got publishers can stick with their own CDN.

You can use any CMS. There's really less friction in terms of them being able to adopt your services.

Whereas before you had a solution, it not only required a little bit more steps on their part, but even on your part, you had one other thing to have to maintain.

Ideally, you want to kind of eliminate that.

And moving into something like workers, this is the great thing I hear time and time again, is that people are going to these types of technologies because they don't want to have to maintain the infrastructure to be able to support a lot of this.

Yeah. So we're not strangers to maintaining high scale infrastructure.

We still have the ad platform infrastructure behind the scene that the worker needs to contact.

But yeah, it's one less infrastructure to manage.

Again, for the publisher, it's just easy, right? It's extremely easy, simple to do.

And what we realize is that if you think about it, this is really where things should have been.

It should have been all this clutter on the browser and the CDN and this capability of computer on edge, the workers capability.

It literally, the way we see it is that we can offer more and more services on that layer.

You mentioned CMS agnostic, that's super powerful to be able to do. We didn't stop there.

We actually have the capability because we're sitting on that layer.

And think about it, when you're for ad serving ads, you always, every time you see the user again, and if it's a different user from different place, you might serve a different ad, right?

You have different targeting criterias, different pacing, what campaign should run when.

And so every time you have to make a new decision, when you think about the content, the editorial content, it's the same every time.

Every time I go to the site, I'm going to see the same exact article and it will be the same as you see it.

And every once in a while, a new article comes up, but what if I already read the article?

Why would you waste such valuable prime space on the homepage?

I already read the article. Why are you showing me it again and again?

I'm not going to read it again. So what we found is because we have all those capabilities of customizing the message we're sending to users on the ad side, we actually can do the same for editorial.

So we can actually, hey, you know what?

Maybe we surface a different article at this point because the user already read this article.

So we can actually, on the worker side, can also improve and customize the experience of users, kind of like social media, right?

You go to your account, you see something, somebody goes to their account, see something else.

So we can customize also the editorial. And again, we rely on this worker capability and doing everything in real time, very, very quickly.

The worker has very, very powerful native code that we can run that is extremely fast that we're leveraging.

Now we were trying originally, we brought our own library because we didn't know if we would be able to accomplish everything with the native one.

And we saw a huge difference. Once we switched to the native library that Cloudflare provide on the workers, huge benefit in performance.

And so there's endless amount of things.

And I can think about a lot of other things that can be built on that layer that can be offered to your customers.

Think about it. The way I see it is that Cloudflare really costs money to the users.

You have to pay for reliability. That's one, right? And you have to pay for that, obviously.

You promised them security, right? Like from bots attacks and all those things.

But what we can bring is the third thing, which is you can actually also get money from this.

Once you sign up with Huffer and you just pull the switch on, now you also get money.

So it's a revenue generator, not a cost center anymore.

Yeah. So I think the relationship, I mean, again, if I'm a publisher that's very compelling, if I know that I can make 20, 30, 40, 50%, maybe even more revenue, and I give better user experience to my customers, it's almost a no-brainer.

You know, I'm almost like, as I hear you kind of talk through this, and earlier I asked you this question, surely realizing that there's JavaScript at the browser is causing slowdowns, regardless of your best effort or what you're doing at the server-side infrastructure as such.

Do you think any of your competitors have also realized that same problem?

They're trying to figure out how to solve that.

And have you sort of landed on a goldmine and through this session, maybe gave away what your secret sauce is going to be?

So obviously we have some patents around, but I don't think that's the main issue.

I would say, you know, the top competitor, so you need, the core thing is you need publisher tech.

It's not just the demand, it's the publisher tech.

And so there's not a lot of publisher tech out there because of the dominance of Google.

And a lot of the publisher tech out there right now is legacy.

It's built in a way where, I wouldn't say they can't change it, of course they change, everything is possible, right?

But the core of their tech relies on things that actually will make it impossible to do what we're doing.

So we don't need to know how the page is structured and all the elements of where the ad should show up because the response from the publisher, and we kind of render the ad and render the page and figure out.

We have all this information already in the backend.

So we can actually in parallel, both fetch the page from the publisher and make the ad decisions and merge them together.

Some of the legacy tech relies on looking at the page and figure out where things should be.

And so it's possible. I mean, I assume if that gets adopted and widespread, there might be, but it's not as easy of a task as you would imagine it.

Obviously everything is possible. A lot of the competition is not just, so on the publisher side, again, there's not a lot of strong competition because again, Google has like a history of killing all the competitors by just, they're just dominant enough to do that.

So. Right. Very interesting story. I mean, it's like, I keep hearing again and again, what people are doing with workers.

And it's always great to hear like in real life, like what you're actually doing with it.

Because I think that's one thing a lot of potential customers struggle. It's like, yeah, we understand the concept of what it is, but what can I do with it?

So we heard a simple A-B testing, but you guys are actually doing things really to help improve and sort of offload a lot of that JavaScript that happens at the browser level.

And it totally makes sense that this should sit at every Cloudflare Edge data center.

And just the experience is going to be certainly far superior. Yeah.

There's also another trend that is happening in ed tech right now, which is everything around privacy, everything around, no more third-party cookies, not being able to track you across sites.

The Edge actually fits this solution that we're talking about, doing everything on the server side on the publisher domain, right?

Because you're fetching the publisher and we're doing everything on the publisher domain.

It actually fits really well to that. You do everything on the server side.

You don't have to communicate anything to the client, to the browser, to send back some sensitive information.

Everything can remain outside of that scope.

Also, it's a cross-browser solution. There's other solutions. Some companies are suggesting, oh, maybe the browser should take some of the responsibility of some of those privacy, but then you have different browsers and different browser behaves differently.

And so this solution is same as CMS agnostic, it's browser agnostic.

So it works perfectly well with all browsers and it actually thrive in an environment which is privacy is a high priority.

So it's also very good, that trend.

Yeah. We're in an age right now where who would have think consumer data has become, it's really a goldmine, right?

And there's a lot of sensitivities now.

You've got GDPR, California having their data privacy laws. And maybe this is something just on- It's the entire business of Google and Facebook.

It's based on the fact that they know things about you. Even Amazon, their ad business is based on the fact that they know what you want to buy.

And so an advertiser willing to pay for it.

I think, again, I wouldn't look at it as a bad thing always.

Think about it. If you want the open web to remain open, right? If you want somebody to send you a link for an article and you can read it and not have to register and pay for it, which you see more and more now in some sites, then you need the advertising industry to flourish.

And you want the publisher to focus on creating good content and good user experience.

And you need the way for advertisers to reach their consumer.

You just want to do it in a respectable way. Good user experience, maintain some level of privacy for customers being able to opt out.

So it's not always bad, right? I mean, you go to Google search engine, it's free because- Everything has a price.

You say something's free, there's a- It's not evil.

It's just that it could be done in a better way, in a more respectable way. And that's kind of where the industry is going to.

Got it. All right. So we are closing out.

We've got about maybe about three minutes left. So I want to make sure we have enough time to kind of fluster.

So it sounds like you guys have had a great experience with Cloudflare, particularly working with Cloudflare Workers.

Here's a question I usually ask all of our guests, is that, okay, so if you were to talk to a colleague or a peer of yours about some of the things that you're doing with Cloudflare, what would be sort of the top reasons why you would sort of recommend Cloudflare to them if they're considering?

So for us, it's definitely, we see you guys are ahead of all the competition in terms of the everything around compute.

This was the focus point for us.

Believe me, we've talked to all the major CDNs. We know where they are at, like, there are different stages.

For us to see how advanced the solution you guys have already, and you're already starting to think about the next things or how we can elevate it.

Believe me, there's more things to do also on your end.

It's definitely where you want to be. If you need those capabilities, it's definitely where you want to be.

We've talked to other companies that were trying to do not exactly what we did, but also rely on that.

And they basically said, yeah, it's Cloudflare. That's the only one that we can use right now.

They're ahead of everything. So that's definitely a key aspect.

Got it. And since it is birthday week, any interest in what we announced on Monday in terms of the workers releases for like DurableLogix or CronTriggers?

So I'm not sure it's exactly perfect for us in our use case.

It could be in the future as we continue and evolve.

I did see some of your web analytics kind of a solution. I kind of like the fact, the concept of, hey, move it to the edge.

We're talking about privacy, right?

Right. And also some of those capabilities we can actually leverage.

I mean, you have a lot of things around bot detection and things like that. We have our own challenges in the ad tech industry.

How do we filter from? And by us sitting on the edge, we can actually kind of leverage some of those capabilities as well.

So there's a lot of interesting things, but time will tell. Yeah. Great. Well, I would absolutely appreciate your time too.

I think it was a very great conversation talking about some of the things that are happening in the ad tech space and in particular, Nativo, some of the advances you guys are making and the work that you're doing in the space.

So I appreciate you sharing that with us today. And I hope the audience enjoyed this last 30 minutes.

And I don't think we have any questions that came through live.

So I think we will close out for the session today.

So thanks again. Thank you. You have a good rest of your day. Thanks, everyone.

Bye.