ABOUT VIOLETA: LinkedIn  | SoundCloud  | Twitter  | Blinkist




ABOUT ROCKETSHIP HQ: Website | LinkedIn  | Twitter | YouTube

KEY HIGHLIGHTS

📈 The change in user perception of native vs non-native ads

🖊️ Ad to app: the extended user flow of a native ad campaign

🦙 There’s a lot of BI in the background for native ad performance optimization

🍚 The subtle components of a native ad campaign

🎚️ How to decide if native ads are the right channel for your product

🎭 The performance of different native ad placements on Facebook

🏦 How to use Instant Experience as a litmus test

🖼️ What the testing framework for native ads looks like

🤝 The efficacy of automated vs manual testing

🧪 The surprising result of testing Taboola and Outbrain content flows on Facebook 

🖊️ Headlines, creatives, article content and more: the variables of a native ad campaign

KEY QUOTES

Ads for the Soft Sell

What’s also interesting is that only around half of the people who are being served these kinds of ads know that they’re being served ads. And of those, only half, about half are skeptical. While the other ones are actually fine with it. So, it’s essentially a soft sell of your product.

Placing ads in the News Feed

On Facebook the placement that works the best is one that looks, quote unquote, native, where you don’t usually find the usual ads or ad placements. The news feed placement is actually the one that works the best. I think that’s because it can be mistaken for a post that someone may do like a brand made that is organic.

Catchy Titles for Strong Articles

The main piece of creative here is the article. And then, testing out different images and titles to go with it. Because the title and the copy and the combination of title and images are what gives you the CTR. And you first have to reach a certain CTR. Obviously, it’s always good to have a high CTR in order to send them to the articles. So, I would first start with a strong article and then test out different images and titles.

Reuse Creatives Across Platforms

As we were testing a lot of different articles on all of the platforms, we realized that it’s the same ones that work for Outbrain, Taboola, and Facebook.

The Human Gaze

One thing that is surprising is that images of people usually work the best. I think there’s some psychology around it. And it’s also within angle. So it’s the front angle where maybe you don’t necessarily want someone staring right back at you. But you can see the eyes. I think I remember reading something, probably on LinkedIn, about the psychology of this, where instantly, just because of the way that we’re built, whenever we encounter another pair of eyes – even if we’re scrolling on Instagram – we’re just instantly attracted to it as humans. 

FULL TRANSCRIPT BELOW

Shamanth: I’m very excited to welcome Violeta Pirnog to the mobile user acquisition show. Violeta, welcome to the show.

Violeta Pirnog: Hi. So nice to be here.

Shamanth: Thrilled to have you on the show today.  We’ve only crossed paths digitally. And every time I’ve learned a lot from  you. You’re always been very, very insightful. I’m excited to dive into some unique perspectives you’re going to bring on a topic that hasn’t been talked about much. So, definitely thrilled to have you on the show. So welcome, again.

Violeta Pirnog: Thank you so much. It’s really nice to be able to talk about things that haven’t been covered before. To be fair, it is kind of special as well, because I’ve been following your podcast, and your newsletter before that. You always seem to cover really interesting and insightful topics. I’m really genuinely honored to be invited and to be able to contribute to this. So thank you!

Shamanth: You’re welcome. Let’s get started. We’re going to talk about winning with native ad campaigns, particularly on Facebook. That’s not a combination that a lot of people talk about often. So let’s start with the basics. Can you tell us what native ad campaigns are?

Violeta Pirnog: Essentially, native ad campaigns … It’s kind of an oxymoron, isn’t it? It’s paid content, by definition. It’s a piece of content that would usually seem to be organic. A really insightful team of brilliant marketers behind this have optimized the end user journey so that they would actually make the user who’s being served that ad interested in getting to know more about their product. It’s a relatively unused format, especially for Facebook.

In general, I think brands and advertisers love native ads because the click through rates tend to be higher, and engagement is a bit stronger. Also, with the added side effect that it increases your brand awareness.

What’s also interesting is that only around half of the people who are being served these kinds of ads know that they’re being served ads. And of those, only half, about half are skeptical. While the other ones are actually fine with it. So, it’s essentially a soft sell of your product.

That’s how I would describe it.

Shamanth: That is crazy. I didn’t realize that that was the breakdown. And I would be curious as to how that breakdown compares with non native ads. I don’t know what that comparison looks like. 

So you said about half the people watching native ads don’t think that’s an ad, but I imagine everybody watching a non native ad knows. And you said of the remaining half, hate it and half like it. I’m curious, for a traditional ad, how many people would you say like it? And how many people would just hate it? If you had to take a guess?

Violeta Pirnog: That’s a really good question. I think it really depends on the platform. On some platforms – like Google for example – they’ve become less obvious about signaling their ads and the sponsored search results. That’s one of the best things to look at, the evolution over the years of how they basically signal on Facebook. Again, it’s different, and in the end, it all really depends on the platform. It also depends on the target audience because people who are more tech savvy or are used to being on social media more often, and used to ads, are able to identify them easier. Whereas if you just aren’t there a lot, maybe they don’t look into that as much.

Shamanth: Right.  And the way people react is very contextual. 

So, in a native ad campaign, if you’re running this on Facebook for a mobile app, help us understand what the user flow looks like.

Violeta Pirnog: The user flow is not that different from the usual app install campaigns, but it is a bit longer. Essentially what happens is from the ads and the platform people are taken to a magazine article. What happens there is there is a smart banner that is also sticky. So it stays with them the entire page, and of course, there’s a CTA there, and then takes them to the App Store. So it’s a bit of a longer funnel. It doesn’t take you directly to the App Store. But you do eventually go there, ideally.

It’s a bit of a longer funnel, and it provides context. It’s a great opportunity, especially for products where this is needed. Especially content driven products, like Blinkist, for example. It’s also a great way to preempt any kind of concerns that a user might have towards trying out that product, because then you can add elements in that page that help them to go towards the decision that you’re trying to get them to go to.

Shamanth: So the content and the selling on the page has to make up for the longer funnel and a longer journey. And therefore, there’s specific kinds of products that could benefit from that kind of education and selling that this feels more right. And then you describe the journey. They go into the landing page, and there’s a sticky banner, and then they go to the App Store. I would think the more obvious way to optimize these campaigns is to run a web conversion campaign that optimizes for clicks on the smart banner, because, once they go through to the App Store, that’s a different experience. Are you able to optimize these campaigns for postinstall events, trials, or purchases? Or even installs?

Violeta Pirnog: Yeah, that is possible because it can be web and mobile events. You always have the MMPs you can use for the mobile events. And for the web events, we use Piwik. But it basically does the same thing as Google Analytics and tracking the events. And then we just bring them into our in-house backend system. And then feed them back in using Facebook’s Pixel and their conversion API. We’re able to also then optimize for lower funnel events.

Shamanth: I see. Would you care to elaborate on how the Facebook Pixel and conversion APIs feed the purchase events, which are tracked by MMPs to Facebook?

Violeta Pirnog: It’s kind of a seamless journey. And that happens internally. I think a BI engineer would be more suited to explain. But essentially, in our backend system, it happens after they’re filtered through all of the attribution rules as well. They, I think, go through several other platforms and into our database, and then back to Facebook. So it’s a long process. 

Shamanth: I see. So would you say there’s no off the shelf way? So let’s just say there’s a small developer that wants to build this and optimize for purchases. This would be really hard for them to do it in an off the shelf manner? Or is there a simple solution that could be possible?

Violeta Pirnog: Yeah, you’re right. It would be a little bit more difficult to do it, but it’s definitely worth the effort.

Shamanth: Yeah, yeah, I can imagine, right? Because again, if you’re optimizing just for the leads, you’re getting very top of the funnel metrics. And you haven’t as much visibility into what happens downstream afterwards. Right?

Violeta Pirnog: Yeah. Definitely. At the same time, I think even just having the leads is very different. Because the CTR is generally higher. You do get a lot more leads, than maybe necessarily installs, with app install campaigns. So that helps.

Shamanth: And, speaking of the user flows, for a web campaign for a mobile app, there are a number of different flows that are possible. One, of course, is what you described, which is take them to an article page that goes to the App Store. But there certainly are other possibilities. You could take them to a landing page that isn’t an article. You could take them to a page where they have to input their phone number and you send them a text message. So, there are a number of different landing page flows that are possible. In what cases is the landing page with an article the best? And when might some of the other flows be better?

Violeta Pirnog: It all comes down to it being a native-looking ad. So essentially, when you have a landing page, it’s very clear what the purpose is; it’s very sales driven in a very obvious manner. Which is a great angle. However, for this particular flow, where you’re serving people an article, it’s a piece of written content where you’re not outwardly asking them to convert immediately. It makes sense that it’s a bit more subtle. It’s a kind of soft sell, whereas the landing page is more of a hard sell. Both approaches work.

For example,

on Facebook the placement that works the best is one that looks, quote unquote, native, where you don’t usually find the usual ads or ad placements. The news feed placement is actually the one that works the best. I think that’s because it can be mistaken for a post that someone may do, or a brand made, that is organic.

For this organic approach, I feel like the soft sell is what works and continues to work. Whereas for landing pages, it’s a bit more of a hard sell.

Shamanth: So you’re saying, if the product is something that’s more of a softer sell, that needs more education, more information? That makes more sense.

Violeta Pirnog: Yeah, depending on the flow as well. Meaning that also depending on the channel, because I imagine for, you know, a higher intent sort of platform, then you can send people to learn from a landing page. Because they’ve already searched in Google for a similar product, or on Pinterest, or so on. But for Facebook, where you want to have a different approach, and kind of be inclusive to all types of intents, then it makes sense to try to be a bit more subtle about it. 

Shamanth: And I think you talked briefly about placements. And did you say, just the news feed works the best? Is that what you said?

Violeta Pirnog: Yeah, so when I started testing this out, we used automatic placements, of course, to see which one performs the best. It’s always been the newsfeed placement. And also, every now and again, challenging this best practice and coming to the same results. So I think that’s definitely very, very interesting.

Shamanth: And do you find some of the other placements, let’s just say Stories, to not work at all? Because that’s something I find curious because Facebook pushes automatic placements, somewhat aggressively. Why do you think some of the other placements don’t work nearly as well?

Violeta Pirnog: To some extent, other placements do work, but they don’t work when you want to scale them to a certain status. So, you can fight that fight. But essentially, when you know that this is something where you can easily spend a lot more and in an efficient way then of course, it makes it a bit easier. Another type of placement that we found was more successful was the instant experience, if one actually got that format—actually, that’s not a placement, sorry, but a format. 

Because there you can have a similar experience of taking people to a landing page or another article without actually taking them out of the platform. So, if you wanted to start with these kinds of ads on Facebook, that is something that you could use to gauge if there’s a good opportunity for you there without having to go through all of the engineering and the BI.

Shamanth: And then the instant experience or experiences. Is it plural, or singular? 

Violeta Pirnog: I think it’s singular. 

Shamanth:I think so too. Anyway, so an instant experience, what does the instant experience landing page look like?

Violeta Pirnog: You can build this inside of the Facebook Ads Manager to mimic the appearance of the landing page or the article that you have. Of course, it’s a bit more limited in terms of the elements that you can use on it. But essentially, you can add your quotes and your images and your blocks of text and have CTA buttons and have a similar experience to the one that you have in the article.

Shamanth: Would you put in the entire article in an instant experience? 

Violeta Pirnog: Yeah.

Shamanth: Oh, really? Wow.

Violeta Pirnog: Of course, that’s also something that should be tested, because maybe sometimes you don’t need to have a story. But knowing what works, I would start first with what I know works, and then see if that can be improved on.

Shamanth: So the user goes to the instant experience, scrolls through the entirety of the article. And then there’s a CTA that takes them to the App Store page. I think that makes so much sense. I can see the parallels between instant experience and the landing page.  And I’ve seen a lot of success for instant experience in e-com. So what you’re seeing makes so much sense, even for a product that’s not e-com, but could benefit from that sort of softer sell. 

Another aspect of all of this that I’m very curious about is how testing works. Because, certainly, there is an additional variable in testing here, because you could be testing different ads, or different articles, you could be testing a winning ad with a bad article, and that could just lose, or you could have a great article, but the ad may not be the best for it. But couldn’t or wouldn’t do so well. So how do you approach testing with these sets of variables, articles, and ads?

Violeta Pirnog: Testing is always an interesting topic. And the way that I generally do it is taking a proven article that we already know has a good conversion rate. And then if we already have an image that works, then that one as well. Essentially starting from the strongest elements that we have for conversion.

The main piece of creative here is the article. And then, testing out different images and titles to go with it. Because the title and the copy and the combination of title and images are what gives you the CTR. And you first have to reach a certain CTR. Obviously, it’s always good to have a high CTR in order to send them to the articles. So, I would first start with a strong article and then test out different images and titles.

Shamanth: And does Facebook’s dynamic creative work well? Where you basically give them twenty images and let them do their thing.  Or 20 different ad texts or headlines. In this case, because the headline and the image do need to be specific to the article, do you find the dynamic creative tends to work well, or what’s been your experience?

Violeta Pirnog: In my experience, even though it will be so much easier to just feed all of the potential combinations in there, it isn’t the most effective approach. In my experience, it’s always been the sort of manual, separate testing that has yielded the better results. But this is one of those things that should be challenged periodically, in case you can somehow automate this.  I’m all for that.

Shamanth: Right, right. And so from what you were saying earlier, the landing page is the key creative asset. The image and the headline are the hook. They’re important, certainly, but they’re the hooks. You take a winning article, then you try to maximize the click through rates to the winning article. Which makes sense, because the article is doing the heavy lifting of converting, and the headline and the image, they’re eliciting the click, but they’re not actually convincing or persuading the user. So that makes so much sense.

You guys also run campaigns, very significant ones, on other paid content channels, Outbrain and Taboola. How does the testing work on Facebook as compared to Outbrain and Taboola? Do you take a winning article from Outbrain and Taboola, and roll that to Facebook? Do you find completely different things work? What’s the interrelationship between these platforms?

Violeta Pirnog: Initially, we wanted to have different flows and essentially different angles for the articles that would go forward within the platforms.

As we were testing a lot of different articles on all of the platforms, we realized that it’s the same ones that work for Outbrain, Taboola, and Facebook.

Given that that is the theme, generally, where the processes of the creation of all these pages can be simplified. 

There’s a lot of work being put into it, and into creating the best converting combinations of elements in those pages. There’s extensive brainstorms, where we don’t always try to come up with new angles or titles for the entire contents as well. What is happening now is that, initially, I wait for the initial test for Outbrain and Taboola. And then the promising ones, I will also go on to test on Facebook, in terms of articles. In terms of images, I would test maybe the same or different images, because it can happen that there are different images that work for different platforms. But that’s also kind of stuck very close together throughout the years. 

Shamanth: Interesting. It surprises me. Because you would think algorithmically, the kinds of users that read a news site would be very different from those that look at Facebook. But it sounds like the kinds of people you’re attracting for the product are roughly similar. 

Violeta Pirnog: Yeah, it’s definitely been a surprise.  The budget I would do for testing these things would then go towards the actual spend of the usual campaigns, instead of continuously trying to test entirely new ideas, just resorting to having one process for creating these for Facebook as well.

Shamanth: Right, it’s certainly much more efficient that way. 

In your time running and testing a lot of these native ad campaigns, what have been some of the bigger surprises that you have seen? Are there other examples that you can share?

Violeta Pirnog: I think most things to do with the best practices have been surprising, including the fact that the audiences react similarly on Facebook compared to Outbrain to Taboola. 

The fact that it’s always, all the time, been the newsfeed placement that works the best.

One thing that is surprising, though, is that it’s usually images of people that work the best. I think there’s some psychology around it. And it’s also within angle. So it’s the front angle where maybe you don’t necessarily want someone staring right back at you. But you can see the eyes. I think I remember reading something, probably on LinkedIn, about the psychology of this, where instantly, just because of the way that we’re built, whenever we encounter another pair of eyes – even if we’re scrolling on Instagram – we’re just instantly attracted to it as humans.

I guess back in the day, we had to look out for danger, or for enemies. Or for animals like elephants. So now it’s being used differently in that it’s used as a hook to get your attention on these platforms, to check out traffic.  

Shamanth: Interesting. It’s not just an abstract psychological theory, but you actually see that when you show more humans making eye contact, it works well. And would you say there are video ads that work better than just the static photos, for a product like this? Or does it not matter? What has been your experience?

Violeta Pirnog: I would say it depends on the product. I don’t think there’s one size fits all. But what I can tell you from my experience is that it’s always been static that has worked the best. Also, for the longest time, even though I attempted some half illustration, half real photo tech creatives, it’s always been real images that have worked. My hypothesis is because it’s in line with the whole native appeal. Most recently, I’ve seen that change a tiny bit. What we did was just add a bit of branding elements to the usual top-performing photo. That has seemed to work as well. In the end, this creates quite a smooth experience, because then you see the same kind of elements then on the page that you’re taken to and then into the App Store. It creates this sentiment of brand familiarity,

Shamanth: Yeah. I also sometimes wonder if static images can work well in a native context. Because users know they’re not going to get any interactive content. If you put a video, they may know there’s going to be an app, or there’s going to be some amount of interactivity. So videos work well. And in this case, they’re going to have to read, so, not making it flashy makes it actually work well. I wonder if that’s perhaps one reason as well.

Violeta Pirnog: It’s very possible. You know, in marketing we talk about best practices. But the reality is that they change because people get used to one thing, and then, at some point, the opposite works. So, back in the day, when we just were starting up on ads, things that were very in your face were working. But then, people understood that their attention was distracted. So then, now, it’s stuff that is maybe more low-key that is working, and so on. So from that point of view, it makes sense. 

Shamanth: Just from having seen your ads, I also get the impression that the headline also matters very significantly. Because the static can just be a stock image, or certainly it can appear that way, even if it isn’t. But headlines can be a bigger hook. And I’m just curious if that’s really the case that you see in your data and your performance as well.

Violeta Pirnog: Yeah, so about the stock photo aspect of things. Yes, it’s easier to… Well, it’s actually not easier to use stock photos because when you know exactly what you’re looking for, you’ll find that it’s actually very difficult to find in all of the databases out there. So, we did a lot of our photos in-house. A lot of the people featured in our ads are colleagues. So it’s also nice to see articles featuring your workplace, where people who actually work for them are front and center. 

When it comes to the title, that’s definitely something that’s very, very important. So I think there’s an equal amount of testing in the content of the article, and also matching it into something catchy. Because this can definitely make or break if someone clicks through, or not.

Shamanth: Yeah. And the ad title generally has to be different from the article title. So that these are two different things that you’re testing at different times. Is that accurate?

Violeta Pirnog: Actually, what I found was that that was not necessary. So you will have the same ad title. And then it’s the title of the article as well. Sometimes there’s variations, but there’s not been an uplift in trying new ones.

Shamanth: So, the repetitiveness isn’t a problem. If anything, it’s reinforcing that, “Okay, this is exactly what I came to see.”

Violeta Pirnog: Yeah, I don’t think it’s a problem. Because essentially, you’re flipping on an ad that has a title that promises to talk to you about something and then it just simply reappears. And it just confirms that that’s what you’re reading about.

Shamanth: It certainly makes sense, Violeta. This is all so fascinating. Certainly, something I will go back and listen to. I feel like I’ve learned a lot from you today, just as I have in the past as well. This is perhaps a good place for us to start to wrap, Violeta. But before we do that, can you tell folks how they can find out more about you and everything you do?

Violeta Pirnog: Yeah, of course. Follow us on Blinkist, on all the Blinkist socials. And the Blinkist dot com. You can download the app. And you can always find me on LinkedIn.  

Also on Soundcloud sometimes. 

Shamanth: Yeah, we can link to all of that. Yeah, on SoundCloud is that your music?

Violeta Pirnog: It is my music.  I’m not competing with you for podcasts.

Shamanth: It’s not a contest, but you’re a DJ just as well so people can check that out. Check out your SoundCloud just as well.  We will link to all of that in the show notes. But, Violeta, for now we will let you carry on with your today. It was a pleasure having you on the show.

Violeta Pirnog: Awesome. Thank you so much. Have a great week.

A REQUEST BEFORE YOU GO

I have a very important favor to ask, which as those of you who know me know I don’t do often. If you get any pleasure or inspiration from this episode, could you PLEASE leave a review on your favorite podcasting platform – be it iTunes, Overcast, Spotify or wherever you get your podcast fix. This podcast is very much a labor of love – and each episode takes many many hours to put together. When you write a review, it will not only be a great deal of encouragement to us, but it will also support getting the word out about the Mobile User Acquisition Show.

Constructive criticism and suggestions for improvement are welcome, whether on podcasting platforms – or by email to shamanth at rocketshiphq.com. We read all reviews & I want to make this podcast better.

Thank you – and I look forward to seeing you with the next episode!