Aarki: making the most of interactive advertising

Candy Yang Aarki

Interactive ads are seen as high-quality ads by the users themselves, and result in higher engagement. In this episode of AppInTop Mobile App Marketing Podcast we talk to senior director of business development Aarki Candy Yang about interactive advertising and Aarki Studio. Listen to this podcast episode or subscribe on iTunes.

With Aarki, mobile game developers and brands can create interactive ads for different platforms simultaneously then measure engagement and return of investment across web and mobile platforms.

What’s the advantage of Aarki Studio?

Mobile space is about personal engagement and interactive ads are missing.  Aarki Studio helps engage those users on a more personal level.

When you are looking at your mobile device and you see a video, sometimes that engagement level is not as deep as what people are hoping for because a lot of the users are desensitized with respect to videos. By adding interactiveness into the ad the level of engagement increases.

Let’s say you have a mini-game version of your app placed in an ad.  Now users can interact with your ad unit. They can play the mini-game or they can view a funny video clip and interact with that clip in a different way.

Our clients are very engaged in the process of creating an ad.  There’s a complex review process we go through because some of them are brands with strict definition of what they want and what branding aspects they want inside these interactive ads.

Can you give a couple of examples of ads you have recently created?  What you would consider to be the more successful cases?

Some of the ads that we’ve put together were for Zynga and EA.  The Zynga team has been using our Studio very actively for brand engagement within their apps.

For example, with Words With Friends we’ve had a couple of very interesting ways of delivering the ad units as well as with Draw Something.

We have these ad units that look almost like the native apps themselves so that the branding is almost as if the user is playing a part of the game.  But in the background the brands are allowed to express themselves.

In an ad for Honda we had a 360-degree view of the car within games like Words With Friends.  And then there’s a scrabble game on top of the car.

What about utility apps or business apps? Do interactive ads work for them as well?

We have not tested that yet. We really do believe that there is value in those apps as well.  Maybe the format would be a little bit different than for gaming apps.

We find that the creativity level is truly very high-ranking in terms of engagement and when there’s a ‘wow’ factor to interactive ads. If a really ugly-looking ad shows up in front of you, you are probably are less likely to engage with it.

What’s the process of creating an interactive ad?   

Creating a mock-up and a storyboard is actually very effective.  It’s just as if you were creating a commercial where you have to lay out what you want the users to do.  Afterwards we piece it together.

If we get special customer requests we would entertain that too as long as it still falls within the deadlines.  Generally, we would just provide suggestions of what we’ve seen that has performed very well.

To what degree can an interactive ad be tested before it is run?

It’s really up to the advertiser.  We always suggest test campaigns. Sometimes these ads are run on a specific date. This is especially relevant for brands because they want the ad to coincide with their commercial or product launch.

If we talk about testing the ad effectiveness, to see what ROI an app install campaign generates, small test campaigns are very effective.

In the near future we’re planning to put in more A/B testing and a dashboard that will give an in-depth view into performance of the ads.

Currently, Aarki Studio does have a analytics dashboard but it is a little bit more basic.  You can embed pixels and you can test the effectiveness of the interaction rates between the first page and the second page or the third page.

These pixels return rapid fire postbacks. You can look at it and see, “Oh, so maybe users like this button a lot more.  Let’s keep the button and remove everything else because nothing else interacts.”

Then you make changes and then send them to a lot of ad networks in one click so you analyze your ad performance and update in real time.

You enable publishers to target users.  Where do you get data for the targeting?

The publisher actually gives us guidelines on where they want to target users.  Aarki Studio just provides the platform to generate creative ad units.  The targeting is done by the publisher.

They can use third-party trackers such as HasOffers – any tracker that allows embedding pixels or tracking URLs.

For example, if you put up an image and you want to target certain users for that image, you embed the tracking link.

Aarki Studio has its own pixel-tracking. For example you want users to download the app.  You want to track the downloads through HasOffers, while we can embed those pixels that track your URL right into the ad.

When you’re giving it to ad networks or other partners to run your ad, you just give them a simple ad tag, which has everything embedded inside.

Is there a future for CPI ad networks?

There is definitely a big future for CPI.  I don’t think CPI will ever go away.  It’s just a matter of delivery.  Marketers most basically want to know how much bang for their bucks they’re getting.  And CPI is the most effective way because then they don’t take on any risk at their end.

What is the current market CPI for games, productivity apps, business apps?

For gaming, it’s a very highly competitive market.  A couple of years ago it went crazy when Gree & dNA said, “The CPI for a holiday should be 3 to 4 dollars.”

Nowadays, CPI could be as low as a dollar. It depends on if it is incentivized or non-incentivized.  So I would say CPI between $1 and $2 is your happy medium, and $3 is where you get your quality users.

For utility apps that could range greatly because of the categorization of the limited number of users that you can target.  You can probably get around $3 as well.

How do you split your traffic between different platforms?

I would say that a lot of people are using Facebook and Twitter.  That is definitely a large source.  But that is not the full hundred percent because you can never get enough installs just from Facebook and Twitter alone.

There are many users that might not use Facebook or might not even see your Facebook ad because Facebook shows a lot of other ads.  And also it’s not very interactive so people usually miss those kind of ads.

Facebook and Twitter probably account for 60 percent, if not a little bit more, of purely non-incent installs.  Incent is totally different monster.

About 40 percent would probably come from other sources such as recommendations, blogging, other ad networks, and other partners that you might have direct contact with.

But not all developers have Facebook and Twitter embedded. Some users just correlate with or respond better to rewarded units. Facebook and Twitter do not offer that.

Having a brand that’s related to rewards can sometimes boost effectiveness of the ad. This is because your users are more likely to remember receiving a reward from a sponsor than remembering where they got the install from.

Do CPI networks have their own traffic?  And what’s that share compared to third-party traffic?

For CPI ad networks, they basically have their own publishers that they run.  For example, those ads that you see in EA for example. They have a list of offerwalls.  Aarki is actually included in some of those.

We have our own publishers that we work with very closely to provide offers.  CPI networks on iOS are definitely not allowed inside the apps.  But on Android they are.

In terms of the other 40 percent, I think it is a mixture of having own publishers or apps for example.

In Asian countries they definitely own a lot of their own applications to show this traffic.  For example Line is a very big CPI network in Asia.

I am a young developer who’s working on a game and I want to look for traffic. What are some of the other places to look for traffic?

Aarki is definitely a great network to advertise on. We have interactive ad units. You can create your own video which helps the app go viral. Our ads also work on browsers.

Other ad networks, too, because you want to diversify your traffic. I won’t mention names because they’re competitors.

Other places are real-time bidding networks.  A lot of the inventory has been consolidating in these networks. These networks proved to be a great source of CPI traffic because they have a lot of inventory you can test out.  Unfortunately the billing is done on a CPM basis.  So not a lot of these guys actually provide a CPI into a CPM or a CPM into a CPI.

What do you think the biggest concern or challenges facing CPI and the model for the future?

The biggest concern is if there’s any changes in Apple or well… Apple more likely than Google. iTunes Store is changing.  There was that scare about incentivized installs.

And then the second biggest concern is that a lot publishers might not want to take the risk for the developers in actually giving them a CPI model.  Obviously there’s always a business aspect otherwise it wouldn’t even make sense for networks to sell ads on a CPI basis.

And there are other factors that play into CPI such as user lifetime value or retention rate.

So you might see CPI give way to CPE – cost per engagement or some other more interesting or more in-depth model where you’re actually testing for lifetime value or the buying power of those users that you’re getting.

The volume of users you can get through CPE-type ads might be too small simply because CPE means a deeper level of engagement than just an install.  And you simply don’t find a lot of networks being able to provide CPE without the proper targeting capabilities.

On the other hand you should not only optimize based on lifetime value. If you are a user, you want to play with other people and not necessarily with paying users because the game becomes very boring when everyone beats you (in a pay-to-win scenario).

What do you think about Apple fighting incent traffic?

It’s going to be very challenging for them unless they change the algorithm.  And I think that there’s enough people to protest that they won’t do it immediately.

Maybe they will change the algorithm to the point where incentivizing it isn’t as relevant as, say, user review or the session duration but I don’t believe that they’ll make incent traffic go away.

What about the concept of watching video ads in exchange for virtual currency?

A lot of people watch commercials.  I don’t see any problems in incentivizing users to watch videos.  What I do think is wrong is incentivizing them directly to force them into watching or doing other actions.

But for Aarki it doesn’t really affect us that much because for us we believe that our videos are actually engaging.  So whether we rewarded the user or not, incentivizing is more about getting the users to watch the video rather than having the users interact with the videos itself.

I believe that if we make creatives engaging, whether it’s rewarded or not becomes irrelevant.

I don’t think that Apple is going to affect the incent model as much as they think because at the end of the day a lot of the users cannot pay. That’s why the incent model was built in the first place. A lot of users are just little kids.  They don’t really have a lot of money in their pockets.

There is a lot of value in rewarded units because otherwise there’s not going to be as many users and that would affect Apple’s ratings and usability.  They’re already losing quite a lot of shares to Android.

If the incent model went away that would cause a drastic decline in the Apple user base.

In Asia Android is more popular because there is a choice. But an iPhone is an iPhone.  There’s no flexibility there.

Another thing is, you really can’t find anything in the Apple Store.

There’s a lot of buzz right now about native ads. What is your view on that?

Native ads are very important but I don’t think that they look better.  The good thing is that you retain users in your app as they install an app from a native ad. It’s a plus both for the publisher and the advertiser.

And I don’t think that it’s going to change.  It’s not hype.

Do you have a success story that you can share in regards to native ads?

A lot of Aarki’s ad units are native.  What we find is that native experience does engage the users a lot more.

You have user engagement because these ads feel like part of the game. You definitely see users respecting the quality of the ad units.  They don’t look tacky or out of place.

I like to use the Zynga example. We placed a Kaiser Permanente ad on Draw Something. With Draw Something you could pick certain pictures.  And you can draw on that and you can post it on the social feeds like Twitter or Facebook and then show it to your friends.  And then you would go back into the ad unit.

Another one that we did similarly was Spider Man.  It’s the same concept but you can draw an amazing Spider Man on the ad unit itself.

We find that once we find a successful unit, we give the advertiser the flexibility of exchanging creatives for these native ads.  And they can just replace them with whatever their branding is.  And it works very well.  The engagement level was amazing.

Everything is templated so that they can quickly bring it back up and use it again.

What about your anti-fraud algorithms, how do they work?  And is there concern about fraud in video ads?

Basically we’ll look at a spread of information that we received from the publishers. If something looks very suspicious like a consistency in the device for the same ad over a period of time then we flag it as a fraudulent user. We would toss that user out and ban them from receiving anything.

And for videos it doesn’t apply as much because usually we have frequency caps for the videos. It’s usually more or less on whether the user quality is any good.

Is there anything that advertisers can do to fight such fraud?

I would suggest they use third-party tracking especially HasOffers and Appsflyer and AD-X.  These are the major three that are really very good at weeding out fraud.

Obviously on their own side advertisers can monitor the users and pass back users that have already received the reward or that are using the game.

A tracker can state ‘here are users that you should not reward because these users are either, a) tagged as fraudulent, or b) have already received the rewards.’ This would greatly reduce the amount of re-spend of the same campaign especially when you’re working with multiple networks.

What is the future of incent ads?   And are advertisers going to make a full-on switch to native?

Incent ads are not going away. Incent volumes are simply a lot higher.  And it’s something that users like and I don’t think that Apple will remove it because users really do enjoy having rewarded units.

Some advertisers will switch to native. It’s not really just the advertisers’ side.  It’s actually the publishing side that’s more relevant to the native ads because these publishers do not want these ad units to disrupt their user flow.

So if you were to say who is going to switch first, I would say that the publishing side that host all these ads would switch to the native side more or less because the publishers see the value of native ads.  Advertisers only care about the ROI but publishers care about the aesthetics.

But they won’t have the same type of native ads as what Aarki is providing to Zynga right now.

These publishers will probably make movie theater boxes and host the videos inside in a frame rather than say, what Aarki is doing for Zynga which is creating ads that look native.

Name one thing that needs to change in mobile marketing that will make your job easier.

If we can get to a targeting level similar to Amazon in being able to predict what users might want, that would make my job a lot easier. Also if we can recommend ads that we know users are most likely to look at instead of what users, the audience type, are inside an application.

This would cut down not only the advertisers’ ROI spent, it also makes my job a lot easier because I don’t have to go back to advertisers and say, “Well, we met your target but their lifetime value might have suffered a little bit.”

I would love for us to get to the level of targeting capabilities.  At the end of the day, because of privacy issues, we might never fully achieve that.

This is an abridged and edited interview from the AppInTop mobile app marketing podcast produced by AppInTop, an automated mobile app marketing platform. Listen to this podcast episode or subscribe to the podcast via iTunes.