Gangnam Style Insights, Chartboost Tips, CPI Trade-Offs

“True popularity comes from acts of kindness rather than acts of stupidity. ” – Bo Bennett

November and December have been busy! Man oh man. I’m up to 78 apps live in the store now with 9 more in production (much bigger apps) that should roll out in the next two months. As my growth slowly began to compound, I had some really cool stuff happen. A whim of launching a Gangnam Style app, changing a few things in Chartboost, and setting up some new deals has really changed the game for me. Having the opportunity to speak at Chad Mureta’s App Empire event was also a highlight and has shown me how much passion and fire there is out there. All great stuff.
In proper fashion, I wanted to share all of this with you guys. Well, as much as I could at least. Sometimes I think that it’s impossible to get all the ideas and findings down onto paper, but the AE event really got me in gear to start helping everyone more.
So, here’s what’s up.

Gangnam Style Runner – The Little Source That Could

Looking at my apps page and you probably see that a bunch of my games are around the runner source code I have. Some of them have done really well (Tiger, Panda, Elephant) on their own, but definitely nothing outstanding. I consider an ROI of 200-300% a really good investment. I think some people are looking for 2000% returns in the mobile market and that’s why they get so disappointed.
The news came out about Psy’s video beating out Bieber (thank God) for the top YouTube video of all time. Knowing that YouTube is the world’s second largest search engine, I was willing to bet that people were searching the app store as well. I had my guys whip something up in a few days and I published it. Shit, I didn’t even bother making screenshots because I thought it was a joke!
I decided to release it in most countries outside the USA just to see what would happen. Even though I knew this was kind of a longshot, I thought it might blow up like others had. I was right. The first night it went out in most countries except the top 8 and collected more impressions on RevMob than I had on all other apps combined that day. Woah.

I released it the next day to the masses and things went bonkers. 150K impressions on both Revmob and Chartboost each day, and revenue to match. Check out what my CB account looked like! I hadn’t really used Chartboost much before (only made  me $200-300 a day) but what this exercise taught me was very important.
What’s even more crazy was that the Flurry analytics basically showed me that people were playing this game 5-6 times. This wasn’t a game that people picked up then dumped – or if it was, they got their fill beforehand.
So here’s the interesting part.
I was getting a lot of five star reviews and a lot of one star reviews.
Some people thought it was pretty cool or funny or whatever that I made such a ridiculous game, and gave my an iHighFive with a good review. Others cried foul, specifically about the ads. The biggest complaint was definitely about too many ads and them popping up in weird places.
Now, when I released this game, I figured I’d make some money but nothing crazy, so I loaded both Revmob and Chartboost at the same time. Like at the start of the game, it would show Revmob, then Chartboost right behind it. I do this on all my other games and it works out well with little to no repercussions on the review side of things.
But with this, people were getting pissed. They saw this hot new app that skyrocketed to the top of the charts so they wanted to see it. When they opened it, they saw lots of ads. The psychology was “I want to be part of this cool new hot app but this guy is just trying to rip me off or take advantage of me.” Whereas when my other apps went out in the middle of the pack, the psychology was “Hey, this guy is just trying to make a buck.”
That’s also why my eCPM on my mid-level games is $15-20 and it was close to $2 on this app (both RevMob and Chartboost). In fact, I talked to the guys who made the other Gangnam Style apps and they had the same results – high eCPM on their lower traffic apps, low eCPM on this.
Made me realize something very important as we all grow as appreneurs – be sensitive to your volume. Monetization strategy changes dramatically as your popularity increases. If you’re not prepared to adjust, you’re going to lose all that ranking volume from negative reviews.
So I tried turning off Revmob. The result? Less people feeling violated about Ads. And a lot less money for me.
But I was buying myself another day that I would probably stay on the top of the charts. Classic opportunity cost and one that was VERY hard to control since I knew I could make a killing if I just let Revmob go nuts.
Now I had a ton of volume on Chartboost and back to regular on Revmob (which still makes me more money, surprisingly).
What does one do when they now have 300-400K impressions a day to mess around with?
Well, here’s what I did that gave me a 25% bump in revenue.

Chartboost Tips – Control The Machine and Force The Intelligence

Yeah, dramatic title I know, but that’s basically what I did. We all go into ad networks with the same assumption – the marketplace serves up the highest paying bids first, when those run out, the next come in, then the next, etc. This all happens at lightning speed and naturally makes us the most money, right?
Well, not really.
I’ve talked candidly with some big advertisers and they have basically said that it’s impossible to buy enough installs, no matter what your bid. Trust me, if they could, they would. Even at $5-6 a pop, advertisers want the organic rankings bad enough…but they just can’t get it.
So I start thinking. If advertisers who pay all these premium prices are literally flooding the market, why do I ever see a $0.50 CPI ad? Or anything under a $1 for that matter? Now, I’m not saying I know all the nuances of how this works, but I have to imagine that it’s because the marketplace just can’t handle the serving well enough to understand all this. Meaning the low price stuff will always be serving, then it gets trumped by the higher stuff, not the top down model. It’s a better business model to start, but it would also use less CPU power, I assume, help fill rate, generate more advertisers, and incentives for up-bidding the price.
This isn’t good for the average developer/publisher who just wants to display their ads and get the best possible eCPM.
Ever notice this option in the Chartboost dashboard – called Priority?

This is the manual way to do what the marketplace should be doing for your already. It was originally built for Prioritizing Direct Deals, Internal Cross Promotions, and regular Publisher Campaigns, but I figured I might as well test out creating CPI limits of 4 campaigns and then creating priority levels accordingly.
Level 1 – $0.50+ CPI, Low Priority
Level 2 – $1.00+ CPI, Medium Priority
Level 3 – $1.50+ CPI, High Priority
Level 4 – $3.00+ CPI, Highest Priority
What this means is that I am forcing the opposite of what I think the ad networks want to serve. Not perfect, but it’s another way for me to control the ads my apps receive. The top campaign looks for anything with a premier bid, then it moves down from there.
The result?  A 25% increase in my eCPM over the next week. I could attribute this to increase bidding during the holidays, but I honestly think it’s just because more of my installs were coming from higher paying advertisers. I was kind of doing the Revmob model on the Chartboost platform (FYI that’s why RM eCPM is way higher – they do this so well). In fact, the eCPM was very clearly highest at the top tier Level campaign, with the eCPM the lowest on the Level 1 campaign.
So I have  my campaigns optimized. I have a lots of impressions. What’s next?

Intra Network Snake Oil – Build Your Army For Advertisers


During all this Gangnam Style hoopla, I released Jolly Journey – an awesome (in my opinion) kids game for the holidays that I really enjoyed making. It’s my first step towards creating games that have higher lifetime value built on in-app purchases and not on advertiser dollars.
One of my immediate reactions to the Gangnam success was that I would have the ability to drive my own installs now. You may have heard marketers talking about the “power of building your network” for a while now. The idea being you build a portfolio of apps that you can re-direct at any moment to drive installs for your own game for FREE. The sexiness of this lies in the ability to control rankings, etc.
Of course I thought this would be amazing. I had my entire Chartboost account ready to go. I was going to point all my apps at Jolly Journey for a rankings bump and also to get users in there.
But then I stopped for a minute.
I was making over $1,500 a day on Chartboost by selling installs that made me anywhere from $0.50-3, averaging $1 or so.
The minimum bid for installs I could buy on CB is $0.50.
Using my own network to drive my app installs would COST me $0.50!
Now, the assumptions may be that the users would be more likely to download another game by the same publisher, which would make my overall install # higher and thus make my per install opportunity cost drop, but there is almost no factor I can think of that would make up for $0.50.
Having your own cross-promo widget/popup is a solution to this (instead of using Chartboost), but that also means you will be giving up the install revenue even further from residual loss.
And therein lies the rub of it – building your app network is not about being able to drive installs to your other games. It’s about having more revenue positive assets that can be re-invested into more revenue positive assets so that you can buy cheaper installs on the open market and monetize your traffic with the rich advertisers. That’s just how the markets work, at least for now.
In a nutshell, because there are so many rich gaming companies out there willing to spend $5-6 CPI, it drives up the average eCPM for all publishers.
A lesson learned, saving me thousands of dollars.

Final Thoughts

1. Gangnam Style Game (Free and Paid to date, launched 12.1.12) – 925 Paid/192 in-apps in Paid version / 623,344 Free, $2,674 in-apps and iAds,  $17,380.12 on Chartboost / 4.9M impressions on CB / $1,223 on RevMob (this was only about 3 days) / 782K impressions Revmob / eCPM $2.01 for all / 2.0% install rate
2. Jolly Journey HD Free (iPad version launched 12.6.12) – 4,860 free sales / $280 in-app / $277 MoPub & TapIt (banners) / 2.5M requests for banner serve.
The reason I’m putting these out there is to show a few insights I am seeing in the app world:
1. iPad is going to crush iPhones in terms of monetization over the next few years. I’m going all tablet by the end of 2013. Every one of my games that is iPad makes WAY more money. The volume is lower but the market is much less saturated. Plus, I can design way better for iPad.
2. Though the JJ revenue numbers pale in comparison, they’re growing steadily every day. And look at the impression ratio – I have almost 130x more downloads for Gangnam and only 2x more impressions!! People play the Jolly Journey game for an average of 8 minutes. Crazy.
There are four versions of the game, all of which make money, but this one has the most potential. It’s compounding as people play day over day. This is a MUCH BETTER long term strategy.
In my App Empire presentation I discussed how development is the key to user retention. This is a case in point. BTW – working on getting that video for everyone to watch. It’s all about flipping apps and people seemed to dig it.
 
Happy Holidays to everyone! If you’re not a Facebook fan yet, please Like my page and Follow me on Twitter to stay up to date!
Keep rocking.
Carter
 

91 Comments

  1. Tom
  2. matt
  3. Ros
  4. Lev
  5. Ob Oberon
  6. Lisa
  7. Brian
  8. zacharie
  9. Brian
  10. Brian
  11. Sven
  12. K
  13. Lisa
  14. Kate
  15. Mo
  16. Mo
  17. Mo
  18. Mo
  19. Mo
  20. Lisa
  21. Mo
  22. Brian
  23. Lana
  24. Andrej
  25. Charles
  26. Matt
  27. Max
  28. Apps 4 Fans
  29. Matt
  30. max
  31. Lisa
  32. Gerald Lum
  33. max
  34. Hiro Lee
  35. Mike
  36. Steve
  37. Steve
  38. Joe
  39. Shark
  40. Karen
  41. Karen
  42. Pravin
  43. Ross
  44. Karen
  45. Dm3
  46. Henry
  47. Volkan Kutlubay

Leave a Reply to Brian Cancel reply