Making Sense of Ad Networks

Ad networks, waterfalls and mediators and attribution partners

In an earlier post about SDKs, I wrote about ad networks. This post will expand upon that ecosystem.

It was likely simple once. You had a game or website with a lot of traffic, and someone wanted to show ads to all those eyeballs. So you did a deal. Like this:


Now there are three layers in between. At least.

The problem with direct sales are the communication channels. There’s more than one game with a lot of players, and there’s more than one company that wants to advertise. If everyone is talking to everyone then they all waste huge amounts of time just discussing and cutting deals. If a thousand advertisers talk to a thousand game developers, then there are a million deals to be made. Enter the ad networks.


The ad networks take care of this problem. All the advertisers just talk to an ad network, and all the game companies also just talk to an ad network. The ad network has made some sort of standard deals that they offer everyone to be efficient. Usually it’s some sort of auction that they are running.


Now, with a thousand advertisers and a thousand game developers, there are just two thousand discussions taking place, and they are mostly automated by the ad networks systems. Clearly this is useful and saves a lot of wasted effort. It’s so useful that more than one company got this idea! Now there are a few hundred ad networks around.

From the game company’s perspective, it is now a question of what ad network to use. Some of them have a lot of advertisers but pay only a little for each ad you show. Others have fewer advertisers but pay more for each ad. As a game company, you can make several times more money if you do the right choices with how to handle the ad networks.


The solution is to integrate more than one ad network into your game. Your game will then say to the ad network that “here’s a spot where you are allowed to show an ad”. If the network has a suitable ad to show, it is sent over, and they log what the user did (watched the video, clicked the link, or ignored it completely). Your game can basically ask the highest bidding ad network to show an ad first. If they have one to show, that’s great. If not, you go on to ask the second highest bidder if they have one, then the third and so forth down the line. This is called the “waterfall”.


Unfortunately, it’s a lot of work involved in optimising your ad revenue. Who has the best paying ads is always changing, and is different depending on what country your player is in. Which means that you have to have several waterfalls and constantly monitor and update them. Fortunately, there are now companies that offer this service. A mediation platform basically does this for you. There are several around, some run by ad networks (they swear on their grandmother’s grave that they do not favour their own ad network), and others from independent companies.

At the advertisers end, there is a similar problem. There are lots of ways to advertise your service or app, how do you know what’s the best one? You could be running small campaigns on each possible channel, measure how they did and then decide where to spend the big bucks based on that. This requires quite a lot of skill, math and time to do. Fortunately, there are companies who can do this advertising for you. Companies such as Fiksu and Twigeo are there for you when you want to outsource this job to professionals. To keep track of where you’re getting your installs, you need an Attribution partner, such as Adjust, Tune or AppsFlyer.

Now we are left with this picture of the market:


Currently, I’m mostly thinking about which network mediator to choose when we’re trying to make money from ads. You know, there are several, and we want to optimise between them…

And I have the same problem when wanting to buy ads. Someone really should offer me a service to choose between the service providers here…

Let me Spoil Every Movie for You

All Blockbusters have the same script! What can a game developer learn from that?

There are some good books to read if you aspire to become a screenwriter in Hollywood. A bunch of them all say the same thing: there’s a pretty fixed structure to the script of a blockbuster.

Let me spoil most movies for you now: when everything is finally going just great for the hero, pause the movie and you will most likely find yourself at the midpoint. When everything thereafter falls apart (bonus points if some minor character dies), pause again and your will be at ⅔ of the movie.

There are a number of books that will teach you the classic story telling model. It’s called “The Hero’s Journey”, and you have hopefully heard about it if you’re developing games with a narrative component (ok.. let’s say “story” instead). As a reminder, it goes like this:

Act 1:

-The story’s world is introduced. It’s in a stable state, we’re introduced to the characters.

-The theme of the movie is introduced. Hints about what this is going to be about.

-Something disturbs the stable world

-Our hero should go on an adventure, but hesitates

-A mentor comes in and gives our hero a swift kick in the behind, and the hero gets going


Act 2:

-Love story & side plots are introduced

-Our hero is trying to adjust to the new world and new rules. Likely feels like a fish out of water. Most of the movie’s trailer clips are from this part.

-MIDPOINT: our hero has finally learned the new rules and is victorious (but it’s a false victory)

-Problems appear, stakes are raised

-ALL is LOST! Something dies.


Act 3:

-Fresh inspiration allows the hero to find a new way

-At the midpoint, our hero had learned the new rules, but forgotten about his/her real self. With a new combination of the original personality and the new things learned, our hero is finally victorious for real.

-Ending with how the world changed as a result of the adventure. Everyone grew and learned something (except for the bad guys). Contrast the new world with the world at the beginning.


There you go: just about every Hollywood blockbuster works according to these rules. You can use it to annoy friends and relatives by telling them what comes next. If you want to learn more, here are some books about it.

Entertaining books:

Save the cat!

How to write a movie in 21 days


Serious (=Dry and Academic):


The Writer’s Journey

The Hero with a Thousand Faces


However, the point I’m getting at here is not the straightforward idea that game stories should be just like movie stories. Rather, it is that a related art form has developed some rules for what works. Rules that have had a remarkable staying power. You see, this structure does not only underpin movies, but stories overall. Joseph Campbell abstracted The Hero’s Journey by looking at stories in literature throughout human history, from early myths to recent books and movies. He found the popular ones so similar in structure, that they basically tell variants of the same story. Human psychology evolves very slowly, and therefore certain patterns appeal to us over millennia.

Likewise for other art forms: there is music theory and there are colour and composition theories for visual arts. The real point is that there is likely to be some design patterns for successful games too, and we need to discover them. We need to understand where to innovate, and what parts are so fundamental that fighting them is futile.

On mobile Free-to-Play (F2P) games, I believe that designing for short sessions, several times per day, every day for months is a fundamental thing that will not change (see the number of the beast post for more). How specifically to achieve that end is still likely to evolve.

We might also see more stories in mobile F2P games, but it’s hard to produce hundreds of hours of storyline cost efficiently. If we do see stories, I think they will likely adapt the TV series version of the Hero’s Journey. That is, they run several of these story arcs overlapping. There’s always one story about to wrap up, one under way, and one just starting. Some arcs are longer, some shorter, etc. F2P games and soap operas are both never-ending.

This blog is about what we have so far learned about making games in general, and mobile F2P games in particular. When reading about ideas and best practises, please think about if the practises presented are only the current fad, or are they the fundamental Hero’s Journey rules of our art form.

Marketing vs Market Engineering

-Free-to-play is a more honest business model than the traditional premium model. And that has consequences for how you market.


-Let me explain!

With the traditional pay-up-front “Premium” model of selling games, the game company or publisher has an incentive to build up expectations before the launch. They need to make people desire their product. It’s all about the customer expectation. What happens after the customer bought the game is less relevant. If they never even play the game, or test it only once, that’s fine. They already payed for it. I do this myself all the time with Humble Bundle games. I like to think of myself as a person who has the sophisticated taste to enjoy a fine indie game in my spare time. The problem is that I do not have spare time. So I end up buying the games, but never playing them. Which, as I pointed out, is just fine for the indie developers. They got paid for building up the expectations.

Free-to-play is wildly different. The clue is in the name. A good F2P game will let you play the game for free forever. It is better to keep you playing for free than having you not play at all. You might, after all, tell your friends about it, and they might tell a friend who tells a friend who ends up paying. Since you can play for free for several months, before eventually deciding to pay, the players know exactly what they are paying for. And therefore, expectation building ahead of a game launch is much less relevant.

A way to think about F2P is that we are giving the player unlimited try-before-you-buy. We don’t optimise for expectations up front, we optimise for retention of the players who ended up trying our game. If we can keep them, they will slowly spread the game to friends, who spread it to more friends and eventually we end up making money. Have a look at the resulting revenue curves for successful games in both models to see the difference. GTA V and Fallout 4 both made a lot of money in their first week. After that, the sales numbers quickly drop. Candy Crush Saga and Clash of Clans are the opposite. They have made similar money over the games’ lifetimes, but it is much more spread out in time. I looked at the public sales data for GTA V and Candy Crush Saga, and sketched out roughly how their sales were distributed. The curves here show how both GTA V and Candy Crush Saga made about $3 billion in the last 2 years. As far as I can tell from public data, they should be very close to each other in terms of total revenue generated over the last 2 years. That leaves out the original ramp-up of CCS, of course (it was released in April 2012, about 1.5 years before the start of this data series). While the totals are about the same, the distribution in time is very different. GTA has peaks at original launch, and when ported to new platforms (XBox One/PS4, and PC). Candy Crush Saga, on the other hand, is very

If you want the premium curve of high expectations leading to quick revenue at release, you need to do traditional marketing. Creative stuff, buzz building, talking to press. A good game will certainly be needed to convince people that you are worth the buzz, but what’s a good game will be decided by press and early players in a matter of days.

In this mode, you have a lot of activity based on best practises, but they are hard to measure exactly. How much did that magazine interview do to help you? Was it more or less effective than that huge poster at Gamescom? It’s hard to know, but that is marketing. A traditional saying is “I know half my ads aren’t working, but I don’t know which half!”

Free-to-play marketing is different. Just about everything you do is measurable. You have different ads and measure click-through rates on them to optimise. You try different channels and and measure cost-per-install on each. And most of all, it’s not about a huge effort concentrated around the launch date, but about a long-term process that can run over several years. My friend Thorbjörn Warin who has been at Wooga, Grand Cru Games and AdColony had a nice way of saying it. You market premium games, but F2P is not really marketing – it’s market engineering.

If you’re an indie developer working on a new title for Steam, you want to do marketing. The guerilla kind, but anyway. You speak to game media, you’re active on discussion forums, you try to attract Youtubers, etc.

If you’re doing mobile F2P, you might as well not bother with that. Our Benji game gets tens of thousands of downloads every day. A news article about the game usually has such a small effect that we cannot even see it when looking at the download charts. The chart below show tens of millions of downloads for Benji over the last 2 years (the drop in April is just missing data on Appannie). This is market engineering territory.2yearDownloads

Ironically, at the very highest levels of user acquisition budgets, you’re back from market-engineering to marketing. Just look at the TV ads of King, or Supercell’s Super Bowl ad.

At their level, they have maxed out what they can get from ads in other mobile apps. Who hasn’t already seen several ads for the top-3 grossing games (Candy Crush Saga, Clash of Clans, Game of War)? That means that they need to bring in fresh users to the ecosystem rather than take users from other mobile games. But, if you’re the guys making these TV ads, I don’t think you need my advice for how to do user acquisition for mobile games anyway.

Premium on Mobile Today

Premium is making a comeback! Or at least some people hope so..

There are a few companies with an interest in keeping premium alive. That might be good for indie developers.

As everyone knows, the mobile market shifted to free-to-play (F2P) games a few years ago. Back in the Angry Birds era, people seemed happy to pay $0.99 for hours upon hours of entertainment in their phone. Then, a few years later, everyone collectively realised that paying such princely sums for mere entertainment is clearly unreasonable – at least before they thoroughly get to test play the game before being asked to pay. During 2011-2013, the market shifted so quickly that by the end of 2013 between 90% (iOS) and 98% (Google Play) of the revenue was from F2P games.

Since then, there have at times been talk about the comeback of premium. Unfortunately, the data does not support that claim. When looking at the top charts, things have moved even more clearly in the direction of F2P dominance. The current top-100 grossing chart on iOS is 99 F2P games, and Minecraft. It has been pretty consistently like that for the past 2 years. Some hit games that have an existing brand outside of mobile (Console titles, Steam titles, etc.) can briefly make it to the top-100 list – think 5 Nights at Freddy’s or Grand Theft Auto. But the only one that has stayed there is Minecraft.

On Google Play, it’s even more clear. The top grossing game without IAPs today is This War of Mine at position 269. The next is GTA San Andreas at position 363, Geometry Dash at 416 and Star Wars: KOTOR at 431. Of these 4 that are in the top-500, only Geometry Dash was a mobile-first game. That is, one single pure premium mobile first game in the entire top-500 grossing list. the others who got there without using IAPs all had brands from outside mobile.

The best premium games today can make revenue of some millions of dollars. This is for games that get selected by Apple as Game of the Year, such as Monument Valley and Badland. While that is by no means bad, these games also required a lot of talented people to develop. For instance, Monument Valley cost $1.4 million to develop, and generated $5.9 million.

That’s a bit more than 4X return on investment. And that’s the best that you can hope for with this model. In practise, it means that every 4th game that you make needs to be Apple’s Game of the Year.

No investor who can count is going to put their money on such a business. Which means that if your company has VC money, you make F2P games. No discussion! Which in turn might open up some opportunity for indie companies.

You see, there are some companies who would still like to see premium succeed. Apple for one. They are a very premium company and the pay-once-up-front model appeals to them. Premium apps also have the upside that they increase the cost of switching between iOS and Android phones. Your F2P game can be downloaded for free on the other platform, and often you can just keep on playing as if nothing happened. With a premium game, you have to pay again – thus raising the barrier to switching.

Closely allied with Apple in the quest to rescue premium on mobile is the games media. The people who write for games media are often hard core gamers themselves (otherwise, how would they have ended up in that job?). And hard core gamers usually don’t have warm and fuzzy feelings about F2P monetization. In addition, it’s just in their self interest to promote premium games. After all, if I’m required to pay up-front with real money for a game, I want to at least read a review before paying. Which gives games media an audience and a revenue stream. If, on the other hand, I am curious about a F2P game, I will just try out the game instead of reading a review. Goes a long way towards explaining why lots of the top grossing F2P games are completely ignored by the likes of And also why TouchArcade is in financial difficulties as a result of the rise of F2P:

For a large game company, the strategy is clear: develop only F2P games. For a small indie, there might still be a niche for premium. You know you won’t be competing with the big budget companies, and you also know that you have a few good allies if you do premium. A tradeoff of a way smaller best case revenue (by a factor of almost a thousand) versus much less competition could still make it worthwhile. Just remember to keep your budget reasonable. And you might still end up like Zombie Match Defence, if you don’t get a good featuring by Apple.

Or, you might have tremendous success, like the guys who did The Room. They’re another game of the year, though.


Even more reading about the topic:

I’m on CNBC!

CNBC make a short article about the Finnish tech industry – games in particular – and interviewed me for it as well. Over here:

And, I’m pretty sure Tero is ironic in his commentary:

Going Global

Apparently, this blog is attracting quite a global audience. It’s getting read by a few hundred game developers every week quite consistently so far. And have a look at the map – you come here from just about everywhere in the world. Readers in over 40 countries so far!

Thanks and enjoy!