Have you noticed that there are a lot of analytics companies in the game space recently? They offer pretty looking charts and require you to integrate an SDK. This article tries to answer the question of why many of them seem to become advertising companies.
For example, the following companies started as game analytics providers and then very quickly added advertising to their offerings:
- Flurry - Analytics and AppCircle Acquisition
- Apsalar - Analytics and Advertising
- iQU - started out as a online game analytics system targeted towards MMOs, and now they are a game advertising network
- Kontagent - originally a social app analytics outfit, teams up with Tapjoy to offer mobile advertising
- Localytics - originally a mobile analytics company, now offering acquisition advertising.
- Playnomics - originally an advanced machine learning company, now offering advertising to get the best players likely to monetize
These companies are holding out for now, will they change?
- Sonamine - advanced player predictions (disclaimer, I founded this company)
- GameAnalytics - a Denmark based company offering dashboards. (Updated 2014-08-18) As predicted, no pun intended, GameAnalytics started to offer advertising services, to help mobile game developers acquire new users....
- Lvlboost - the improved version by the makers of Playtomics.
Data needed for advertising (Supply side argument)
One hypothesis is that perhaps advertising is made more efficient by using data. On the face of it, this seems like a plausible argument. If the ad networks know that user U is a monetized on game A, then they can show ads to this user for game B. And the game B developer would be willing to pay a lot of money for user U.
If we let different game developers bid up the price for user U, then the middle-man advertising business would be a booming one.
Cult of user acquisition in free to play (Demand side argument)
Businesses do not spring up in a vacuum, so what's driving these game ad networks? There's clearly money to be made. The obvious answer would be that free to play game developers and publishers need to "get" users. The huge influx of new gamers, casual or mobile, has drawn a lot of investment and created a land grab mentality.
These same F2P game developers are trying to find a way to monetize the "non-paying" users in their games. As a result, they are lured by the idea that they should show advertising to these non-paying users.
The need for short term growth, fueled by investors
Advertising to get new users costs money, and this money is provided by investors. In addition, the investors in the game analytics companies are also demanding faster growth. Investors are looking for revenues to ramp up to $100M in 3-5 years - "venture scale".
If we consider the economics of the analytics companies, they start out free (flurry) or about $200 per month. If they manage to get 10,000 game developers to pay $200 per month, that is a $24M annual revenue business, hardly a venture scale business. And this assumes one gets 10,000 paying game developers!
On the flip side, if they get 100 game companies with $1M user acquisition budgets, then a 10% ad network cut means $10M in revenue. Much easier to sell and scale.
This type of economics drives some of the migration towards becoming an ad network.
Facebook advertising Google Plus, never
Whatever the business drivers of the new game advertising ecosystem, if we take a step back to reconsider, then some of the second level issues start to crop up.
Recently I was playing Glu's awesome new Samurai versus Zombies Defence 2, and up popped an ad for Supercell's Clash of Clans. These are not exactly dissimilar games so the rationale for advertising your competitor is that you can monetize the non paying users.
Although not a perfect analogy, a game that has advertisement for a competing game is similar to Facebook allowing ads for Google Plus. This would NEVER happen. Similarly, you will not see Amazon ads on ebay. Even considering the idea would be a career death sentence.
Imagine how the stock market would react if ebay announced that they were monetizing non buying visitors on their website by sending these users to Amazon!
The truth about user cannibalization
When a monetized user of a game sees ads for other games, and starts playing the other games, we call that cannibalization. You sacrificed perhaps $20 in lifetime value of that user from microtransactions or in-app purchases, in order to obtain perhaps $0.40 in revenue share for that "ad-conversion". This is the most painful type of cannibalization. We at Sonamine have discovered that at least 10% of monetized users will click on cross-promo ads and churn out, before reaching their full LTV. Game developers should look for your own numbers.
But there is a another type of cannibalization that is more insiduous. For the user that is just about ready to convert to being a paying user, losing these players has a large opportunity cost that will never be recovered. Unless you use Sonamine's ConvertSoon score, you would not even know who these users are.
No developer we have worked with wants to tell the ad networks who the monetized users are, even to "not show them the ads". The reason is that ads can still be targeted to these users in other apps, just not within the games of that developer.
How do I keep my own data and users?
Open sourced game analytics packages such as http://count.ly are the natural evolution of the ecosystem. Developers can easily within a few minutes get their own analytics package up and running, keeping their own data and users private. Google analytics will always be a viable option, especially with their new app analytics SDK.
Does the long tail data really help?
So why do these analytics companies offer things for free or a low cost? One answer is mindshare and to "get" the data. The premise is that by getting the data, they will be able to offer a more compelling and efficient advertising offering.
As is always the case, deeper examination reveals the problems behind these assumptions. Game play traffic follows the same pattern as internet traffic -- a scale free network. In other words, the bulk of the traffic flows to a relatively small number of websites, or in the case of games, a handful of game developers. The "long tail" of game traffic, made up of small indie games and single game developers, contributes about 30-50% of the overall game traffic. And when that is spread out over many games, the "signal" in this data is marginal. We at Sonamine have analyzed this for web based games and found very marginal improvements in the 10% range by adding this long tail data to our user predictions.
The fat tail is made up of large games like WoW or Clash Of Clans, companies that would never give out their data to be merged with other games.
Building a business on data that is not your own
So we come full circle to the original question: why do most game analytics companies become ad networks? Because they are trying to build a middleman business that grows quickly using the data that game developers give them. If you read their TOS carefully, they usually say something like they have the "right to retain, use, and publish" (Apsalar), or the "right, for any purpose, to collect, retain, use, and publish" (GameAnalytics).
Is this premise a viable sustainable one? Let's look at some internet examples. Google Search crawls the web to make it easier for consumers to find relevant content. Advertisers target ads to keywords based on these searches. In this case, Google is not selling users of a particular type (eg. monetized), and one could argue the data is actually public source as the websites, are by definition open to the public.
NDC Health Info or IMS health purchase prescription data from the pharmacies, repackage and resell them. This type of aggregation was a valuable service in the past, since the collection and collation of paper prescriptions was manual and expensive. It is arguable whether such a aggregation service could be created in this internet connected age. Note too that just 5 companies (Caremark, Walgreens, Walmart, Riteaid and Kroger) employ 75% of pharmacists in the US. Another scale-free network. In games, App-Annie follows this model.
My personal belief is that game companies will realize that their data is their greatest asset, and they will build their own analytics capabilities. User acquisition is going to be a continual problem, but the days of easy in-game cross promo will be over soon.
Contact me at nick_at_sonamine_dot_com