Don't miss
  • 2,266
  • 6468
  • 6097
  • 20

ARPPU in freemium games

By on November 23, 2011
Print Friendly

 Averages are dangerous things. Most people hear “average” and assume a normal of Gaussian distribution. This can blind you to the reality of the situation. To take a practical example, imagine 20 people you know and imagine lining them all up in a row. Their height is likely to average out at around 5’7, depending on the male/female split. Now add Bill Gates to the line. His presence is not going to change that average very much.

Now rearrange those friends according to their net worth. Add Bill Gates back in, and his immense worth will create a massive distorting effect to the average. So much so that unless you have 20 very rich friends, the average (by which I mean the mean) will be essentially meaningless. Most media people think in terms of normal distributions. If the price of a game “averages” £15 over its lifetime, total revenue = number of units sold x average price. The entire industry therefore fixates on number of units sold.

In the free-to-play world, revenues don’t follow a normal distribution. They follow a power-law curve. (Or, to put it another way, the handful of whales – people who distort the spending average in the way that Bill Gates distorts the wealth average – are critically important.) My post on Whales, dolphins and minnows – the beating heart of a free-to-play game covers this concept in more detail. You may also want to read Whales, true fans and the ethics of free-to-play.

What ARPPU should I use?

How should you estimate Average Revenue Per Paying User (ARPPU)? In the GAMESbrief online game spreadsheet, I’ve approximated the power-law by splitting the users into three categories: minnows, dolphins and whales. Real whales can spend an enormous amount of money. Social Gold reckons the highest group of spenders have a lifetime value (i.e the amount they spend over the time they play your game) of over $1,000, with some spending over $20,000 on a single game. Flurry, meanwhile, says that on iOS and Android in the US, the average transaction value for an iOS or Android in-app urchase is $14 and 51% of revenue generated from in-app purchases comes from transactions of over $20Superdata research has whales as the top 15% of spending users and dolphins as the next 25-40%. Whales generate 50% of the revenue, while minnows – users who spend between $1 and $5 – generate 15% of a game’s revenue.

I recommend that you use the following splits to approximate the curve:

  • Whales: 10% of payers, ARPPU of $20
  • Dolphins: 40% of payers, ARPPU of $5
  • Minnows: 50% of payers, ARPPU of $1

These are dependent on your game. Not just which platform or genre, but how you design. For your whales to reach an ARPPU of $20, some of them must be spending over $100 in your game. Is this possible? Your dolphins need to have a good reason to keep spending a little bit of money each month. Have you created one? Your minnows need to be converted from freeloaders to buyers. What will make them make the jump?

Different types of IAPs generate different levels of revenue. For example, Roger Dickey of Zynga estimates an ARPU (note: not ARPPU) of $0.03 for energy mechanics, $0.02 for decorative factors, and $0.05 for competitive gameplay.

In other words, this spreadsheet is not just a forecasting tool. It is a design tool. If your design team don’t understand the financial model behind your business, they won’t be able to design a game that will deliver against it.


Because few people break out the split between whales, dolphins and minnows (not least because it is an arbitrary approximation to a continuous power-law curve), I have included a derived stat in the spreadsheet of overall ARPPU. As you refine your spreadsheet, you may want to sanity check this derived ARPPU against the stats below.

Trevor McCalmont from W3i says, “we’ve seen typical figures from $5 to $20, but of course there are games with an ARPPU below $5 and others that exceed $100. As with conversion rates, titles with very high ARPPU usually do not have mass market penetration.”


  • Superdata gives an ARPPU of $21.60 for US mobile games, and $32.46 for Chinese mobile games. (source, May 2014)
  • Playnomics sees their ARPPUs ranging from $32.01 in Europe to $270.55 in MENA (source, January 2014)
  • The average paying Candy Crush player spends $40 a year (source: Pocketgamer)


  • ngMoco: $1.50 (ngMoco estimates it makes $0.03 per DAU per day, and 2% of DAUs spend each day. I estimate this equates to $1.50 ARPPU per month)
  • Tiny Tower: $10 (source: Pocketgamer 15/12/11)
  • Funpark Friends, Tag: £6.50 (Source: Pocketgamer 1/12/11)

Other Mobile

Darya Trushkina said at Login conference 2012 that there is an Android game with an ARPPU of over $100.


  • Bigpoint: [$10] (Bigpoint says its ARPPU is equivalent to that of World of Warcraft)
  • Puzzle Pirates, Three Rings: $50 (Gamasutra)
  • Unnamed collectible card-game that uses Superdata: $55 (source: Joost Van Dreunen, Superdata)
  • Habbo Hotel: $30 (Engage Digital)



Reloaded’s Bjorn Book-Larsson stated at GDC 2012 that their benchmark for a free-to-play RPG ARPPU is $40-100, and for a FPS game $20-50.



The ARPPU for social games in Japan for 2011 was estimated at $42.18, and is projected to reach $53.03 in 2013 (Source:, 23/6/11)
Brazilian users on average spend $10.29 per month on mobile games (source: Darya Trushkina speaking at Login conference 2012)

About Nicholas Lovell

Nicholas is the founder of Gamesbrief, a blog dedicated to the business of games. It aims to be informative, authoritative and above all helpful to developers grappling with business strategy. He is the author of a growing list of books about making money in the games industry and other digital media, including How to Publish a Game and Design Rules for Free-to-Play Games, and Penguin-published title The Curve:
  • Pingback: Chinese mobile games’ ARPPU is over $30, and other new benchmark data - Games()

  • Pingback: How much does an American spend on mobile games? Less than a Chinese person - Games()

  • Pingback: Don’t release a 99c game #f2ptoolbox - Games()

  • Pingback: What’s an impressive conversion rate? And other stats updates - NativeX()

  • Pingback: Daniel Parente Blog » Plantilla de cuenta de resultados de un videojuego freemium()

  • UnknownSubject

    And yet the article is still very relevant.

  • RichS

    The thing you are missing are at a minimum the fees, fines and costs associated with the chargebacks from people using stolen credit cards to purchase the virtual goods. These costs can add up to be a significant expense. Most people don’t realize is this fraud is typically performed by a very few people using multiple user accounts. If you can figure out the very few bad guys and attribute the costs to them, they have a very high negative ARPPU.

  • Nicholas Lovell

    Good point. Fixed. (A year later, but still…)

  • Nicholas Lovell

    I’m not sure I seee how this is “negative” ARPPU. Since we are selling virtual goods, the specific loss is negligible. They can inflate your ARPPU figures, sure, until your accounting team strips them out, and malicious behaviour is bad for your community. But negative ARPPU? I don’t see that. Am I missing something?

  • RichS

    I would suggest another class of users following along the aquatic theme. Sharks these are users with a high negative ARPPU because of their fraudulent payments and malicious behavior. These “sharks” can have a negative ARPPU of hundreds or thousands of dollars.

  • Anonymous

    If the ngMoco revenues are $0.03 per DAU per day, and 2% of DAU spend each day, how can we get a $1.50 ARPPU per MONTH? Shouldn’t it be per day? Or am I misunderstanding something?

  • Pingback: What is a "good" RPDAU for a social game? - Quora()

  • Ayhan

    check the glossary for definitions…

  • Pingback: Evidence of ISD()

  • Pingback: How many downloads? | Nerd Gamers()

  • Pingback: Social Games Summit Stats | Nerd Gamers()

  • UnknownSubject

    I’d be careful of the overall ARPPU of APB since “purchased time” could (and probably did) include box cost factors.

    Plus at a glance I can’t see the term ARPPU defined anywhere on the page (or the other pages I checked, either). Although you are talking to an audience you’d expect to know, but it doesn’t hurt to define terms like “average revenue per paying user” if that is the chief topic of discussion in an article.