Low conversion in free-to-play is a feature, not a bug

Posted on May 12, 2014 by Eric Benjamin Seufert


An article excoriating the supposed plague that is free-to-play made the rounds recently, citing engagement statistics from Swrve’s April report in support of an argument asserting that, because average engagement rates for free-to-play games are low, and the percentage of players that pay anything within free-to-play games is very low, consumers clearly don’t enjoy free-to-play games. From the article:

Recent data shows 20 percent of mobile games get opened once and never again. 66 percent have never played beyond the first 24 hours and indeed most purchases happen in the first week of play. Amazingly only around two to three percent of gamers pay anything at all for games, and even more hair-raising is the fact that 50 percent of all revenue comes from just 0.2 percent of players.

This is a statistically insignificant amount of happy gamers and nothing that gives you a basis to make claims about “what people want”. I think it just as likely that mobile’s orgy of casual titles is due to simple bandwagon-ism or, in other words, not knowing what people want.

The problem with this line of reasoning is that ratio metrics (and, more broadly, industry benchmarks) are meaningless for free-to-play games absent the context of the entire metrics portfolio or knowledge of the game’s marketing mix (ratio metrics are obviously important from the perspective of the developer, but presented individually to outsiders — eg. game X’s conversion rate is Y% — they reveal nothing about a game’s overall performance). The calculus behind beginning the development of a game under the freemium model (and the decision should take place that early) revolves around aggregate numbers: total addressable market, total daily active users, total daily revenue, etc.

Ratio metrics become relevant as the total addressable market decreases as a result of changes to the premise of the game (ie. if the game has 20% fewer users, but they’re more targetable with marketing and monetize 30% higher…), but these changes are optimizations and should take place only after the core mechanics and format of the game have been established. The fundamental decision of whether or not to pursue free-to-play shouldn’t entertain concerns over conversion rates; by definition, conversion is low in freemium.

Abstracting this argument away from games and onto consumer technology in general, freemium isn’t meant for niche products: it’s a business model designed to benefit from scale, network effects, price differentiation, and varying levels of consumer enthusiasm, with the confluence of these factors producing more aggregate revenue than if the product had adopted an up-front price. Conversion rates for freemium consumer technology products are low across the board:

  • In its IPO filing, Skype indicated that only 6% of its users had ever purchased premium talk credit;

No freemium software developer releases a product expecting high conversion rates; low conversion in freemium is a feature, not a bug. The goal of a freemium product is to drive more aggregate revenues through a much larger user base than could be achieved through a paid download / install pricing strategy. This is facilitated by high virality, a small but highly-engaged minority of users that is willing and happy to spend large amounts of money on the product, or both.

Even if one accepts the notion that the 66% of players that don’t return to free-to-play games after Day 1 didn’t like those games, it’s impossible to claim that they were offended or angered by them. More likely, those games simply didn’t match the tastes of the affected players; do people suffer longstanding negative emotions when they don’t like the taste of a free sample at the supermarket?

Contrast this with users who paid to download games they ultimately didn’t like (it’s impossible to estimate the number of players for whom this holds true, but it’s certainly more than 0%): they probably did experience anger, or at least regret, which means they took negative emotions away from the exchange.

Why would the potential to arouse those negative emotions — waste, bitterness, disappointment — be preferable to presenting a player with the opportunity to make an informed decision over the games on which they spend their time?