One of the more interesting applications of the social game “sticky factor” that I introduced last week is the ability to look at the life cycle of some of the most popular games and identify some of the key feature roll-outs that greatly increased engagement. The sticky factor is simply the daily active users (DAUs) of an application divided by monthly active users (MAUs).
Let’s take a look at two of the largest games on Facebook, by DAUs, that have been around over three months and also have a pretty accessible view of product feature launches: FarmVille and Mafia Wars, both from Zynga.
First, it’s critical when looking at the data to not really pay too much attention to the first thirty days of data, especially when trying to compare different games: By definition the MAU requires a month’s worth of data. Also note that because of the growth in Facebook, an application launched in September of 2009 is going to be exposed to a lot more users more quickly than an application launched a year, or even six months earlier. So you can see after launch that the sticky factor had dropped down to around 32% one month after launch (July 21st). Over the following month, you can also see that FarmVille increased its sticky factor to somewhere between 37-38%. Let’s take a look at some of the main features launched in the game that might have impacted engagement:
- On July 24th, Zynga launched its 2nd expansion of the farm (the amount of land you can plow, harvest and decorate) but this seemed to have little impact.
- On August 7th, Zynga launched in-game achievements, with the ability to send a notice to your friends once you had earned the white, red, blue or yellow ribbon.
- On August 18th, The sticky factor was at 37% and Zynga then launched tractors to help make game play a little less tedious on those larger farms
Of these three game feature releases, it would seem that achievements had the biggest impact. And of course achievements generally increase how viral a game is, helping to bring in new users into the fold.
For Mafia Wars, I don’t have reliable launch data prior to March of 2009, so I can’t pinpoint what helped increase the sticky factor from late December 2008’s low point, but we can pinpoint a couple other key feature launches:
- March 26th: Gifting of items is launched, allowing users to send useful in-game items to their “mafia” of friends, greatly increasing re-engagement and driving new users. FarmVille launched with gifting from the beginning, so we couldn’t see it in that comparison
- April 30th: Achievements are launched, again seeing an increase much like we did with the earlier FarmVille example
- The broad releases of the Cuba expansion (June 11), new Cuba achievements (June 18) and a daily lottery ticket (June 25) may have helped contribute a bit to the increase in the sticky factor to nearly 30% over the Fourth of July weekend, but that also may be due to the fact that during that weekend Zynga launched special loot that could only be earned by engaging with the game over that holiday weekend. This looks like it was a huge boost to re-engaging users and has been repeated nearly monthly (the next red dot was Labor Day weekend’s Loot event)
- The launch of the Moscow expansion pack (September 25th) doesn’t appear to have had as much impact, and that may be because these expansion packs are focused as drains on some of the highest level users who are already extremely active.
There is no question that achievements and “gifting” have been enormous engines driving viral growth and re-engagement – the sticky factor analysis lets you visualize the actual impact on engagement those tactics bring and can be used as a benchmark for measuring the impact of future initiatives.
Does a More Viral Game Mask the True Sticky Factor?
One criticism of the sticky factor is that it can be greatly influenced based on how viral a game is – that the influx of new users skew the DAU number. That’s definitely true during the first 30 days (as mentioned in the FarmVille example above) or where there is a large promotion (huge ad buy or a cross promotion from another large game that drives a spike in new users), but after that I think this is somewhat mitigated by the fact that with a rolling 30 day MAU you are not going to see huge swings.
Ultimately, the sticky factor is an approximation of churn – the secret sauce that helps you understand your customer lifetime value and maximize revenues. As a developer, you’d have access to much more granular churn rates with an understanding of the differences of newbies you are acquiring virally and those that have been with you a number of months. Plus DAU really looks at JUST daily usage, where you may find quite a profitable experience with users that play every other day – to that point Playfish’s vice president of product, John Earner, mentioned at the Virtual Goods Conference on Friday that 50% of the users who ever played Pet Society are active each month, and that’s 65% for Restaurant City.
Still, I wanted to look at the impact of a game’s viral rate to drive new users, to see how much that can impact the sticky factor.
As an example, last week I mentioned that it looked like a 15% sticky factor was the breaking point for a game’s success. I took a pretty straight forward model assuming steady viral growth rates (e.g. of the users you have on the service x% will post something or invite a friend that induces another user to join the game) and a steady churn rate (you lose y% of your total users every day through attrition). These are the sticky factor curves based on a core churn rate and a viral rate.
When you have a churn rate of 10%, you easily reach that break-away velocity (the 15% sticky factor), regardless of the viral ability of the game. But as your churn rate increases, you need increasingly more viral pull to reach that 15% sticky factor:
- With a 12% churn, you need a 5.9% viral rate (1 in every 17 users bringing in someone new)
- With a 15% churn, you need a 13.4% viral rate (1 in every 7.5 users)
- With a 20% churn, you need a 23.9% viral rate (1 in every 4.2 users)
Generally, for every 1% increase in churn above 10%, you need a 2.3% increase in the game’s viral rate.
Again, this is a very simplistic analysis but I think it shows that while a game that is more viral can help mitigate churn, when that churn rate gets above 12%, it will take a great deal more viral growth to make it sticky enough to sustain that level of Stickiness.
Is the Facebook News Feed Change Impacting the Sticky Factor and Exposing Churn?
As I noted above, some of the big gains in the sticky factor for Mafia Wars came from gifting and achievements – two tactics that post to a user’s wall and would get replicated across their friend’s news feed prior to the recent changes to the Facebook news feed. Now that the changes have been in force for two weeks (and developers are exploring tactics to get these valuable cross-promotion and re-engagement tools back), what are we seeing?
It’s only been ten days since the change and there is a lot of noise in the data, but roughly:
- FarmVille and Mafia Wars actually look like they are up since the change, but there were a great deal of Halloween promotions (specifically limited item drops and Halloween design contests in Mafia Wars and FarmVille respectively) during the last week that may have temporarily re-engaged users.
- Newer games (Café World and Roller Coaster Kingdom) are still coming off their launch highs and hadn’t found a stable sticky factor prior to the news feed change, so it’s hard to attribute the declines solely to the news feed at this point.
- YoVille had been on a decline in its sticky factor prior to the news feed change and appears to have stayed on a similar trend since the change, so again, hard to attribute it solely to the news feed changes.
- Texas Hold ‘Em had a fairly consistent sticky factor prior to the change and appears to definitely be feeling the impact of the change, trailing off in the last week.
At this point, the sample size is too small to really tell if the news feed changes have had an impact. Because viral tactics do support re-engagement, I would expect that higher churning games will see their sticky factors decline at a slightly faster rate over the weeks ahead. Until then, we can continue to look at the sticky factor as a strong indicator of user engagement and as a metric to measure the impact of initiatives on moving the engagement rate.
Eric von Coelln was the vice president of marketing at Oberon Media, a leading multi-platform casual games company, and most recently the vice president of Marketing at PowerSoccer.com. He is now a New York based freelance consultant to games, e-commerce and social media companies — including some of the largest social gaming companies on Facebook. While Mr. von Coelln does write about some companies for which he has done paid consulting from time to time, this post is based on publicly available information and in our view is an unbiased analysis of the industry. You can find his blog here.