How much do guilds matter?

Today, we have another in a continuing series of articles written by the highly talented Sanya Weathers for GamerDNA. In this, her latest column, Sanya runs down some interesting and illuminating statistics about players and guilds, and just how they fit in to the macrocosm of MMOs.

A long time ago, in a fantasy world far far away, I wrangled guilds as part of my job. At first, I wrangled them because it was terribly efficient for one person struggling with a beta. Why deal with thousands of individuals when I could deal with fifty, and put them in charge of their own groups? Guild leaders will always be more informed, more tuned in, and better suited to herding their own cats than a studio representative could ever be.

But what started as expediency turned in to more.

As a community weenie, I had my dearly beloved "frequent flyers" – people who sent in feedback, usually through email no matter how many systems I built for them to use, with a regularity previously reserved for clocks, robots, and dripping faucets. If it was 3 PM on a Thursday, it was time to get mail from him, him, her, and what I was pretty sure was a him but hadn't ever asked.

And of course, whether the player was happy with my employer or so furious he was frothing, the time would eventually come where they would all move on to other games or hobbies. If a player has gone beyond the free trial, they usually stay for a period of months, and then move on.

My frequent flyers in active, engaged guilds stayed on for far longer than my frequent flyers who were unguilded or in casual guilds with indifferent leadership. So I started keeping track. Nothing fancy, because (/old lady voice) in my day we didn't have fancy metrics, whippersnappers (/old lady voice).

Still, I found that players in guilds that had regular events, strong communication outside the game, and charismatic leadership typically subscribed for six months longer than players who did not boast such social connections.

Go ahead, do the math in your head, I'll wait.

I know! Isn't that a crazy, absolutely crazy amount of money? I can barely count that high, and yet I dearly want to learn how.

So, here at GamerDNA, we're exploring ways of proving this connection with something a little more powerful than a scratch pad and an Excel spreadsheet. Today's column is the first of several such pieces.

First, let me say that in today's data, we cannot tell the difference between an active, engaged guild and one that is basically a pickup group that shares a name. Future columns will draw that distinction, we hope, but today's simply differentiates between the guilded and the non-guilded.

Today's column is written from data that we started gathering in November of 2008.

Also, in order to be sure that the randomly selected "pools" were the same size, we had to stick with World of Warcraft. Because of GamerDNA's origin as a guild tools site, many of our original members signed up at the request of the guild leader who had chosen us as a host. Only WoW has a big enough population that we could randomly select a thousand people to be in our "unguilded" pool – a few other games might have offered that many unguilded players, but it wouldn't have been a random selection.

Here we go!

(Ed note: Click on images for larger versions.)

Right away we see a difference in something key to player engagement: the time they spend in the game. Everyone saw their average hours peak during the release of the expansion. Unaffiliated players put in less than two minutes more than guilded players did – the one and only time that was true.

The winter holidays allowed the two groups to draw close to one another, and the week of New Year's saw them get as close as they've ever been besides the Lich King launch week – guilded players put in less than ten minutes more than their brethren who stick to pickup groups. I'm guessing that the holidays allowed players who have no social ties in the game (ties that would cause them to prioritize the game more highly in their daily activities) to spend more time logged in.

The guilded and unguilded haven't been that close since. Guild averages dropped during the week of the 3.0.8 patch, and solo logins rose that week, but the guild averages were still higher than the solo. Generally, those with social ties in the form of a guild put in between a half hour and an hour more every week.

Here's another standard measure of player engagement – their level, and how quickly that level rises.

The average level of the guilded players in our sample was 66.7 when our data gathering period began in November. The average level of the unguilded players was 60.5. In that first week, both groups rose in level at a rate so similar as makes no never mind: the guilded player increased by an average of 0.089 of a level, and the unguilded player rose by 0.085.

The difference a guild can make is apparent by the jump during the expansion launch week. Guilded players went up an average of 2.57 levels that week, whereas the unaffiliated players – who, if you'll recall, put in two minutes MORE time that week, only went up 0.85 of a level.

Which is still kind of staggering at the 60+ level, but that just shows you how much support matters in a massively multiplayer game.

Guilded players continued to advance a hair faster than the unguilded players for the next two months, but by the end of January, the two groups were roughly even in their leveling speed. Unguilded players occasionally leveled faster, if less than a tenth of a level counts as much faster.

Interestingly, the expansion – widely hailed as vindication for the non-raiding, non-uberguilded player – seems to have widened the gap between the two groups. They were six levels apart before the expansion, and are now separated by ten levels.

This one is fascinating, I think. The OOC channel in any MMO often hosts discussions of what classes are best for soloing. Lately, I've been forced to play almost entirely solo, but my preferred class – tank – is never the one the channel experts advise newbies to choose. Hybrids inevitably get the nod, because of their ability to do a little bit of everything. My own arguments for choosing a tank as a solo – great armor, decent damage, plus sufficient hit points to handle multiple aggs without the kind of timing that mezzing or fearing requires – get mocked.

Well, guess who's laughing now. Booyah.

Though I admit that my own private guess for the most likely to fly solo was the Rogue. Assassin players sure seem the type to prefer working on their own in the shadows, but I guess that's a stereotype. Unguilded rogues have numbers less than 2% more than the guilded. The unguilded class of choice in World of Warcraft is warriors by a long shot.

So what have we learned?

It's not just anecdotal evidence, anymore. In WoW, and I think any game built with similar mechanics (which is practically freaking everyone coming down the pike) any guild membership at all increases engagement as measured by time played per week, levels achieved, and leveling speed.

In the months to come, we at GamerDNA think we're going to prove that it's not just engagement that rises with guild membership, but subscription length as well. We will also be looking at the differences between casually formed guilds, and highly motivated close knit guilds. So stay tuned!