20 Jan 2012

SWTOR: The Retention Rate Question

Earlier today I was listening to Talk of the Galaxy over on Galactic Public Radio, when they reminded me that the month of game time bundled with the full version of the game was about to expire and that subscriptions were just about to kick in. There was a little bit of debate about how many players would actually maintain their subscription after the end of the month, with the show hosts generally feeling that it’s a difficult thing to predict.

Trying to forecast renewal rates is pretty challenging at the best of times, mainly because there’s very few benchmarks on these figures for MMOs. The most common figure that’s announced is the total subscriber base, although sometimes we’ll hear about how those subscribers are distributed by market. We’re also not given churn rates, so we have no clue about the number of new subscribers coming in versus the ones that are leaving.

So what exactly is there to go on? Well, assuming that the best renewal rate BioWare can hope for is 100%, let’s have a look at what will help them keep that number high and what’s dragging it down like a lead weight.

The Pros

Kicking off with the most obvious, BioWare organised two very large beta weekends just before the game launched, inviting everyone who’d shown interest in the game to download it and try it out. This had two major benefits – it helped persuade people waiting for a 10 day “free trial” to preorder the game, but it also convinced people who didn’t like the game to not pick up the box at launch.

Content starvation is also a critical thing. One month is enough time for someone to hit level cap on one character if they’re playing regularly. For those being more relaxed about it or who have their attention split between games you’re looking at twice that. Plus there’s the reroll factor – people hitting 50 on one character only to roll and start working on the next.

There’s also more subtle methods that BioWare have used to retain subscribers, like requiring a payment method or gamecard before activating the game. The reason behind this is simple – players are more likely to cancel the subscription if they’re prompted to enter their details at the end of the bundled month.

And what about competition? The heavyweight rival World of Warcraft still doesn’t have any new content, and in fact Blizzard decided to make the latest Dragon Soul raid even easier for players to complete. While this might help raiddgroups get over a difficulty hurdle, I can’t see it being a great lure to pull people back. “Hey guys! You know that frustrating boss we were stuck on? He got nerfed! You can come back now!”. For anyone who wanted to see the content, LFR has given them exactly that.

And what about the single player market? Skyrim took an arrow to the knee, Mass Effect 3 is out in March (although the demo is going to destroy Valentines Day) and Diablo 3’s eruption from hell has been delayed. MMO and RPG gamers could be left in limbo until either Guild Wars 2 or The Secret World comes along to pick them up.

There’s also the news that anyone who bought the game and buys a one month subscription or game time card will qualify for the in game title ‘Founder’. While it’s not on the same level as a companion pet or weapon crystal, it’s an added incentive for people to stay in the game till patch 1.2 arrives.

The Cons

The Old Republic’s Achilles Heel right now is end-game content. It exists and it’s built to a high standard (the new Kaon Under Siege flashpoint is incredibly good fun) but it’s very difficult to get in to at the moment. With a population spread widely between the early levels and the cap and many guilds straddling between SWTOR and Warcraft, finding groups is an exercise in frustration. A standard population server might only see one or two groups formed through general chat channels in a 2 hour window. Unless you’re in a large guild populated with level 50 characters you’re going to get very bored very quickly.

There’s also been a couple of observations that have done BioWare no favours on the last day of subscribing. First was stock analyst Todd Mitchell coming out and cutting his price on EA stock, citing concerns with SWTOR’s sales performance and first month play observarions. Mitchell’s no newcomer to the videogames industry, having looked at their performance since at least 2006. Then there was the issue players started having with the unsubscribe page, again drawing attention to the date and fuelling conspiracy theories.

Also, just because Warcraft is asleep at the wheel doesn’t mean there’s no competition for SWTOR at all. As Sypster and Tesh reminded me on Twitter, Star Trek Online has just gone free to play. If you’re prepared to trade Lightsabers for Phasers then this might be a smooth (and cheap!) move. Alas, force powers are not included.

The Prediction

We know things are good but not great in SWTOR at present. We have an analyst cutting predictions and problems with the accessibility of end game content. But by the same token, we have a market with very little competition, either from newcomers or existing stalwarts. So what’s the figure?

Looking at other industries, churn rates for other industries tend to be fairly low. Verizon’s wireless subscriptions tend to have a churn rate of around 1.5%. Netflix use a rolling monthly contract and their churn rate recently rose to 4% following various blunders this year.

Meanwhile, analysts have historically pegged Warcarft’s churn rate at around the same as Netflix with 4-5%. That article’s from nearly 5 years back though, so it’s unclear if this figure is still accurate. There’s also some research from Nick Yee (a research scientist at PARC). His own research (PDF link, ‘Building an MMO With Mass Appeal’, pg 292)  indicates that “only 40% of new subscribers remain in a game for more than two months”.

So what’s the number? With the work that BioWare have put in before the game’s release coupled with this being month 1 and not month 2, I’d put the retention rate at somewhere around 60% to 70%, but probably on the lower end of things. After month 2 I’d expect another drop to about 40-50%, in-line with previous MMOs. This means that from the 2 million in sales at launch, SWTOR will end up with a stable base of about a million players from which to grow. This is broadly in-line with other predictions from a year ago, although we might not agree on the final number of subscribers.

Just like Chris mentioned over at The Daily Blink, subscribing to an MMO once the bundled month is up isn’t an automatic thing any more. With the current economic climate many of us are feeling the pinch and tightening our belts. Although I’ve made the decision to maintain a subscription to both SWTOR and WoW for now, there are a lot of gamers who can’t make that choice. Which one will work harder to earn our cash?

Like this? Try these other related posts:

Tags: , ,

17 Responses to SWTOR: The Retention Rate Question

  1. Rodd says:

    Question for ya Gaz: do you think a more robust space game along the lines of Jump to Lightspeed would have raised you retention predictions?

    • Gazimoff says:

      I don’t think it would have hindered things, but personally I think that the time could have been better spent tightening the interface and social experience.

      My first priority would be ensuring that end-gamers had easy access to end-game content – currently L50 flashpoints (normal or hardmode) are a non-starter if you’re not in a large guild. PvP queues for the L50 bracket are 2-3 hours at peak times. These all need to improve to give those who aren’t fans of the alt-reroll something to do.

  2. Ravious says:

    Great post. Very well written and thought out.
    Ravious recently posted..[GW] Things You Get Used ToMy Profile

  3. ellori says:

    “…the month of game time bundled with the full version of the game was about to expire and that subscriptions were just about to kick in.”

    “…requiring a payment method or gamecard before activating the game”

    I’m pretty sure they were actually required to sub for at least 1 month right off the bat before they could even activate the freshly purchased game. I remember seeing Stabs raging about that. http://stabbedup.blogspot.com/2011/12/swtor-great-subscription-scam.html

    If so, then wouldn’t it mean SWTOR’s time of reckoning for retention rate, so to speak, wouldn’t be after 1 month, but after 2 months? Since everyone that’s playing was presumably forced to sub at least 1 month on top of the free month?

    • Greygamer says:

      Although they required the payment details, you are not required to subscribe for an additional month.
      If you do not wish to subscribe after the 30 free days, you can simply cancel your subscription and you will not be billed on your renewal date

  4. ~Kate says:

    Well, I for one, will be one of those retainees! Love SWTOR! The drop-off of the wow-babies not be a bad thing for my fave game. Those of us who love the immersion and stories will just game on!

  5. Douglas Stewart says:

    Great post Gazi 🙂

    I’m more concerned over turnover versus Lucas IP costs (whatever the hell they may be, but we can be sure considering past franchise costs they are ‘Lucadriss’) factoring into the calculation of EA deciding if the game is worthwhile to them or not. EA aint exactly known for their tenderness when it comes to aggresive speculation affecting their market share, as the news revealed yesterday.

    If we base loyalty on content expansion, the statement by Bioware in today’s preview vid said they had been working on the Rise of the Rakghouls content for over a year, indicating to me they are aware at least of deployment time and how voice acting delays factor into it. A good sign to me.

    All I know, personally, is me and the Lady are keeping our subscriptions and having an immensely good time in the game. Might help we do not give a toss about PVP though 😉

  6. Steve says:

    This is an interesting read, well put together. I agree with most things.

    I am wondering about the prediction though because it seems to assume that subscription numbers will remain fairly stable after the first two months.
    As pointed out however, SWTOR has very little competition at the moment. So wouldn’t more drops be expected as large titles are released in three to six months time?

    Also, a drop of 50% in subscriptions after two months could be accurate. I have no evidence to suggest anything else. It is also quite substantial. It seems that the problem of finding people to do end-game content with will exist for quite a while. Although more people reach level 50, losing half the population will most likely compensate for population gains at level cap made.

  7. BenBos says:

    The average retention rate of every MMO published in the last 5 years was clear cut.
    60% after 2 months, 40% after 4 months and 20% beyond 6 months.
    In these times of free to play stuff, it is even getting harder to obtain those figures.

    And SW TOR is not exactly considered as a “real” MMORPG by the hardcore players. More a kind of single player on line game with very few things to do in the end game.

    In fact the retention rate in the first weeks are… problematic. You see a 20% degree on player numbers on sample sites like XFire, BUT … this is in a time that weekly sales are still around 100 K NEW added players. So instead of a decrease of 20% players (and 30% activity) watching server loads you HAVE to take into account these latest weeks are still a “BUILT UP” phase .

    If you loose that kind of stats in the building up phase of things, you’ll meet problems later on when your game will sell the usual 30-40K per month copies.

    That’s why a 400-500K sub by summer time (and after D3, GW2 and WOW MOP) would even be a huge success, rendering the reaturn on investment impossible before mid 2013 and after 2012 anything can happen to this subscription base market.

    • Gazimoff says:

      I’m curious where you got those percentages from. Is there a new study or research that’s presented these? I’d be interested to know, thanks!

      • BenBos says:

        Xfire stats showed a max sample of 11.600 Xfire players (and 2 times Xfire 11.200 players) on the first 10 days.
        The last sunday that was down to 9.400. I took that figure because it was a sunday (today it was 8.200, but that’s a week day).
        That shows a 20% decrease in player numbers for a NEW game, that still added 600K NEW players (or 30%) between the first peeks and the last days over the last 3 weeks.

        This decreasing player activity is also tracked on the offical website: http://www.mmo-junkies.net/statistics/
        Normally when you would add 500K-600K new players over the first 3 weeks of a game you would see a small increase or break even point (as players play less), but a DWONWARD trend of at least 20% while adding masses of new players in its launching week is VERY problematic.

        It shows a lower than expected retention rate.

        • Gazimoff says:

          There’s two problems with these stats.

          Looking at xfire first of all, it’s showing activity and not subscription. Playing less once you hit level cap but maintaining subscription is normal behaviour for an MMO. Xfire also has a user group bias to particular types of player – its not exemplar of the entire subscriber base.

          By the same token, the MMO-Junkies stats are indicative of server load and not subscriber level. Player activity may have dropped off but that doesn’t automatically equate to a drop in subscriptions. Their own FAQ even states that the data should not be used for this purpose.

          • Steve says:

            It is true that MMO-Junkies stats do not directly reflect subscription numbers.

            I don’t think anyone has tried to calculate subscription numbers based on these stats though. It is more about the general trend.

            I am not sure that people play less once they hit level cap if they like the game. For WoW most of the game seems to happen at level cap these days, so it is unlikely people play less.

            If what you are saying is true though, new (more active) players coming in should have fully compensated in activity for that of players at level cap. Obviously this is not the case.

  8. BenBos says:

    And another (even more powerful tool can be added to the above measurements).
    It show the average activity over one complete week
    EU: http://www.torstatus.net/shards/eu/stats

    US: http://www.torstatus.net/shards/us/stats

    You can see straight away that the amount of full servers have bottomed out to nothing, the heavy servers are on a massive decline and the amount of light servers are going up.
    That’s a process on a weekly basis (with the last few weeks showing steep declining trends over ALL servers). … And since it was shown in the latest week also, it (of course) has nothing to do with adjusting server caps. No one will increase a 10% server cap when 60% of your servers over the week period show “light status”.

    You can track the server you are playing on btw by clicking on it.

    prefer this one to check the server status:

    EU: http://www.torstatus.net/shards/eu/trends

    US: http://www.torstatus.net/shards/us/trends

    • Gazimoff says:

      I think we’re coming to a fundamental disagreement here. I disagree with using server activity charts as an indicator of subscriber retention. Let me explain:

      If someone unsubscribes from the game then unless they’re replaced the server activity drops. I think that we can both agree on this.


      If server activity drops, that doesn’t necessarily mean that someone has unsubscribed from the game. There are other reasons for a drop in activity, such as:

      * An end in holiday (SWTOR was released during a holiday period)
      * An end to levelling activity (day 1 players are known to play intensely up to level cap then ease off)
      * Completion of existing content and waiting for further content/patches to arrive
      * Other real-life or non-game factors.

      While it’s easy to predict that a drop in subscriber numbers of x% would correspond to a drop in activity of x(y)%, the reverse doesn’t hold water as a logical statement.

      Using server activity as a predictor on subscriber numbers or retention rate is just plain wrong, no matter the source of the data.