The problem is that CCP's development cycle for each expansion is quite short. And being quite short, they don't have the time to really give depth of features.
Player expectations (which are always high), simply don't mesh well with the short development windows that CCP gives itself.
The development time for a summer expansion is approximately five months. January to May. Tack on another couple months for fixes and iterations once the expansion is released. Then there are two months of summer holidays, when CCP development slows to a crawl. The winter expansion cycle is much shorter, starting around the end of August. A three month development window, with another two months of fixes and iterations once the expansion is released. This is the CCP development year.
What seems to happen, especially with the big features, is that CCP announces a really exciting upcoming idea at their fanfests and player meets. Then come release day, that reality always seems to be a pale shadow of the original idea. The short development cycles should likely be blamed. There's never enough time to deliver everything, and some of the features of these bold and grand ideas have to be left on the cutting room floor.
Take Inferno's faction warfare, as an example. This probably should have been re-developed from scratch, given a complete rethink. But due to the short window of development, existing features were simply massaged and iterated upon. What resulted was a system that focused on PvE and loyalty points, rather than a system that focused on PvP. The five month summer expansion development window precluded anything but the massaging of existing features, rather than the development of all-new features.
I would suggest a single expansion cycle per year. Each expansion will focus on either a broad theme (i.e., industry or war) or a single big exciting (jesus) feature (i.e., POS revamp or ring mining.) This allows CCP to really give a theme or single concept the development time it deserves and to really pack on the features, deliver to expectations.
As well, throughout the year, various smaller iterations, fixes, and rebalances can be released. Every team will not be working on the main expansion full-time, and when they're not doing so, they can work on some of the low-hanging fruit. The little things. These are released to the players as they are completed.
What we end up with, is a fully-fleshed out concept/theme, once per year. As well, throughout the year, a random collection of little things released in a relatively constant stream.
It would seem to me, with the upcoming POS revamp, that a change in development cycle would be advantageous. The POS revamp, as envisioned at fanfest and player meets, is a massive project. A rethink of what currently exists. Modular structure-cities in space. It will require a lengthy development window. It will require work from most of the CCP development teams (though not all of them simultaneously.) So, while CCP still get the time to develop POSes in the correct way, working to deliver the expectation (rather than the pale shadow), there is still development time throughout the year to tackle little things, release them as they are finished. Players are thus not waiting a full 12 months for one big feature, plus a collection of smaller features. Rather, they wait 12 months for the big feature, and are delivered smaller features throughout the year.
Even with the change in development cycle, CCP doesn't lose out on the marketing potential that expansions bring. The expansion concept remains, CCP is just given time to deliver a yearly expansion of greater depth and scope.
Some will complain, that we're being short-shrifted an expansion per year. That's not really looking at the larger picture. Given the short development cycles to begin with (especially winter), we're being short-shrifted on the larger features with two-expansions per year to begin with. I think, in the long run, CCP will be able to deliver better content if they switch to a one-expansion per year cycle.
Certainly SOE went that route with EverQuest. They used to push out two boxed expansions a year, which tended to be rushed and often went live with broken content. So they scaled back to one a year.
ReplyDeleteOf course, for SOE, each box was a revenue stream. For CCP, the expansions might boost subscriptions a bit, but otherwise they are artificial constructs, created to give features a name and a color scheme twice a year.
It might be worthwhile to dump the expansion format altogether and instead adopt a more project-centric approach where there is a more narrow focus on a specific feature of segment of the game and then maybe ship stuff when it is ready rather than waiting for the train to leave the station twice a year. (And the do seem to miss the train and have to trot stuff out late now and again, don't they?)
Of course, then CCP would lose the "Our nth Free Expansion" press release potential.
Which is why I don't suggest doing away with the expansion concept. It does have a lot of benefit with respect to marketing an existing game.
DeleteI know this has been suggested before and CCP has always cited marketing concerns. My feeling is that they already *have* marketing concerns and they need to continue to assuage those - perhaps a middle ground where some teams are on a one-year cycle and others on a six-month?
ReplyDeletewhy not have 2 concurent systems?
ReplyDeletea single team working on big X feature will work in year long cycles where we retain the current cycles for everythign else.
Yes but, as you just said: "players want it all," and that includes NOT waiting a full year for an expansion content. ;-)
ReplyDeleteHow about not shutting down the company for 2 months?
ReplyDeletePeople need their holidays. I don't bemoan CCP on this issue. They still tackle little things during the summer.
DeleteThey need to divide their assets in such a way that one group can fine tune the things they should have tuned over the previous 8 years, while the other begins the longer-term development of actual "Jesus" features or major new content. They then sell the short term fixes as expansions, etc, while they line everything up for the new sequence of long term development, for which they will by then have built up a significant time lead without the pressures of the typical deliver-this-advertised-feature cycle.
ReplyDeleteOh wait, that's exactly what they're doing.
Why can't they work on multiple things at once ?
ReplyDeleteWhy can't we have two teams working on different things ?
It is reasonable for us to expect that the vast majority of our subscription fee's go to supporting EVE ONLINE. Lets make that a reality.
No its not reasonable to expect that.
DeleteYou pay a subscription because CCP has a product. CCP may develop, upgrade or change that product however it sees fit. You are free to cease subscribing. But paying the subscription does not guarentee that your subscription money is going to be spent supporting the product you subscribe to.
One of the key reasons for adopting Scrum as a methodology for software project management is the ability to deliver value often. Good scrum teams will deliver only a handful of sprints at a time, sometimes as few as one sprint as a time. The point is that Scrum focuses on incremental improvement to a software product instead of a large scale plan (which is subject to quite a number of problems, which is why traditional "waterfall" project management simply doesn't work for software) with a large lead time to a release.
ReplyDeleteCCP's strategy of bundling their product into 2 releases a year is somewhat at odds with the concept of iterative and incremental development. If they want to be a scrum team then I would be expecting monthly releases of very small scale features - in some instances partially complete - yet working - features (a common approach is to break a large feature down into smaller component parts which can exist independently of the whole, and release the parts in dribs and drabs until the whole exists in one form or another at the end).
Moving to a single release a year would be a step backwards if this is the approach that CCP has adopted. It does not fit properly with the principles of Scrum development or Agile principles at all really.
It is a tough one to market in the MMO industry, where players are used to expecting large scale upgrades from competing products. Delivering constant incremental change (which is not necessarilly noticable until you accumulate several months worth of changes) vs a big "wow" is always going to be a problematic balancing act...
I like the idea ; 1 Jesus feature by year where you really see content, and lots of iterations / fixes / ships balancing during this time, that can be shipped every 2 months and allow a little something to change often.
ReplyDeleteMy thought on expansions is that CCP should continue with 2 per year but with a caveat. One expansion is the 'Big New Thing' and the other is bug fixing, iteration, and tweaks.
ReplyDeleteSo once per year, either summer or winter we get a "Jesus Feature", that has been worked over for an entire year by one "Team" at CCP. For the release of that "Jesus Feature" the other teams are building ancillary content for it, ships, skills, complexes, new features, whatever. But all that got planned throughout the year by the lead team.. "We're doing this and we want to do this as well for it's release, we don't have time for the big peace and these smaller pieces so we'll farm them out to the other teams with basic design documents". It comes out. Everyone is happy about this new amazing feature that incubated and is well thought out and huge. The next 6 month release everyone but the "Jesus Feature" team goes into working on defects, balance, and minor new content in the game.
So once per year we get a "WOW!" expansion and once per year we get a quality of life expansion. It seams a much more balanced approach to game design than what CCP has been doing. For a long time it was nothing but "Jesus Feature" every 6 months. The last year and a half has been almost completely quality of life. Eve needs both, but not famine or feast quantities. Staggering them seems like the best way to do it.
You make several good points, as do several of your commenters. IMHO, "reining in" to only one expansion a year would be seen (especially by the critical pundits, and CCP's competitors) as stepping away from EvE, possibly to move to new projects.
ReplyDeleteThat said, I agree that the present schema gives the impression of grandiose dreams hammered flat into tawdry product by the dual hammers of limited working-time, and the major complexity of the existing code.
I don't (and don't pretend to) have a solution as such, but I concur with the idea that there needs to be a division between "New features", "polishing" and "Make and mend". New features should be developed to a stable and (relatively) complete state before they are released; existing features which are to be polished (eg ship skins or new warp effects) can be rolled out at any time once they are ready; "Make and mend" (by which I mean "It's *** BROKEN!! and the [griefers] are exploiting it" (insert your own hate-cabal as appropriate)) needs doing soonest and ought not to be waiting for a formal expansion.
In practice, I think I'd see this as (at the least) a monthly release of "make and mend" (possibly under the banner of CONCORD bulletins), polishing perhaps once a quarter (possibly bannered as new corp developments), and the "Jesus-features" twice a year as now, but with each one being *one* feature, thoroughly tested. This would eman that, in oder to get the lead time for the first feature, we might have to endure one cycle without anything really major, but it would then allow a rolling programme of improvements, with something always on the horizon, even if only rebalancing whatever weapon system has suddenly proved over-powered.
My 2 cents
You don't need to cut down expansion count. All you need is optimized project and marketing management. And CCP already does that in a good way. The big feature pos revamp is a long term one, which tackles a few teams over a year or two. The other teams iterate on stuff and develop smaller features (ship rebalance and new destroyers). If features get out unfinished it is not a problem with with expansion management but with project management.
ReplyDeleteBut out of experience it is not easy to pin down a project that precisely in a time frame. To get it all polished and fine it might easy take a few more month than expected and expansion date was announced too early.
I remember an expansion where ccp rescheduled the core expansion and just delivered a small piece at the first date and came back a month later with a bit more polishing.
All in all i think the two expansions a year cycle is good. If you have a feature ready it can soon be shipped and does not have to wait 6 month or more to be delivered. If it happens that 2 large projects are ready at the same expansion you get a really big expansion. If they get ready in the same year you have 2 full blown expansion a year. If only one gets ready you have one full and one iteration/small things expansion. If there is not big feature ready you have 2 expansions with many small thing to tackle and iterate on. And your customers get new stuff more often and are pleased.
Whilst you make some interesting points, it seems to me that you are also assuming a great deal about CCP's development cycle. Do you actually have first-hand information about it, or are you speculating based on how you perceive their public relations?
ReplyDeleteI would be amazed that they have a single team working in a linear fashion, rather than adopting a modular approach that lets them introduce features when they are ready, all pulled together by project management. In fact, many of the devblogs suggest this approach above your assume one. After all, the talk about modular player-owned stations has been around for a little while and will not be released for Winter, so there must be work being done on it already, even at an abstract level, whilst the current expansion is being prepared for release.
I'm confident that CCP can manage their internal projects just fine. Managing player expectations is more difficult, however.