Gamasutra is part of the Informa Tech Division of Informa PLC

This site is operated by a business or businesses owned by Informa PLC and all copyright resides with them. Informa PLC's registered office is 5 Howick Place, London SW1P 1WG. Registered in England and Wales. Number 8860726.


Gamasutra: The Art & Business of Making Gamesspacer
Dissecting The Postmortem: Lessons Learned From Two Years Of Game Development Self-Reportage
View All     RSS
July 9, 2020
arrowPress Releases
July 9, 2020
Games Press
View All     RSS







If you enjoy reading this site, you might also want to check out these UBM Tech sites:


 

Dissecting The Postmortem: Lessons Learned From Two Years Of Game Development Self-Reportage


July 8, 2011 Article Start Previous Page 2 of 4 Next
 

Caveats 

It's important to mention some qualifications that are inherent in the nature of postmortems intended for public consumption. The first thing that's called into question is the sincerity of the report itself. Is the writer deliberately holding back certain information that would be considered "bad PR" for the company to reveal?

Some writers appear to be more forthright about dramatically bad outcomes of their projects than others. It seems safe to say that there is likely some amount of this type of informational restraint, but of course we can never really know who is writing with candor and who has one or more fingers tied behind their backs for whatever reason.

Fortunately, I have been pleasantly surprised about the willingness of many authors to expound on some quite disastrous situations in their postmortems.

The second major qualification is one of completeness of information. Because postmortem authors are free to pick their favorite five "bad things" and "good things" about their projects, and because each postmortem is written independently, there is no guarantee that any given topic will be mentioned at all in any given postmortem.

What's more, the absence of mention of any given topic gives us no information about that topic one way or the other. For example, if there is no mention of "working crunch" in any way, we still don't know if the team worked crunch or not.

Either way, if the team did work crunch and it wasn't mentioned, we wouldn't even know if it was a brief finish-line crunch with everyone cheering to complete their most polished work, or a morale-obliterating death march with far-reaching casualties.

In short, we're at the mercy of the collective of postmortem authors, as the quality of information presented here is a function of their ability to present honest and complete information about their own projects.

Results Part 2: How Things Went Right and Wrong 

As described in the Methodology section, each "thing that went right" and "thing that went wrong" that was reported in each postmortem was classified into one of six categories. Each of these categories generally corresponds to a major discipline involved in game development. In total, this supplied 240 data points: 120 "rights" and 120 "wrongs."

If we look at the all of "right" issues under this classification, we see some interesting results:

What's going on in this pie chart? When things go well, production is most often credited for it, with design running second place and issues external to the team a close third. Art issues appear to take a disproportionately small slice of responsibility for the good things that happen on projects, and testing gets the least amount of representation.

Now, let's take a look at how issues were classified when things went wrong.

Immediately we see that production issues take an enormous share of the problems when things go wrong. What this means is that when things go wrong on a project, most often it's not inherently because of design mistakes, art mistakes, or coding mistakes, but problems in the way the process of development is prioritized, conducted, and managed.

In general, this seems to indicate that development teams are just much worse at planning, coordinating, and conducting the work required to produce a game as a whole, more than anything else. I think this finding will resonate with a lot of developers -- poorly managed projects unfortunately appear to be much more common than well-managed ones. Mind you, our sample of postmortems was all games that shipped, and at least by that account can be considered successfully managed projects.

There were a total of 68 individual "what went wrong" issues classified under the "production" category. Within this group, the most common issues were related to scope, feature creep, and resource problems; this accounted for 16, or about 23 percent of production problems. The second-largest subset of problems within production was team-communication related, which accounted for eight, or slightly more than 10 percent of the issues. Six issues involved various critical events happening later than they should have in the development cycle.

While production increased dramatically in the "wrongs" versus the "rights," all other issue types decreased in frequency of reports (except for testing, which nearly doubled, but was still a small number).

It should be noted that the "wrong" issues related to testing were all about planning, managing, and being unprepared for the logistical burdens of the testing process, and not related to testers doing a bad job. For these reasons, most of the testing-classified items could arguably also be included in the production category instead.

Another interesting statistic is that programming issues decreased only by 1 percent, while design, art, and external categories decreased by 7 percent, 5 percent, and 9 percent respectively. This also seems to make sense, as when things do go wrong, technical issues can easily have much more salient effects on the health of a project. However, design issues are still the second most represented type, which also seems to indicate that designers may be creating more problems for projects than the other traditional disciplines.

Here's what the distribution of issues looks like if we combine all the rights and wrongs together:

This aggregate graph might be best understood as the amount of overall impact or significance a certain discipline can make on the smooth-running of a project's development. As expected, production is almost half of the pie, with design taking the number two spot again, and coding taking roughly equal weight to external issues. Note that since postmortems are explicitly about the process of game creation, this is not an indication of what is most important to making a game successful or artistic -- just to completing it.


Article Start Previous Page 2 of 4 Next

Related Jobs

Insomniac Games
Insomniac Games — Burbank, California, United States
[07.08.20]

Character Artist (Blendshapes Focused)
Insomniac Games
Insomniac Games — Burbank, California, United States
[07.08.20]

VFX Artist
Remedy Entertainment
Remedy Entertainment — Helsinki, Finland
[07.08.20]

Technical Director
Wooga GmbH
Wooga GmbH — Berlin, Germany
[07.08.20]

Unity Game Engineer





Loading Comments

loader image