[PIGMI] Survey

Wed Aug 3 21:58:01 PDT 2011

I care a lot about what proportion of foreign and/or IE6 users I have
visiting my site if I'm in charge of browser compatibility and localisation
of the site.

If I can correlate a higher bounce rate for these users than the average
bounce rate I now know have some ideas on how to improve retention. Rather
than having no idea why the bounce rate is high and being sad about it I can
now assume (and test) that it's because the site is broken in IE6 and that a
lot of people are arriving who likely can't read the content.

On 4 August 2011 12:08, Doug <douglas.linder at gmail.com> wrote:

> I'm not arguing that demographic information is bad, or wrong, or
> irrelevant.
>
> I'm just saying: randomly collecting data in the hopes that magically you
> will be able to see patterns it _afterwards_ is fundamentally the wrong
> approach to analytics.
>
> Start with a hypothesis and then test. You can't action data.
>
>
> If you have a good reason to collect demographic data, sure.... go for it!
>
>
> Web analytics is a different story. Seriously who gives a toss if you have
> male IE6, foreign visitors who like eating banans while they browse your
> website, or more US visitors? or more visitors from ufos with 50x50 pixel
> screens?
>
> Correlate visitors to revenue and optimize for high revenue visitors.
> Correlate marketing spend to revenue and optimize for effective marketing
> channels. Literally everything else about web analytics is irrelevant
> noise.... but I digress. That's not really relevant to game survey stuff. :)
>
> ~
> Doug.
>
> On Thu, Aug 4, 2011 at 11:46 AM, Wesley Lamont <pigmi at raez.net> wrote:
>
>> Any survey can only be related to the demographic of the audience
>> surveyed. Without knowledge of who was surveyed the results can't correlated
>> to anything but speculation. But I'm sure there is information from this
>> survey which can be used. Already we know they had to do the survey online
>> which is limited to people with internet access and/or mobile phones. Where
>> the survey was posted would also determine where these results relate.
>>
>> As Paul said he did it for a reason so I assume he put the survey in the
>> position that would most suit find the answers to his questions.
>>
>> Some other statistics on game use from the US over 2010:
>> http://www.theesa.com/facts/pdfs/ESA_EF_2011.pdf
>> they have a break down of computer game sales, video game sales and other
>> media sales for 2010. And units sold from Video versus computer games. But
>> this is a different demographic surveyed
>> On 04/08/2011, at 11:14 AM, Doug wrote:
>>
>> > Can't really say I agree.
>> >
>> > Just as in website analytics 90% of the data collected is never used,
>> and thus totally irrelevant, unless you have someone dedicated to actually
>> trying to interpret your results and look for interesting relationships,
>> it's not worth bothering.
>> ahh but if you want to look at the relevance of the data in question the
>> information is useful. What countries are viewing your site, what browsers,
>> how much are people looking through your site, how long they are staying.
>> All this can alter your design and development.
>>
>> >
>> > You can talk confidence intervals, statistical relevance, domain bias
>> etc. etc. but really unless you (or someone working for you) actually has
>> the experience and knowledge to apply these concepts, you're better off
>> ignoring them; in appropriately applied they just mess the data up and you
>> end up over fitter to outliers.
>> >
>> > Broadly speaking, increased sample size --> increased survey accuracy.
>> >
>> > I don't really think including demographic breakdown is relevant in many
>> surveys just on general principles.
>> >
>> > Want to know about surfing games? Ask a question: Do you like surfing
>> games?
>>
>>  and depending on who you asked the results could vary from 0% like to
>> 100% like? so without a demographic your results are generally..... I won't
>> say useless..... but hard to correlate to anything but the people surveyed.
>>
>> Not that you would ever know nothing about those surveyed. It is
>> impossible to get a completely random assortment of people globally to do a
>> survey. Would almost instantly be limited to a country, a language and a
>> literacy demographic.
>> >
>> > Trying to figure out, oh, surfers are 'usually' 20-24 year old males
>> (O_o) and then reverse correlate that to results is just asking for trouble.
>>
>> now if you asked the survey and they could add their age you would be able
>> to look at your result and state. Thos aged 5-10 like surfing games 15%,
>> 10-20 like it 25%, 20-30 like it 45% and 30-40 like it 30%
>>
>> so you can then use that information to tailor the development for a
>> narrower target audience.
>>
>> >
>> > :) Just my $0.02. Been involved in a fair bit of web analytics and the
>> number of times I get asked if we can collect demographic data to see "if we
>> can see some interesting relationships" is just silly. No. You can't.
>>
>> such as where people are viewing your website from what country when you
>> are selling stuff overseas? or people are looking at the home page and
>> leaving after 2seconds? (maybe the flash file is too big, maybe there is a
>> flaw on certain browsers etc etc)
>>
>> Wesley Lamont
>>
>>
>>
>> >
>> > ~
>> > Doug.
>> >
>> > On Thu, Aug 4, 2011 at 10:51 AM, Ellen Jurik <
>> sunlightsylph at yahoo.com.au> wrote:
>> > Theoretically, every game appeals to everyone ever.
>> >
>> > And yet effective marketing and advertising relies on knowing who is
>> likely to be a surfer and therefore where they will be and how they find out
>> about it. While I agree that it can go overboard and become superfluous,
>> demographics can point out false assumptions- maybe a surfing game would be
>> more likely to be accessed via Facebook or a flash game portal, etc.
>> >
>> > It's not just me thinking this- almost every division of the
>> entertainment industry does it, and I know for a fact that Tantalus
>> approached "Pony Friends" that way, and it explains why there are so many
>> "little girl" games on the DS- because little girls are more likely to have
>> a DS, and DS consoles are more likely to be owned by little girls.
>> >
>> > Besides, sometimes it just puts your survey results in context, or can
>> lead to interesting data or links you hadn't expected ;)
>> >
>> > Ps, you know me and I'm in no way attempting to undermine the validity
>> of the results. I just do believe that demographics of age, sex, income etc
>> shouldn't be dismissed as completely superfluous.
>> >
>> > Sent from Yahoo!7 Mail on Android
>> >
>> >
>> > From: Paul Turbett (Black Lab Games) <paul at blacklabgames.com.au>;
>> > To: <pigmi at pigmi.org>;
>> > Subject: Re: [PIGMI] Survey
>> > Sent: Thu, Aug 4, 2011 2:34:29 AM
>> >
>> > I respectfully disagree regarding demographics. Whilst it is possible to
>> generalise about certain age and gender profiles gravitating to a particular
>> platform, I think it's more useful to focus in interests. If somebody makes
>> a game about surfing, it is going to appeal to people who surf. That covers
>> a wide age range, and both genders.
>> >
>> > L8r, Paul
>> >
>> > On Thu, Aug 4, 2011 at 10:12 AM, Ellen Jurik <
>> sunlightsylph at yahoo.com.au> wrote:
>> > Demographic breakdown is certainly important for making design decisions
>> and choosing which platform for which game (eg, 9 year old girls might not
>> play on PS3, so the game would never sell). It can still be anonymous while
>> gathering that data- "anonymous" just means "without name" :)
>> >
>> > Meanwhile, ScreenAustralia's Audience Participation Survey didn't even
>> give the options of PCs (or macs) or non-mobile/phone handheld gaming
>> devices... Which data sources were you looking at, Wesley? Sometimes surveys
>> that look totally fair, scientific and impartial aren't actually designed
>> very well.
>> >
>> > Sent from Yahoo!7 Mail on Android
>> >
>> >
>> > From: Paul Turbett (Black Lab Games) <paul at blacklabgames.com.au>;
>> > To: <pigmi at pigmi.org>;
>> > Subject: Re: [PIGMI] Survey
>> > Sent: Wed, Aug 3, 2011 9:21:02 AM
>> >
>> > Hi Wesley
>> >
>> > The survey was anonymous, so I didn't collect that data. To be honest,
>> > I don't think age and gender is really important (and certainly wasn't
>> > relevant to the goal I had for the survey which was to assist with
>> > some decisions about a project I have in the works).
>> >
>> > L8r, Paul
>> >
>> > On Wed, Aug 3, 2011 at 3:18 PM, Wesley Lamont <pigmi at raez.net> wrote:
>> > > Thanks for sharing Paul.
>> > >
>> > > I also am unsure of the respondents. Could you put up a break down of
>> the people who replied? Age, platform reply was made, gender? If you
>> collected that data that is. From most other sources Consoles beat a PC for
>> games share by a rather whopping margin so I'm curious as to the reasons
>> (apart from those you have already mentioned in the article).
>> > >
>> > > Wesley Lamont
>> > >
>> > >
>> > > On 03/08/2011, at 2:29 PM, Paul Turbett (Black Lab Games) wrote:
>> > >
>> > >> The survey was web-based, so it could be taken on any web-enabled
>> > >> device. The other thing is that it asked how frequently each of the
>> > >> platforms are used for gaming, which I think it independent from the
>> > >> device being used to answer the questions.
>> > >>
>> > >> However I think the results are skewed for sure - by the sources I
>> > >> used to direct people to the questionnaire. But like I said in the
>> > >> post, that was intentional, because the sources are related to the
>> > >> people I'm seeking to make games for.
>> > >>
>> > >> L8r, Paul
>> > >>
>> > >> On Wed, Aug 3, 2011 at 1:25 PM, Michael B <michael.sg at gmail.com>
>> wrote:
>> > >>> Hi Paul, i just had a thought, don't you think that the results
>> might be
>> > >>> compromised by having the survey on a PC?
>> > >>>
>> > >>> Ie if the survey was given to kids in school, high school, tafe,
>> even
>> > >>> university whether the kids and adults who couldnt afford PC's or
>> iphone's
>> > >>> would make up a larger majority than PC users ?
>> > >>>
>> > >>> Michael
>> > >>>
>> > >>>
>> > >>> On 3 August 2011 12:16, Paul Turbett (Black Lab Games)
>> > >>> <paul at blacklabgames.com.au> wrote:
>> > >>>>
>> > >>>> Hi All!
>> > >>>>
>> > >>>> For those interested in stats (and lets face it, that's everybody
>> > >>>> right?), I've just got around to my first post on the survey I
>> > >>>> conducted in June. http://blacklabgames.com.au/blog/2011/08/stats/
>> > >>>>
>> > >>>>
>> > >>>> L8r, Paul
>> > >>>>
>> > >>>
>> > >>>
>> > >>> _______________________________________________
>> > >>> PIGMI mailing list - http://pigmi.org/ - pigmi at pigmi.org
>> > >>> Unsubscribe: http://lists.pigmi.org/listinfo.cgi/pigmi-pigmi.org
>> > >>>
>> > >>
>> > >>
>> > >>
>> > >> --
>> > >> Paul Turbett
>> > >> Black Lab Games
>> > >> www.blacklabgames.com.au
>> > >> _______________________________________________
>> > >> PIGMI mailing list - http://pigmi.org/ - pigmi at pigmi.org
>> > >> Unsubscribe: http://lists.pigmi.org/listinfo.cgi/pigmi-pigmi.org
>> > >
>> > > _______________________________________________
>> > > PIGMI mailing list - http://pigmi.org/ - pigmi at pigmi.org
>> > > Unsubscribe: http://lists.pigmi.org/listinfo.cgi/pigmi-pigmi.org
>> > >
>> >
>> >
>> >
>> > --
>> > Paul Turbett
>> > Black Lab Games
>> > www.blacklabgames.com.au
>> > _______________________________________________
>> > PIGMI mailing list - http://pigmi.org/ - pigmi at pigmi.org
>> > Unsubscribe: http://lists.pigmi.org/listinfo.cgi/pigmi-pigmi.org
>> >
>> > _______________________________________________
>> > PIGMI mailing list - http://pigmi.org/ - pigmi at pigmi.org
>> > Unsubscribe: http://lists.pigmi.org/listinfo.cgi/pigmi-pigmi.org
>> >
>> >
>> >
>> > --
>> > Paul Turbett
>> > Black Lab Games
>> > www.blacklabgames.com.au
>> >
>> >
>> > _______________________________________________
>> > PIGMI mailing list - http://pigmi.org/ - pigmi at pigmi.org
>> > Unsubscribe: http://lists.pigmi.org/listinfo.cgi/pigmi-pigmi.org
>> >
>> > _______________________________________________
>> > PIGMI mailing list - http://pigmi.org/ - pigmi at pigmi.org
>> > Unsubscribe: http://lists.pigmi.org/listinfo.cgi/pigmi-pigmi.org
>>
>> _______________________________________________
>> PIGMI mailing list - http://pigmi.org/ - pigmi at pigmi.org
>> Unsubscribe: http://lists.pigmi.org/listinfo.cgi/pigmi-pigmi.org
>>
>
>
> _______________________________________________
> PIGMI mailing list - http://pigmi.org/ - pigmi at pigmi.org
> Unsubscribe: http://lists.pigmi.org/listinfo.cgi/pigmi-pigmi.org
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.pigmi.org/pipermail/pigmi-pigmi.org/attachments/20110804/f7f53072/attachment-0002.htm>