2010-09-02 MEETING PROPOSAL: discussion on Annual User Survey

Robyn Bergeron robyn.bergeron at gmail.com
Thu Sep 2 07:00:43 UTC 2010

On Wed, Sep 1, 2010 at 9:34 PM, Jon Masters <jonathan at jonmasters.org> wrote:
> On Wed, 2010-09-01 at 22:23 -0500, Mike McGrath wrote:
>> On Wed, 1 Sep 2010, Jon Masters wrote:
>> > On Wed, 2010-09-01 at 19:47 -0400, Jon Stanley wrote:
>> > > On Wed, Sep 1, 2010 at 7:34 PM, inode0 <inode0 at gmail.com> wrote:
>> > >
>> > > > What is preventing you from just going off and doing it now? I hope
>> > > > you find encouragement but I also hope the board doesn't get in the
>> > > > business of issuing mandates.
>> > >
>> > > Correct me if I'm wrong, but I read what Jon was looking for as more
>> > > of permission than a mandate of "you SHALL do this" - obviously he
>> > > already wants to. I think we should give encouragement to this,
>> > > because I think the data is extraordinarily valuable. Also, he
>> > > mentioned that there are outside companies that could be contracted to
>> > > do this - implying a financial cost. While the board has no finances
>> > > as such to direct, we can certainly say to CommArch or more directly
>> > > Fedora Engineering (as anyone else can) that we think this is a
>> > > valuable use of money.
>> > >
>> > > Feel free to correct me if I'm wrong :)
>> >
>> > You are exactly right Jon. What I want from the Board is:
>> >
>> > 1). "We would like this done". That says I'm not acting alone and that
>> > they agree with doing this. That makes it easier to get others to help
>> > because they will see that this is helpful to the project as a whole.
>> >
>> > 2). The financial cost exists if the Board prefer to have it done
>> > independently of the project, which should be considered, and is another
>> > reason why I have asked that they discuss it.
>> >
>> > I gather this discussion will actually happen on Friday, not tomorrow.
>> > The Board wiki page has had an incorrect meeting time for some time, but
>> > (other) Jon just corrected it at my request. I am unable to attend on
>> > Friday, even if there were a public IRC, since I will be driving to
>> > upstate NY for a weddding. But don't let me gate the discussion. Have at
>> > it, and let me know what is decided.
>> >
>> I'm all for this, I'd like to get someone that knows how to do these
>> things, I suspect we don't have the inhouse expertise to create this sort
>> of survey.
> So, firstly, I knew about the SuSE survey I think. And I suspect
> Canonical have done surveys, too (I can find out sometime). I suspect
> Red Hat might have some expertise in this area, but we first need
> support for doing it. Then we can go find out what resources we have to
> make it happen. I'm quite sure we can figure something out.

Yes, there's been an annual survey for the -at least- the server group
in 2008 and 2009.

As someone who used to do market research stuff for a living (both as
an industry analyst at a market research firm writing surveys and
reports on servers and PCs and usage in business markets, and on the
other side of the fence as strategic marketing analyst harassing, er,
contracting industry analysts for better data so I could do realistic
long-term strategy and planning), I have the following 8 zillion words
of wisdom:

Using an outside house is a good way for us to be unbiased in our
questioning, ensure that we can get a good representative view of the
"market" (ie: we don't want to have 300 male respondents between the
ages of 18 and 35 located in North America who primarily use Fedora
for uploading photos to failblog, and 2 respondents in China who use
Fedora for emailing friends, and that's it), and most importantly,
sanity-checking the questions we come up with for gaping holes ("You
guys don't want to know about $x??? Yes, you do!") as well as being
the unbiased group that can pare down the question list from 150 to a
more manageable number that people are actually likely to complete
without making people wonder why their question is lower priority than
someone else's question.  All that said, custom research is not cheap
(as in, at least 4 zeroes after a number, and can go up to the cost of
an official Fedora Blue Ferrari.) We'd have to find some middle ground
if we went that route, and we'd also have to expect it would be not an
overnight process. Which means that, if we are spending good money, we
should know what decisions or parts of the process are hinging on
having this data before we start, what the timeline is, etc.

I would recommend that if Red Hat has a market research group, we ask
them the following things:
1) Do you have random leftover unused custom research budget that you
are looking to put to good use?
2) Are you already contracting with any research houses on custom
research that we could sneak a few questions into?
3) Are you a customer with any research houses that would offer us a
discount for their services?
4) * Do you have any existing market research that mentions anything
about Fedora *at all* that is licensed in such a way that we / at
least one person could look at it and absorb said information? (No
point in wasting valuable question time in asking things that we have
the answers to elsewhere.)

Speaking of "what we know" - we should also look at what we have from
both Diana's research that she did (or figure out what's interesting
there that we might want to know *more* about), and also figure out
what is going on with the Duke study (is there anything preliminary
that we could please please look at).  Heck, we could ask if Duke has
some people who are interested in helping us do our own study.


* If someone is interested in pursuing this further, here's a list of
reports that actually mention Fedora:
IDC doc #lcUS22416410 - RHEL 6: Watching Linux Grow Up
IDC Doc #219380 - Worldwide Linux Operating Environment 2009–2013
Forecast: Can Linux Prove Resilience in an Economic Slump
Evans Data Open Source / Linux Development Survey 2009

> Jon.
> _______________________________________________
> advisory-board mailing list
> advisory-board at lists.fedoraproject.org
> https://admin.fedoraproject.org/mailman/listinfo/advisory-board

More information about the advisory-board mailing list