In which Matthew goes to Training (and comes back with Ideas about Marketing)

[Stephen John Smoogen]

On 21 October 2015 at 12:29, Gabriele Trombini <g.trombini at gmail.com> wrote:
> Il giorno mer, 21/10/2015 alle 10.12 -0400, Matthew Miller ha scritto:

>>
>> Punchline
>> ---------
>>
>> The core takeaway — and, basically, this is Spoiler Alert for the
>> training — is that the essential activity of Marketing is _finding
>> problems in the market_ and helping the organization create and
>> distribute solutions to them.
>>
>
> That's fair; but in order to have a market strategy we should track our
> "sales" (not really sales of course).
> That's impossible to get, we are not able to handle what happen *after*
> the release download.
> Maybe Country communities can give us a feedback about the issues
> people are facing. Only in that way we can have a sort of realistic
> feedback; we might consider also people installing Fedora without any
> request to the community.
>

Actually release download is a horrible counte and always has been. It
is worse than no data at all because it is always incredibly overly
optimistic. Too many people set up (for i in true; do wget
http://getfedora.org > /dev/null) type things because they think it
helps us in some strange way. We have better numbers on post release
usage from dnf and other daily tools. Not perfect but probably within
an order of magnitude. [If we went by release iso download numbers it
would like some releases we should have hundreds of millions of users.
That is clearly not the case in other traffic.] The dnf/yum numbers
are limited though because it only shows up in non-metered internet
places. Most of South America, Asia and large portions of Europe seem
to have metered connections and you instead end up with snapshots of
usage because it is too expensive to stay on the daily updates tree.


>> But, Bigger Picture Problem!
>> ----------------------------
>>
>> Refer back up to the punchline, and notice something I left out —
>> identifying user problems that we want to target. This is something
>> we
>> did in very broad strokes with the three editions, but which we
>> aren't so
>> great at doing continually.
>
> So the question is, how we can get statistics? Statistics are the
> solution worldwide for getting strategies.
> I know the Project is full of people really qualified to have an
> opinion where/when we are going, but only an answer from the market can
> tell us if it's really happening.
>

I am working on getting statistics more out of the infrastructure logs
so that they can be more visible than Matt's slides. It is my desire
to have a solution by Feb of next year. The main issues is that I am
wanting to make sure that I don't violate basic privacy concerns. Then
it is also a small part of my job os progress comes in fits and
starts.

-- 
Stephen J Smoogen.