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

[Pierre-Yves Chibon]

On Sat, Oct 24, 2015 at 10:27:41AM +0200, Gabriele Trombini wrote:
> Il giorno ven, 23/10/2015 alle 14.55 -0400, Matthew Miller ha scritto:
> > On Wed, Oct 21, 2015 at 08:29:27PM +0200, Gabriele Trombini wrote:
> > > > 
> [...]
> > This is certainly part of it. Knowing more about the people who
> > choose
> > Fedora and how we are helping them and how we could do it better is
> > important. But, it's a relatively small slice of the pie of people
> > who
> > *could* be using Fedora (even if we narrow that from "all humans" to
> > a
> > more focused segment).
> > 
> > Making existing users happier is crucial to long-term success, but we
> > also need to find ways to reach out to our wider potential user base
> > and find what we can be doing to get them in.
> > 
> > > 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.
> > 
> > 
> > Yes, I think feedback from ambassadors is a good angle. Especially if
> > we can get people having conversations beyond LUGs and LinuxFests
> > where
> > Fedora is probably already well known. (Linux users who have chosen
> > another distribution are yet another pie slice from our own users and
> > from people who aren't using Linux at all.)
> 
> 
> We have different kind of hitch here:
> 
> - we don't know how many people are getting Fedora for the first time;
> - we don't know how many many already got;
> - we don't know issues new Fedora users are facing, beyond QA tests;
> - we don't know if people like Fedora;
> - we don't know the way-to-go people are asking us:
> - we don't know if the way we are following is clear for all the
> people;
> - we don't know if our communications are well received.
> 
> Of course I forgot something but I'd like to get focus on these, at the
> moment.
> Each one of the points above, in a structured enterprise, requires
> different approaches, I mean e.g.:
> 
> - new user -> customer file;
> - old user -> sales file;
> - issues -> after sales service;
> - liking -> fill in the form;
> - not liking -> as above;
> 
> and so on. Then is on the manager shoulders to aggregate data.
> 
> In our case we cannot think such as an enterprise, and we don't want
> to, so we should look at them from a different angle: the "customer"
> one.
> 
> We don't want tracking the user each time he/she gets in touch with us
> (FAS accounts, obviously, are out of this discussion) which means we
> can't use any tool which gathers sensitive data.
> 
> Commops already provides some tools and Stephen J Smoogen is working on
> getting statistics.
> 
> I think we could build a "let us know ...." webpage which has to be
> translated in each language we have in L10n group to fill with replies
> to questions (to be renew every 3/4 releases?).
> Of course we don't need random replies, so they must be chosen from a
> provided list (combo box?) because we have to aggregate them.
> 
> This approach has two issues:
> 1) not each fedora user will fill the form;
> 2) how we can avoid duplicates?
> 
> The first question is easy: we don't wanna force anyone, so we can
> advise that "the form is to help communications between user and the
> Project, if you don't fill the form you are not able to change thing,
> so don't protest in the future".
> The second question is really I problem, since we don't track users, so
> we have to trust in them. Of course funny people will complete the form
> more than once, but I'm pretty sure the percentage will be low. 

I am unfortunately not has trusty as you are there and there will be people
scripting their answers.
Could FAS be a solution for this? It would make the entrance barrier higher but
also much higher for bots/scripts.


Pierre