There's a more general problem here we should fix:

We already know that effectiveness of any single banner drops off
dramatically after the first few views.  So there's rarely a reason to run
a continuous banner -- certainly not if there are other banners to run.

We should improve the effectiveness of all banners by showing them
conservatively.  If most pageloads are completely banner free, banners will
likely have a greater impact when they are shown. And we should be able to
measure this: tracking how effective each % of visibility is at getting
clickthroughs for each banner.  Making that sort of data immediately
visible to everyone who designs banners would be a great step forward.

Are WLM projects doing A/B testing of banner messages?  Do you have access
to A/B test frameworks and results?


On Tue, Aug 18, 2015 at 3:18 PM, Nathan <> wrote:

> On Tue, Aug 18, 2015 at 2:59 PM, Romaine Wiki <>
> wrote:
> > Hi Antanana,
> >
> > And I forgot to mention, this same issue existed in 2014 as well, with
> also
> > there the downside effects.
> >
> > This subject is of banners has been discussed internally with the local
> > Wiki Loves Monuments team, after I tried to gave some insights in the
> > matter. I think this is done so because me and others have always thought
> > and assumed that it is possible to find a solution with understanding of
> > both sides. With these outcomes I think I can safely say that that
> > assumption and thought can't be considered realistic.
> >
> > I think it would be better in future to have the community decide somehow
> > how they perceive this matter. After all, they create the content of
> > Wikipedia and bear the bunt as result of it.
> >
> > Romaine
> >
> >
> It seems like there are other communication channels you could take
> advantage of - other types of banners, bot-distributed talk page messages,
> WMF-assisted mass e-mail campaigns, social networking messages (FB,
> Twitter, etc.) and so on. Is it really true that having to share banners
> with fundraising will result in an unavoidable loss of 90% of contributors?
> _______________________________________________
> Wikimedia-l mailing list, guidelines at:
> Unsubscribe:,
> <>

Samuel Klein          @metasj          w:user:sj          +1 617 529 4266
Wikimedia-l mailing list, guidelines at:

Reply via email to