First, what kind of decline in direct visits (i.e. what %) are we
talking about?

Second, look at the long term correlation between direct entries and
"navigational" keyword searches from search engines i.e. searches in
Google for your site's URL or brand to determine if it broke down
during/after the period in question.

Third, keep in mind how the tools classify traffic. Any session that
begins on your site for which an HTTP header is absent is classified as
"direct."

As such, I'd suggest investigating the following of the top of my head:

- check to see if email campaigns were active during the same period
that the SEM campaign was underway. If someone forgot to tag the links
pointing to your site, click-through traffic will generally show up as
"direct."

- confirm that tags are present on all your entry pages if possible

- further to the above, confirm with your developers that no new
redirects have been inserted in the code. If a hop on the client side is
introduced for certain traffic entering your site and the entry page
isn't tagged, this traffic will typically show up as direct unless your
developers create a workaround

- confirm that there's no long-form content, such as a film or tv show
or magazine, with a duration of > 30 minutes (i.e. greater than the
analytics tool session timeout period) w/ no additional tracking pixel
requests in your code, has attracted a significant number of your
visitors. If 30 minutes elapse with no further tracking pixel requests,
subsequent activity is classified as a new session. This will show up as
a referral from your own site for real pages, but it may appear to be
direct if you're e.g. in a Flash application like a video player because
the previous tracking pixel request may not include associated HTTP
header variables since the URL never changed.

- Re: other campaigns, check to see if any banner (or offline!!)
campaigns were active while the SEM campaign was underway. Many people
who don't click on banners visit the site through other means, such as
direct or with brand-aware search keywords. Hopefully this will have
shown up in your analysis of correlation btw direct and
navigational/brand search keywords but it might not have...

I'm sure they are others, but I can't think of any at the moment

HTH

David

RomyM wrote:
>
>
> Hey,
>
> I would first check to see if the pages are tagged properly. One of
> our clients experienced a drop in traffic and found a couple of pages
> not tagged well.
>
> Then,try and segment across different factors to try and identify
> where the problem is. Is there any particular region, keywords,
> referring sites, new vs returning visitors which is particularly
> affected?Then dig deeper and find out what the cause could be.
>
> Good luck!
>
> Romy Misra
>
> --- In webanalytics@yahoogroups.com
> <mailto:webanalytics%40yahoogroups.com>, "jlpalmerbel"
> <jasonlpalmer@...> wrote:
> >
> > R. Karhu,
> > One explanation could very well be found in looking at visitor
> traffic rather than visits. As you look at traffic to your website, if
> you believe that 10% of your visits were coming from a large SEM
> campaign, those same visitors you were exposed through this campaign,
> may be coming back as direct traffic on their second visit and
> therefore representing the decrease in visits to the site.
> >
> > There are a couple ways to drill into this issue to validate. If you
> had set up the previous campaign as an adword campaign that tracked
> across visists, you should be able to see what the traffic looks like
> across visits. If you average visits per visitor during the period for
> the campaign comes close to 2 you may have your answer.
> >
> > Regards,
> >
> > Jason Palmer
> >
> > --- In webanalytics@yahoogroups.com
> <mailto:webanalytics%40yahoogroups.com>, "galdratyr" <rami.karhu@> wrote:
> > >
> > > Hi,
> > >
> > > During the last four weeks or so, my company's website has seen a
> drop of almost -20% in visits. Half of the drop is explained by a
> fairly big SEM campaign stopping, but what more concerns me, is that
> the other half of visits drop is within direct traffic.
> > >
> > > We are not doing less offline marketing than normally, so that
> cannot be the explanation. Also, Autumn should normally be an active
> season comparable to spring time.
> > >
> > > We had a rise in visits after summer holidays but that dropped
> back almost to summer's level in the last week of September and has
> remained there since. I cannot seem to grasp what the problem is
> simply by going through GA reports, so I thought maybe one of you guys
> has figured out a similar situation or have a good rule of thumb that
> I might be ignorant of? There haven't been any major changes to our
> website, not technical or content-wise.
> > >
> > > Best Regards,
> > > R. Karhu
> > >
> >
>
>