Issue with Visitors and gaid::-1

[cuoredicapitano]

Hi
I've just found an Issue about Visitors metric and, the segment gaid::-1
If you add this segment (that should be a default segment meaning "all
visits") you'll have different data between Google Analytics and Query
Explorer.
If you remove the segment (that a lot of tools use by default) you'll
have exact correspondance between GA UI and GA API.
I don't know why this segment impacts on Visitors metrics, but it's so.

Is it a bug?

Thanks!
Carlo

[Mike Sullivan]

Just echoing this...a couple of times. Several customers have started
to notice that if you use &segment=gaid::-1 in a query, the
ga:visitors number is different (incorrect) than if you do NOT include
it, even though that is supposed to be the default.

This is occurring over several accounts/profiles.

I can reproduce it in the Data Feed Query Explorer as well. It must be
over a period of a few days at least (the default 14 day period showed
the problem).

[Mike Sullivan]

I now have yet another customer suffering from the same problem, and this time it is affecting the ga:pageviews as well as the ga:visitors metrics.  Specifying &segment=ga::-1 in the query seems to increase the numbers, and taking it out of the query seems to make the numbers reflect what is shown in the web UI.

In testing with an old abandoned blog (profile/table ID 14331662), for the calendar month of January, 2011, I actually have a test case where I can get 3 visitors vs 2 visitors. It looks like one of the visitors may be the same person but with a different browser (IE vs Firefox). My query (copy/paste from the XML response):

'https://www.google.com/analytics/feeds/data?start-index=1&max-results=1000&end-date=2011-01-31&start-date=2011-01-01&metrics=ga%3Avisitors%2Cga%3Avisits%2Cga%3Apageviews&segment=gaid%3A%3A-1&ids=ga%3A14331662&dimensions=ga%3Amonth'

Am I putting it together wrong somehow? Is order important at all?

[Nick]

So we found a little bug here. When you apply the "All Visits" segment no segment should be applied.

We'll fix this and the numbers should align.

-Nick

[mschenkel]

I am glad I came across this posting. Because it explains a major
discrepancy one of my customers came across with regards to
calculations based on the ga:visitor metric. Just a couple days I
posted this:
http://groups.google.com/group/google-analytics-data-export-api/browse_thread/thread/fcdda0cbb450e460

After a "wild goose chase" trying to figure what I may have changed in
my logic, I eventually traced it to the inclusion of gaid=-1 in the
data feed. I have since rewritten all my code, including gaid only if
there is actually a segment to be used. All is well now. Thanks for
the post - now I can sleep better at night.

Thanks All!!!!

[Mark]

More than 2.5 years later, this issue is still around.  What data does gaid::-1 actually give you?  I wrote up an experiment we did with it this morning here:

http://megalytic.tumblr.com/post/66285552398/gaid-1-is-not-really-all-visitors