Which Stat do you trust?

Google AdWords or WebTrends?

[chewy]

I'm noticing a new discrepancy between what Google AdWords is reporting and what my WebTrendsLive account is reporting.

For one day, I get about 500 visitors as reported by AdWords.

For the same day, WebTrends (live and Log Analyzer 7.0C) both show about half of these visitors.

This used to vary maybe 5 - 10%

Which would you trust? And why?

Thanks,

C.

[sem4u]

I would probably trust the AdWords stats. The problem with log analyzers is that they don't always report the information from raw log files in exactly the way they should. In my experience they underestimate clicks and traffic.

It is far better to take a day's log files (or whatever time period you like) and compare clicks from AdWords with the stats from the campaign manager. This should match up the figures. If not, this may be due to the time difference on your server and the start/end times of the AdWords day.

[cornwall]

Deep subject Log File Analysis!

Take caching for example, and AOL user can pick up your page on the AOL cache, it will not show up on your log file, but will trigger an Adword event.

Have a look at the tracking and logging forum to see further examples of discrepencies.

[sem4u]

Yes, even the log files are flawed! Sometimes I really hate AOL - always get a lot of visitors from Reston, Virginia, USA ;)

[chewy]

WebTrends Live is Java Script on all the files - so this is about as good as it gets.

Time synch's seem about perfect - although I don't know how Google sets their time.

I have viewed both the data tracked from JS as well as the actual log files.

What I am trying to find out is if others are seeing this.

Might something have recently changed in the AOL architecture so Web Trends is now undercounting it?

I do get a majority of traffic from AOL although WT is not showing the clear, sharp picture of AOL spikes like it used to.

One other thing that is different is that the AdWords traffic is landing on a specific landing page, not the home page - and last I knew, this was not likely to adversely affect tracking the refferer source + word.

Web Trends seems to be showing the users were there, just not the referral source and word.

[count_zer0]

I am having the same problem with WTL under-reporting PPC referral figures by more than 50%. I am also running these through a tracking script which redirects to a target page. WT support recommended putting a delay of 1 second onto the redirect to allow the tracking GIF time to be generated. I am still waiting to see if this has an effect on the accuracy.

One thing to remember, is that WTL measures visits, not clicks. However I would expect a more or less one-to-one correlation of clicks to visits for PPC referrals.

I would be very interested to hear how you resolved this problem - WTL is great in most other respects.

[chewy]

We are also using a tracking page (which also shows the same problem) however we aren't seeing any changes to more fully reflect what we are paying for in terms of ad clicks from Google AdWords.

I've got a call into Google on this. They're working on it.

Maybe someone else can put a call into NetIQ?

We just did a stats review for 2003 which looks great - then when we hit 2004, things drop off a cliff. (In one specific keyword, it off by a factor of 4.) I usually expect a slop factor of .10% - .30%. 400% tells me something is wrong.

We've made no changes at this end in code or hosting options.

I know WebTrends Live was running behind schedule during December but this is ridiculous.

C.

[chewy]

anyone able to shed any further light on this topic?

I do use landing pages, and the Google AdWords clicks do appear to show up there, but WTL is not picking up the referrer and keyword.

Is it possible that WTL doesn't pick up this level of detail from inner pages?