Adwords v. Analytics

[Margaret]

Has anybody compared these? I recently had a blitz of a few hours on Adwords. Adwords claimed they had sent me 71 visitors to my site, whereas google/cpc (the only thing on the analytics list which I though might be Adwords) said it was the source of 38 visitors. Any sensible explanation for this? Am I looking at the wrong thing?

[mitchellbryson]

Neither are in real-time, so you can only really compare them starting 12-24 hours ago, i.e. nothing in the last 12 hours.

I don't know how often they update exactly (it will be published in their support docs somewhere) but I'd say 12 hours is safe.

If this is already the case and you are comparing historic data from more than 12 hours ago, there are 2 more explanations.

1) Users that clicked on your ad, don't have referrer tracking enabled in their browser (this is how analytics knows they come from google/cpc)

2) Users that clicked on your ad don't have javascript enabled (Analytics relies on javascript, unless you use their noscript gif)

However if you still think you are being short-changed on the number of visits you actually get, take it up with adwords staff and they will look into it for you, and will reimburse you for any fraudulent/missed clicks

[Margaret]

So the figures on Analytics, widely touted as accurate, don't include these 2 groups, not accounting for 50% of visits in my case? This was last week so so chance of updating periods being responsible.

[marcus]

Users that clicked on your ad don't have javascript enabled

Surely if someone had javascript disabled they wouldn't even see the ads - so can't possibly click on them...?

I'm wondering what happens if someone clicks on an ad but then uses the back arrow before a page has finished loading - analytics is usually last thing in the page code, so I think adwords might recognise a click without analytics knowing about it (although I though adwords excluded clicks in these circumstances?)

[mitchellbryson]

The ads aren't created by Javascript. They're static html. So yes they would still see them and click on them.

You're right about the back button thing, although Google analytics gives you an improved Javascript snippet to combat this. it's asynchronous, meaning it starts as soon as a page request is made.

[caroig]

Only about 1% of users don't have javascript switched on, so that's unlikely to be the source of such a descrepancy.

You might want to take a look at invalid clicks that Google identifies.
Click on the main campaign tab in adwords, then on columns (in the row below) and you'll two columns available for invalid clicks and invalid click rate.
In my Spanish campaigns (sounds nearly Napoleonic!), these account for 5-15% of my clicks - a surprising number.

You won't have paid for these clicks. They won't appear in your adword stats, but I'd guess they would appear in analytics - which if anything would inflate your analytics stats, not the other way round.

I've recently reported suspicious clicks to the adwords team and they were surprisingly responsive, giving me detailed feedback on the adwords data I was concerned with. You might just want to pen them a message.