While I know conversion tracking is essential if you’re doing any more than using AdWords for 谷歌购物广告优化, there’s no denying it’s not perfect. Here are 3 examples of its imperfection:
- Analytics “goal” conversion tracking stats will not match the AdWords conversion stats.
Why? I don’t know. How much are they off by? I’ve seen it as high as 25%. This is a great argument for keeping your keyword list simple, or at the very least, identifying your top 10-20 keywords and giving them extra special treatment. When I was managing an AdWords account that had over 50K keywords in it (client’s idea, not mine – I think 2-5,000 is more than plenty) and was spending $12K a day, I did take the top performers and compared the AdWords to Analytics conversion data every 2 weeks or so. You can do this for keywords and for ad copy.
From what I can tell, Analytics doesn’t break out keyword match types. I’ve never really dug into this (note to self: email analytics support about this) and even after digging around ROI Revolution’s site, I still don’t think Analytics shows match types. However, there is a tool that shows the exact search query (http://www.ga-experts.com/blog/2006/11/how-to-get-detailed-ppc-keyword-data-from-google-analytics/). I do wish they could give us better apples apple view. Now it’s more like apples to rose hips.
- Google only counts conversions that happen within 30 days of the original click.
If you’ve got a really long sales process, good luck. The best fix might be to just use AdWords conversion tracking for the first step in the sales process (like downloading a white paper) and then use Analytics goals to track where the user went from downloading the white paper.
The next question is if AdWords only tracks conversions within 30 days, how long is the conversion tracking window for Analytics? There’s a good piece of info comparing the two kinds of tracking here: http://www.google.com/support/googleanalytics/bin/answer.py?hl=en&answer=55535 , but it doesn’t say how long Analytics tracks for.
If you know, please educate me.
- AdWords will assign conversions to the day the original click happened, not the day that the conversion actually occurred.
I do agree with this – if you’re truly tracking conversion activity, it’s not just the keyword and ad copy you’re tracking – the time of the conversion matters, too. But if you don’t know that AdWords is assigning conversions to clicks that happened last week, or two weeks ago, your interpretation of the data is skewed.
Again – I wonder how analytics handles conversions – is the conversion assigned to the day the click happened or the day the conversion actually happened?
If Analytics is attributing the conversion to the day the conversion actually happened, that might well explain the big differences I’ve been seeing between AdWords and Analytics counts. I could probably tell that by comparing the two figures just after the day closed out (maybe add six hours for the reports to update) and comparing the two figures again later… say 3 weeks later.
Lemme go check that…
- Here’s data from May 11th that I actually grabbed on the morning of May 12th (so we’ve got the reporting time delay out of the way)
- AdWords “one per click” conversions: 54
- AdWords “many per click” conversions: 67
- Analytics conversions: 62
Here’s what’s being shown for May 11th per the reports generated today, June 10th (which is just within the 30-day window)
- AdWords “one per click” conversions: 55
- AdWords “many per click” conversions: 69
- Analytics conversions: 62
So no change in the Analytics data, but one extra sale in the AdWords data. I’m guessing, but it looks like Analytics counts the conversion on the day it actually happened, not on the day that the click that generated the conversion happened.
Again, if anybody knows this for certain, please enlighten me. If I’m right, then that might explain at least part of the gap between Analytics and AdWords conversion counts I’ve been seeing the last few years.
Knowing how long the average customer takes between clicking on a PPC ad and actually completing the conversion might be a good metric for an advertiser to track. Even if it’s only measured quarterly or monthly. Of course, most people are not as obsessed with accurate tracking data as I am. But when you’re running $2K a day through the exact match of one keyword, for example, being 10% off every day adds up to $73,000 worth of lost sales if the conversions aren’t being tracked properly. Thus my paranoia.