I’ve been telling customers for years that web analytics packages are essentially mired in a technologists view of performance, not marketing. Now the larger agencies are starting in on the analytics vendors.
We’ve developed our own, behind-the-scenes sets of analytics to identify user behaviour and optimise paid search marketing. If you can’t identify average pages browsed by paid search source and keyword, or calculate an ROI using changing bids, what are you doing?
Some leading packages that we’ve used don’t even separate the type of match on AdWords, blending the separable performance of exact, phrase and broad match. All too frequently concurrent or sequential A/B testing of copy (advert or landing page) is extremely complex or even impossible, an appalling state for this most basic marketing activity.
Actually, I guess I should be mourning this - we’ve been doing this stuff for ourselves, for our small client base, for a few years. Now others will get in on the game!
Merjis blog » Blog Archive » Macros, Analytics, Paid Search Performance Improvement wrote,
[...] How do you tell whether you have actually received visitors from AdWords? You need to use web analytics. There’s two problems with that solution - most analytics are completely useless for explaining what happened with your paid search, and secondly, the search engines only give you part of the information they collect about the adverts you pay for… making it hard to refine your paid search, without being an expert in paid search as well as your professional or vocational expertise. [...]
Link | April 25th, 2007 at 12:27 pm
Amit wrote,
This is almost bul*^t…
What do you mean “How do you tell?”
It’s as easy as setting a URL with a cookie..
that how “all” of them works. Not to mention
Google Analytics integration with Adwords.
I don’t get this post.
Link | March 15th, 2008 at 12:20 am
Amit wrote,
Just to go on the safe side..
Here is a filter and java script file allows you to pull in the EXACT keyword string used by people coming to your website through PPC ads
http://www.roirevolution.com/blog/2008/02/exact_keyword_tracking_with_gajs.html
Link | March 15th, 2008 at 12:27 am
Jeremy Chatfield wrote,
@Amit - Web Analytics Packages will still gleefully give you a breakdown of the operating systems and revision levels of OS and browser. However, finding out what the paid sources are that are involved in obtaining conversions is a non-trivial exercise - even if you tag all your paid sources.
Off-the-shelf packages are mostly configured to report on only the source used to initiate a session. Sessions are mainly meaningful when you are trying to explain to an advertiser the publishing characteristics of a site - they are not so useful to explain how to optimise a marketing mix. And that’s what the Avenua A/RazorFish article referenced in this posting, was all about.
And I still find that most web analytics packages require non-standard setup to discover basic things about the effectiveness of spend. In other words, by default, most web analytics packages will tell you technical stuff easily, but fail to deliver marketing insight in ways useful to control spend. It’s a pretty basic business need, unfulfilled. Here’s an example - uses AdWords, uses Google Analytics, can’t trace the clicks out of the box.
I note that just last week, Microsoft was bragging that they’ve stolen a leap on the analytics industry by releasing a package that is supposed to report on the Marketing Mix, by default. It’s been about 15 years since people started writing analytics packages - and we still make it easier to understand the OS revisions of our customers, than to understand how the spend has worked, out of the box.
Link | March 15th, 2008 at 1:15 am
Amit wrote,
Hi Jeremy,
Well.. a really really basic use and understanding won’t get you there..you’re right.
But when you say “If you can’t identify average pages browsed by paid search source and keyword, or calculate an ROI using changing bids, what are you doing? ”
I’m not sure if you refer to the software or the analyst, cause both ways, if you do it (analyze with a software, let’s say in the level of Google analytics)
all of these are quite easy to understand and access, the interpretation and actions result from tha..that’s a different story.
Link | March 16th, 2008 at 10:20 pm
Jeremy Chatfield wrote,
@Amit - this was a three paragraph item pointing to the RazorFish article. Aren’t you, perhaps, going a little overboard for a critique?
I’m still disappointed that web analytics packages are so difficult to configure. Remember that marketeers generally are not technicians. While my programming background means that I do understand how to configure these tools, most marketeers end up using what they’ve been given.
The current state of the tools is pretty pathetic, off the shelf. Getting them to deliver better results requires marketing and R&D staff working to solve the problems. Usually the packages salespeople say that the tool does everything, the company management don’t want delays in getting reports from the site now, and marketing ends up making elaborate presentations about the OS breakdown of site visitors (because that’s all they can get to without unbudgeted configuration and programming) and can say nothing useful about the evolution of search, marketing mix, etc for their products, and the relationship between A/B tests and conversions.
Note that this comment is now longer than the article.
I think I’ll write up a proper article, with some examples of how packages could be better configured for delivery and use. :)
Link | March 17th, 2008 at 5:32 pm