Google Change Breaks Third-Party Analytics Tools

Tweet

A couple of days ago blogger Michael VanDeMar wrote that Google appeared to be testing AJAX search in the US. VanDeMar noticed that URL of his Google queries switched from the familiar “/search?=” to “/search?#=” The # is a visual cue of a “page change” for the user, and allows the browser’s back and forward buttons to make AJAX calls.

After VanDeMar’s post a lot of people speculated what the ramifications of Google AJAX search might be. From breaking Firefox plugins, to causing search rank checkers to need rewriting, to at least temporarily stopping automated SERP scrapers. But those all pale in comparison to the broader implication that VanDeMar discovered this week. Google going AJAX means that any external analytics package will be unable to read the search queries used by visitors to get to your site (at least from Google — which, of course, is where most search traffic comes from).

Analytics packages get search query information by reading the text of the referral URL that browsers send to servers to tell them how you arrived on that page. The reason Google’s AJAX search breaks this is that browsers don’t send any information after the # hash mark. This isn’t something that analytics packages can create a workaround for — the only way to change it would be to rewrite at the browser level — and it affects log based analytics software the same as it does JavaScript based.

For now, it even affects Google’s own Analytics package, which is showing referrals from Google AJAX searches as coming from google.com. But makers of analytics software are understandably worried. “If this update goes live for everyone, it effectively means that 2/3 of all searches leading to the average web site will be a complete mystery. This is huge,” writes third-party analytics provider Clicky on their blog.

Why would Google make such a change? Well, one possibility is that the test is just a test and won’t go live to everyone. The reason you test before a system wide change is to understand the consequences of the change. Another reason, however, as Clicky suggests, could be more “evil.” Google Analytics is broken just like any other analytics package right now, but maybe not forever.

“Of course, I’d be willing to bet that somehow, Google Analytics will still be able to get the search terms,” writes Clicky. “I can see it now: ‘Oh, sorry, we broke Clicky? No worries guys! Come use Google Analytics, and everything will be fine! Oh, we completely destroyed all our competitors in the process? Whoops!'”

Search blogger Peter Da Vanzo has hit upon a similar line of thinking. He suggests that Google might be gearing up to lock web site owners into Google Analytics if they want complete search referral information. He reminds us that Google has done something similar once before.

“Remember the changes to Adsense? Google introduced a new form of tracking code that can’t be tracked by third party tools. However, that data is available within Google Analytics,” he writes in a blog post on SEOBook.

A Google spokesperson said today that the experiment was just an experiment and only visible for some users. The goal, says Google, is to make search faster for users and that breaking referrer tracking was not their intention. “It is not our intention to disrupt referrer tracking, and we are continuing to iterate on this project,” said the spokesperson.

We shall see.

What are you thoughts? Google being evil or tempest in a teapot? Let us know in the comments.

Get your free chapter of Level Up Your Web Apps with Go

Get a free chapter of Level Up Your Web Apps with Go, plus updates and exclusive offers from SitePoint.

  • corbyboy

    Here’s a question: why don’t web browsers send any referrer information after the hash?

  • lpereira74

    SearchWiki comes to mind here and not so much Google Analytics. The Social Web is the future.

  • DangerMouse1981

    Because fragements are not considered to be part of the URI – https://bugzilla.mozilla.org/show_bug.cgi?id=179400

  • Tarh

    Relying on referrers was an unstable business from the start. You simply can’t rely on referrers (I use the Referrer Spoof extension set to “Forge” mode to purposely mess up most web tracking packages). Whether or not Google intended to do this, breaking tracking applications can only be a good thing.

  • 2MHost.com

    I see now .. I was crazy yesterday because my custom tracking systems showed some visitors comes from http://www.google.com without any paramaters like the key words the visitor used to reach my website ..

    thank you Josh for the update!!

  • Silver Firefly

    I only use Google Analytics, so I’m sure they wouldn’t let this go live AND break their own software, would they? So I have little to worry about. /shrug

    As for the other analytic software, maybe it’s time for them to shift up a gear.

  • nachenko

    Fixing this is actually very easy provide you CAN ask Google for reference, which I’m pretty sure Analytics will be the only one able to do.

  • http://altoonadesign.com halfasleeps

    Because fragements are not considered to be part of the URI –

    Well with current uses of the hash, (ajax and flash deep linking) maybe that needs to be rethought.

    Off topic:
    How do you do a quote on here?

  • corbyboy

    Looking at the RFC it states that the hash fragment should NEVER (in capitals) be passed as the referrer. They then link to the security section, which doesn’t actually mention anything about the hash fragment.

    Does anybody know why it isn’t considered part of the URI?

  • http://www.mpavel.ro TKD

    Good one from Google … or maybe not so good. Anyway, I agree with Tarh. There was another post on SitePoint talking about building a business on another business (had to do with the iPhone I think)…

    Google can do what they want to do with their search engine and data, they shouldn’t actually care about the others. Why should they?

    Overall ‘the game’ is all about the money, isn’t it?

  • glenngould

    Off topic:

    How do you do a quote on here?

    Use b-quote button or <blockquote></blockquote>

  • http://www.mockriot.com/ Josh Catone

    @TKD: In this case, I would guess they might be looking at potential anti-trust violations. They’d essentially be using their monopoly position as the world’s largest search engine to bully competitors out of the market. That’s not legal.

    (Of course, I am not a lawyer, so I could be wrong.)

  • http://www.mpavel.ro TKD

    @Josh Catone

    I’m not a lawyer either, however, Google’s competitors are not the companies who use the results of their search engine. It’s other search engines. It might be an issue with Google Analytics.

    We all like diversity in our lives and more places to choose from. If you’re trying to do a large search on something you will be using other search engines as well. And if Google will disregards aspects of users’ liberty … well … it may be that users will stop using Google.

    My point is that if you provide services based on another business you are playing a risky game. And I think the main goal of a business is to be ‘the best’ in what it does. That’s what Google did. Now it’s up to them to decide what paths to follow (even if we like it or not) …

    I’m not trying to ‘defend’ Google. I’m just trying to be realistic.

  • Rich

    This is why they test things though isn’t it? I can’t see any way they would push this to live across their whole search network with this flaw. The repurcussions would be too huge.