Why web analytics data is better than 99% of other stats

Regardless of the tool you use today to measure your website traffic, even if it’s an antiquated server log parser (boy I’m dating myself here), you can be proud that the information you’re collecting is by far the cleanest statistical data on the planet (booyah!).  Many will argue that there are still a ton of flaws associated with web analytics data, but truth be told, information available to a web analyst today is far more trustworthy than the majority of audience measurement options out there.

This post will probably be WA 101 for some of my readers, but with a new blog I have to cover the basics early.

Opt-in versus Opt-out

Data collection is often a question of fishing for people to participate in a study.  After all, what’s better than observing behavior first-hand from willing participants? Well, a lot as it turns out.  It can be pretty darn tough to phrase the perfect survey, conduct polls, and formulate focus groups.  All three of those can also be super expensive.  In the online world, the equivalent of a focus group is an opt-in panel, such as employed (sometimes literally) by research giants like comScore, Neilsen, and Alexa to name a few. A few obvious problems with any panel-based data is non-representative sampling and bias.

Web analytics is different.

Most users don’t even know information is collected on websites.  Oh sure, upon asking them simple questions they think Facebook is evil because of all the privacy concerns the media has got a hold of, but otherwise the majority of internet users are oblivious. As visitors become more familiar and comfortable with e-commerce, they soon realize that their information is tracked because leaders like Amazon will start recognizing them when they visit, or recommending products other people liked.

To the average person, web analytics is an opt-out service that is beyond their comprehension to disable.  Or the convenience of cookies and other beneficial features that go along with website tracking outweigh the perceived risk.  By default, web analysts have a huge sampling of highly representative (albeit not 100% exact) behavior that is highly unbiased.

Usually highly precise

Precision can be thought of as a measurement’s reproducibility.  Can you get close to the same number again and again, regardless of what tool you use?  If you cannot, does the data at least trend similarly?  Peaks and valleys in the same places over time?  If so, you’re already winning over more traditional measurement of media.  That’s not to say we don’t care about accuracy, but it’s becoming less important.

Accuracy is the ability to find the real-deal number.  Many have said that web analytics isn’t accurate because of this or that…

Wha-chu-talkin-bout Willis?

Last I checked, the accuracy of web analytics data is never used to settle legal matters, meet financial disclosure, etc.  Sure, we can use analytics to give us a ballpark number of orders on an e-commerce site, but your analytics will never account for all activity in a business.

That’s an important hard fact to realize.  You don’t have to sweat it if Google Analytics tells you one thing, and Omniture or Webtrends tells you another.  It’s okay to be different.  Let’s all sing kumbaya.

Correlate shoe size to video views

Heck, why not?  There are so many different ways you can measure an audience, you’re going to want to use web analytics for everything.  Based on subtle signals visitors use to help themselves, we can better understand and apply mechanisms to capture that information and feed it back into creating better user experience.

If you really wanted to, you could capture shoe size on an e-commerce store and keep that information handy for that same visitor over the lifetime of their tracking cookie.  And cookies can stick around for a long damn time.  One of the oldest cookies I’ve tracked was more than 3 years old, and the lifetime value information attributed to that visitor was amazing to behold.

But let’s not limit ourselves to the realm of websites, because today we can correlate and apply measurement of offline (or nonline) response as well.  Coupon codes for in-store offers, inbound call tracking, and even our cell phones themselves, all give analysts a wealth of information.

Aside: it’s no coincidence that Google is entering the mobile phone business.  Food for thought.  Information is power, and they’re quickly becoming Skynet.  Fashion the aluminum foil head wear!  (joking, of course, Google, I love you guys/gals)

Everyone else is doing it, and so should you

I think the stat that is currently floating around on Twitter is that 28% of websites worldwide are using Google Analytics.  Let’s toss another 10% on there for paid analytics tools such as Omniture, Webtrends, HBX is still kicking long after death, Citi, Neilsen, etc.  The majority of webmasters don’t know that the information they possess goes well beyond a comprehension of popularity in a vacuum, or my favorite, “hit counting”.

And that is what we’ll talk about in coming posts, turning a bunch of “prudy numbers” and “purdy graphs” into cash.  Cash cash cash.

2 thoughts on “Why web analytics data is better than 99% of other stats”

Leave a Reply

Your email address will not be published. Required fields are marked *