Every blogger loves statistics. It’s nice to know how many people came to our sites, how they got there, what they read, how long they spent and where they went when they left.
But as neat as all of that is, as anyone who uses two or more systems knows already, these stats are far from perfect. Simply put, there’s a lot of wiggle room when it comes to getting those numbers and, despite having over a decade to perfect and hone our stat-gathering abilities, we seem to actually be getting farther away from firm, consistent information.
How so? There are actually too many ways to count. But here are just five ways of the more important ways that the stats you’re looking at every day are becoming less and less accurate and, more importantly, what you can do about it (if anything).
1. Your Search Stats Are Becoming Less Useful
Back in October, Google announced that they were routing all searches for logged in users through SSL. While this was great news for searchers, the secure connection makes it difficult to snoop in on user activity, it created a real problem for webmasters as it means search terms are no longer being passed on in the referral information in those cases.
Though Google initially said that this would affect less than 10% of all search queries, with the rapid growth of Google+, more and more searchers are likely staying logged in, increasing that percentage daily.
As such, even Google Analytics can’t report on what search terms logged-in users are using to find your site.
How to Beat It: The good news is that the percentage is still fairly small, averaging about 20% on my site, and it’s unlikely that the logged-in users search in a drastically different way than the non-SSL ones. You can also look at aggregate search data in your Google Webmaster Tools.
2. Users Are Stopping Tracking
Users have been blocking ads for years, creating a source of frustration for sites that based some or all of their revenue off of advertising.
However, a recent movement aimed at increasing privacy has a growing number of users blocking tracking even if that tracking doesn’t result in a visible ad. Tools such as Do Not Track Plus let users block or severely limit the usefulness of analytics tools, such as Google Analytics.
This trend is likely to continue to grow as the Do Not Track movement picks up steam and more users opt out of having their visits monitored and more analytics systems agree to participate.
How to Beat It: There’s no easy way to beat this one. As with ad blocking, most visitors haven’t taken any action, leaving the bulk of the data intact. However, if visitors don’t want to be tracked, the data you will get on them will be much more limited, especially from third-party analytics systems.
3. Visitors May Never See Your Page
Back in June, Google announced an update to Chrome that would enable pages to be preloaded and prerendered so that, when the user clicked on the link, the new page appeared to come up instantly.
Initially, this was intended to be activated by webmasters using the “rel=”prerender”” link attribute but eventually, with Chrome 17, began to be initiated by the browser in may cases.
However, this isn’t a new feature by any stretch. Firefox has had a similar feature, called prefetching, for many years and Google (the search engine) has routinely integrated it into its results. Furthermore, extensions such as FasterFox have taken prefetching to another level, grabbing all links in the background.
What this means is that, in many cases, your page is being loaded without the user actually seeing it. While Google Analytics and other “smart” systems handle this gracefully using a new API, other analytics systems can not.
This can help lead to the discrepancy between different systems.
How to Beat It: To a server, a page view is a page view regardless of whether a human looked at it. However, Google’s Page Visibility API helps analytics determine if the page was actually looked it by a human being. It may be worth seeing if your analytics system uses it.
4. Incognito/Private Browsing Messes with Visitor Totals
The truth is that, despite the best efforts of groups such as the Web Analytics Association, there’s never truly been a practical definition of simple terms like “visitor” and “unique Visitor”. Though we have definitions as to what one should be, going from the theory to reality has proved difficult.
However, cookies have been one of the key ways from separating a new visitor from a returning one. But while cookies have been useful, all major browsers now offer some form of private browsing, which scrubs cookies after the session is closed.
This means that a visitor who comes by your site while in private browsing mode and returns later, regardless of what mode they are in, will likely be treated as a new visitor. In that regard, this is much like problem #2, but with a twist in that users aren’t blocking tracking cookies, but accepting them and wiping them clean shortly after.
How to Beat It: Analytics can look at other data, such as IP addresses, to better understand who is a new or a returning visitor, but that data if of limited usefulness. The good news is that, unless you have the kind of site that people would want to browse privately, most likely you don’t have many visitors using this feature on your domain.
5. Mobile Browsing Screws Up Everything
Finally, mobile browsing is changing the way we surf the Web and possibly screwing up analytics.
It’s possible, and likely not that uncommon, to start viewing a site on a laptop, pick it up elsewhere on a tablet and then again on a mobile phone. Three different devices, likely three different IP addresses and three different browsers.
In most cases, this is going to be read as three different sessions by three different “unique” visitors. There’s simply no way around it and it’s almost impossible to track users as they switch between devices.
How to Beat It: Requiring user authentication is one way to defeat this, albeit an impractical one for most sites. Instead, it’s probably best to take comfort in the fact that most users don’t switch devices mid-session though they may visit the same site in different sessions across multiple devices.
All of these changes have one thing in common: They’re great for the user. Better privacy, better security and easier reading all end up making for a better user experience, even if it is much more difficult to track that experience.
To be clear, this doesn’t mean the stats you’re getting are useless, just that their applicability is changing. This shift is going to be gradual but it is happening and there’s no telling what analytics will look like in 5-10 years because of it.
One thing is certain, there’s going to be a much greater emphasis on stats that are still easily tracked, such as social networking engagement, inbound links, comments and so forth.
In short, statistics such as “page views” and “unique visitors” are likely becoming even less important just because they are so difficult to track effectively on changing Web. Furthermore, with the shift toward streaming content and a more dynamic in-page experience, it’s likely those stats are of less use anyway.
It’s all part of the evolution of analytics and, in this case, it’s an evolution being driven in large part by changing user behavior and a Web climate that’s a lot less analytics-friendly.