One Year Without Google Analytics

In mid-2012 I decided I was fed up with Google Analytics’ bogus data and horrible interface. I replaced the free tool with my own self-served installation of Piwik. Unfortunately I run enough Websites that Piwik put a considerable load on my server. It didn’t help that constant robot probing from hackers and spammers kept knocking the server offline.
In late November 2012 I decided to try the “less is more” approach to analytics reporting. Rather than spending hours every day poring over the minutiae of what people were doing on my sites, I decided to try just looking at the high-level data that might be reported by a simpler tool.
As many of my sites are now built on WordPress the Jetpack module/plugin came to mind. So I installed Jetpack, connected my sites to a couple of WordPress.com accounts, and proceeded to depend on that dashboard for my personal analytics solution.
To say that Jetpack offers fewer details and reporting options than Google Analytics should suffice. People who feel compelled to bury themselves in data every day (even if the data is nonsense) have no reason to look at Jetpack. It was designed for folks who want to focus on blogging, not on chasing keywords.
And that’s all I am going to say about Jetpack.
I still ask clients for access to their Google Analytics data because any metric, even a bad one, is still useful if it’s consistent. I’m not convinced that Google Analytics meets that relatively low standard but since the clients are already using Google Analytics I might as well look at the data.
I can eek out some barely useful reports for most clients from their analytics accounts. If they feel they have been penalized or downgraded then GA gives me enough data (assuming the history goes back far enough) to pinpoint referral traffic events that signal such penalties or downgrades. Even gradual declines in traffic have to start somewhere.
A good analytics package would allow you to create a baseline in your data. To date I have never come across such an analytics package, but one can always hope.
Unlike other analytics software GA has always seemed incapable of providing even the most minimal of competent filtering of robotic data. Still, in a recent site audit I found relatively little indication of robotic activity in the GA data. I had access to raw server logs and so was able to calibrate what I found in GA with some real-world data. But one site audit doesn’t prove much of anything about an analytics tool.
Not knowing when an algorithmic downgrade or manual penalty action occurred is annoying but as an SEO you have to be able to deal with the lack of data. There are other ways to see that a Website is struggling in search. In fact, I have looked at many sites this year for which the owners were NOT collecting analytics data of any kind.
So while it’s nice to have that GA trend line when you’re looking for the precise date a Website lost search referral traffic, the fact it’s not there (either due to lack of analytics or insufficient data) is not a deal-killer. Anyone with 2-3 years’ SEO experience should have an idea by now of what to look for (of course, I say that, but many people continue to fuss over nonsense metrics like “time on site” and “bounce rate” so maybe I should say you need at least 10 years’ experience).
If you don’t need analytics data to figure out what to do about penalties and downgrades, where does it help?
One popular crutch was to use the search referral keyword data to calibrate your landing page strategy. At times this has proven to be moderately useful. Most of the time when I looked at clients’ analytics data I found what I expected: that their landing pages were ranking 1st for most of the queries that sent them traffic, and NOT ranking 1st for their highly coveted 2-word and 3-word queries.
When I have been able to tie the income to the keywords, the income is rarely dominated by those highly coveted keywords. And no off-the-shelf analytics package tells you this. I have seen some nifty custom solutions but because the income data collection is separated from the Web traffic data collection most people who know to do this have to do it in a spreadsheet.
People who keep their noses in the numbers give me all sorts of excuses about why they need analytics data but to date I have yet to see anyone other than a serious data analyst explain to me (in coherent fashion, using complete sentences, supported by charts and screen captures) what they actually accomplish toward a business goal with those analytics.
If you’re manufacturing Web content, such that you need to “plug the holes” in your keyword coverage, the analytics data will help you build the map you need to find the holes. Google Analytics doesn’t give you that map but you can download its data and integrate it into an offline process that produces such a map.
Most Websites don’t need to do this kind of analysis, though. Unless you have about 50 people working for your Website there is no way you can possibly maximize coverage of your potential keyword map. A Keyword Map for a typical Website — when done properly — should run to several million queries.
I just struggle with the concept that the average “Internet marketer” has the time and capacity to produce that much content. Assuming you’re actually good at what you do and you target about 20 queries with every article, you’d need about 50,000 articles just to hit your 1st 1,000,000 queries.
Even the legendary Lance Winslow has yet to produce 50,000 articles. The rest of you are just dreaming.

The perpetual trend in search marketing (and social media marketing and content
marketing) is to collect as much data as you possibly can, tidy it up into some sort of report, and then set goals that are supposed to magically produce incredible growth in traffic and conversions.
I can tell you right now, just glancing at the highest level report in Jetpack and comparing that information with information I had to compile by hand when I was relying on Google Analytics, that SEO Theory is coming off its highest traffic year ever.
And I did that without setting keyword goals, building keyword maps, targeting keywords, chasing queries, or trying to plug holes in some convoluted strategy.
Virtually everyone needs to see the map early in their career. You need to understand that relevance is broad, far-reaching, and has almost nothing to do with those highly coveted 2-word and 3-word “vanity queries” that small business owners, “Internet marketers”, and vice presidents love to obsess over.
The bulk of your search referral traffic should be coming from the long tail of search. Of THAT traffic, a hefty portion of it should be coming from previously unseen queries — at least 20%, and on some sites that number should be as high as 50%. It depends on whether you’re rehashing popular topics (“how to play Texas Holdem”) or introducing new concepts to a concept-hungry search public.
You should be able to see this in some sort of analytics report early in your career so that you know intrinsically in your bones that this is actually how search plays out for most Websites.
But once you know that, you need to move on and get away from the analytics. Yes, someone at your company paid hundreds of thousands of dollars for a high-falutin’ analytics tool that captures all sorts of BIG DATA for your company. The executive team on down want to make sure you are getting investors’ value out of that expense.
But BIG DATA has to support BIG PROCESSES. If all you’re doing is scanning trend tools for keywords and assigning articles to writers, you don’t have a BIG PROCESS. You don’t need BIG DATA. Just pay lip service to the analytics dashboard (if it’s not tracking money) and keep your head in the editorial calendar.
If you cannot see the return on investment in analytics that is because there is NO return on your investment. In some cases this just may mean you need to learn how to use your analytics tool better. In most cases it means you have set your expectations too high.
But the bigger challenge in measuring ROI for use of analytics tools is deprogramming yourself from believing that all that analytics data is telling you something useful. If you know which pages are receiving traffic and why then what else do you need to know? Just look at the reports once a month. Once a year may be better unless your sales drop off.
If you’re not actually selling
anything then what are you trying to do with your analytics?
The less time you spend in mesmerizing data that doesn’t tell you a story, the more likely you’ll be able to ask yourself intelligent questions about what your Website is doing and should be doing.
In other words, before you go into the data and the report, you need to know what you need to know. If you have no questions about your site’s performance then why are you wasting your time looking at analytics reports?
Sure, something might
go wrong but until something does go wrong you’re wasting a lot of time in “What If…” dreamland. You need to keep your feet on the ground and your attention on what you need to be doing to be productive.
One of the chief reasons why people created crappy Websites that led to the Panda algorithm was that they were allowing the analytics to drive their search marketing strategies. You’d think some folks would connect the dots and say, “Oh — we’re not supposed to be mindless keyword-chasing robots” but apparently that is too hard an equation to work out.
I don’t expect to reach much of the vice presidential crowd on this score because they have meetings and have to mentor their underlings in how to condense a lot of information down to 1 or 2 Power Point slides. In fact, it is NOT the vice president’s job to come to the conclusion that his team is spending too much time in the analytics data — it’s HIS job to present that finding to the executive board AFTER the team produces the 1-2 slides that show just how much productivity is lost while people twiddle their thumbs in analytics data.
In other words, waiting for someone else to tell you the obvious just makes your problem worse. I was pretty confident when I gave up on Google Analytics that I would not need it. I was right.
If you have been doing SEO for 2-3 years then you should have figured out by now that it’s not all about analytics and measurement any more than it is all about links.
Don’t make me say you need TEN YEARS’ experience to figure that out.


View the original article here