Background

So, I’ve been running this site for around 1 month now without hitches, the homepage sorted and the page structure hieracy as it should be.

I’ve had analytics on it from the start.

My current set-up is:

Apache server 2.*
WordPress v 3.5.1
PHP 5.3.3-7

I’m not new to this game, I’ve been running a blog, this blog since 2003.

This month has been the first full month when I’ve been able to sit back and analyse things being happy that: the platform, the site, the applications and the product as a whole was working as it should be (sitemaps and all).

Method

Currently the method I use to analyse stats falls under two methods:

1. Google analytics: Using the WordPress theme editor, I go in and edit and insert the analytics code below the http header at the opening of the head tags.
   The first thing to hit after definitions.
2. Using log files: Law log files downloaded directly from the server. Used Weblog expert lite logfile analyzer lite in this case, (although I know there are better software’s out there that can provide a better job).

Results

Using the month of April. We are going to look at the disparity and breadth of information, accuracy in information and trust index for the two products.

Looking at the Key Indicators for site metrics: For a blog taking the key data as follows:

Visitors

Taking a look at the visitors graph.

You can see the Google graph scraping along, showing no improvement over the course of the month.

On the right, the Weblog expertlite graph shows a pronounced increase toward the end of the month (co-incidental with my inclusion of sitemaps to both bing and google).

Google shows roughly 12-14 visits per day on a 500 page+ site, Weblog expert showing 200-400.
In this month, I introduced a full set of comprehensive sitemaps in the prescribed methods to both Bing and Google.

Entry pages

One of your key indicator stats.

Possibly the most important, taking time as a consideration, as to the way you want your blog or content to go.
If your metrics software can provide you with an accurate picture of this, (combined with the other key stats on here), it’s the key to the success or failure of your site.

Google analytics: As you can see, effectively, the only URL the Google analaytics only show the /blog/ root directory getting hits of any significance. This URL is effectively a 404, and as a landing page, insignificant.

Weblog expert lite: Shows the wp-cron activities of my wp app. Not very helpful. Stats like this clutter up the real information, and on the full version can be removed by setting up filters. The next stat down is the ‘https://www.michaeltyler.co.uk/ etrex-garmin-gps-to-google-maps/‘, traditionally, as an ex-geoblog, on of the biggest attracting USP’s of the site in terms of content. This post used to tell of how to convert a set of waypoints directly from a handheld device, such as an eTrex to your Google map.

Search phrases

Another key indicator. Where the traffic that comes to your site comes looking for. Having a grip on what people look to you to tell them about? Any use? Maybe not if your analytics software is telling you erroneous or useless information.

Check this out.

Google analytics: Google, as you can see, is passing what it’s managed to gather, which seems to be very little. Taking into account, this is only ‘visitor’ referrals not calls on the server; it appears there’s actually very little of interest on my site. Most pages only gathering a maximum of one search phrase.(!?)

All spam in effect?

Weblog expert lite: Weblog expert lite shows a different picture altogether. Going from requests on the server, as opposed to ‘visitor’ stats, I can see a variety of searchphrases pointing at searches like ‘manastir ostrog’, ‘montenegro’, ‘ostrog’ and all manner of phrases using  the giant walrus I blogged about https://www.michaeltyler.co.uk/i-am-the-walrus/, (which actually exists in a zoo in Dortmund), or the shot of Kim-Jung-Un’s Penis. All of this is helpful to me. I can use these images to promote my site by inserting some accreditation in the way of a watermark. Plus I am paying for the bandwidth, it’s placing a load on my server, and it’s costing me money.
I want to know about it.

Top referring URL’s

A list of where your requests are coming from. If there’s juice being sucked out of your site, this is where it’s going to.

Weblog expert lite: Weblog expert lite traces the top 50. Two of these I’ve tracked down and asked to comply with the sites copyright terms and conditions. The rest are easy to ban using the robots.txt or .htaccess file if they become anything more than an annoyance.

Google analytics: No stats available.

Error reporting

Sometimes, a lot of the time, you’re going to find that visitors are going to find parts of your site which are no longer there. They may have changed, changed location or disappeared permanently.

Whatever has happened, it’s up to you to track these visitors and guide them along their in the best way you see fit.

Weblog expert lite: Weblog expert lite sees a number of 500 server errors, all occurring on the same date. It reports a standard amount of 404, (below the hundred mark), which is probably about the normal level. It also shows me the site was having a WordPress error.

Google analytics: Google tells me, nothing. Yep. Absolutely nothing about who’s reached my site in error. Nothing about about the dates or times they’ve visited and found it on error or the errors they’ve experienced. Not the best, you can imagine. Not even professional, and certainly not acceptable for anyone running a professional site.

I’ve started using this service, pingomatic, based in Sweden, they monitor your site status and response times. Currently one site is free.

Browsers

The browser is the users first point of entry onto the world wide web under most circumstances. Information on the browsers people use are going to be important for issues such as

• Compliance: How your site functions according to W3C standards
• Design: Not much use if your site runs well on only i.e. 3.x (back in the 90’s sometime) and you’re rendering using <font> tags! Nor if you only render on the MOST up to date versions of ie, firefox or chrome. You’ve got to be aware of percentages involved in the various release brackets and cater for the maximum breadth of browser platforms for the minimum potential cost.
• Visitor OS platforms: Modern OS platforms, each has it’s own browser type. This is especially true of the touchscreen technology which is proving more and more popular for those accessing the medium.

Weblog expert lite: Weblog expert lite has some drilldown, from the stats it’s easy to identify there’s been a sudden pick up in the usage stats from safari mobile browser. Also activity there from the Android and Galaxy mobile platforms.

Google analytics: Looking across to the Google equivalent, I can see we’ve got some stats which seem to tell exactly the opposite story. In terms of mobile users, Google analytics tells me their Android platform is providing most of the traffic. Indeed this maybe true of using the 14 visitors a day site sample they use, which I also believe is an error.

Summary

I have some reservations about the way WordPress published the tracking code. But having checked pages on site, I never came across a page where the Google code wasn’t present.

Google is woefully under-reporting my visitor numbers.

It’s not reporting key metrics or reporting them inaccurately.

It’s not able to provide information on who is leaching my site and causing possible infringements.

It’s not able to tell me about potential errors, or when they’ve occurred.

Recommendations

If you’re using a WordPress site and you want to boost your readership, seo effectiveness, metrics analysis and intelligence. Consider using log file.

As my little experiment has proven, the stats you get from logs are uncovering a number of key seo components which Google has overlooked, doesn’t care about, can’t measure, or is looking to make money on in the future and won’t include, (that wouldn’t be like Google would it)?

To take the leading hand in what your blog’s steerage is, you’re going to need the whole and accurate picture. Maybe this means drilling down into the minutiae of your referrers raw logs, getting your meta-hands dirty, spending a little time. In my eyes something that’s infinately better than being handed things by Google which are plain wrong.

For me and my WordPress blog site, log files for now, at least until I hit around the 10,000 per day mark again. :)

ps. If anyone else has had any experience about implementing Google code on the wordpress platform. What kind of metrics you’ve used to make your blog more popular, and whether you think the stats are better or worse than more traditional methods, I’d be interested to know.