This blog has reached 100 articles for now. I'm taking a look of what I have written and also go over the reading statistics.
I started actively writing in 2013 when I realized that a blog is an essential marketing tool for a freelancer. It is not my main channel (those are my network of friends, tech-focused events and IRC) but it works well as a secondary channel and a place to show someone the things I have built and worked with.
The blog also gives me a chance to express myself and contribute back to the community by helping others with the solutions that I have found for the technical problems that I have faced.
With 100 articles per 2 years I get out an article per week on average. Writing is easy for me but some articles take lots of time, especially when your references look questionable and you want to find better ones. I write mostly about (by the Tags) page:
- upgrading something
I prefer to write short and clear articles and use original references while also providing my own opinion. Such articles take less time to write than something that is purely based on my own opinion. Articles about tools and solved/worked around issues fall into this category.
It takes me about an half day to write an article (including Internet research for references).
I use Google Analytics but extracting the reading statistics is not very easy due to referral spam. Referral spam generates huge amount of non-human pageviews that skews the statistics for smaller sites/blogs.
I had to create a custom report for my "All data" view with a filter to exclude referral spam visits. Creating a custom view would not help as it does not apply filters to the past data. My filter expression looks like this (wrapped into multiple lines):
avocatnet|business|buttons|buy|cheap|chinese|darodar| dollars|free|guardlink|kambasoft|market|monetizer|money| offer|semalt|seo|sexy|smailik|tracking| traffic|webmaster|--
This was put together by looking at the Acquisition->All Traffic->Referrals report. Most referrals with the Average Session Duration 0 are likely to be spam referrals. I generalized the domain names to avoid listing many that are specific cases of similar-looking spam domains. This filter expression can be overly broad for other sites/blogs and I do not recommend using it unless you know what you are doing.
Another type of spam is so-called Ghost spam. This is executed without even visiting your site but by launching requests directly to the Google Analytics server using your tracking id. A way to filter out ghost spam is to filter out all requests that have non-matching host name (something that is not your server).
I exported my result in a CSV format and turned them manually into HTML tables below (you cannot use GA API for custom reports!):
The list of top pages is dominated by the articles about specific issues. This is followed by the articles about tools and upgrades. This is quite expected as these topics match the top tags I write about.
|Source||Sessions||Avg. Session Duration||Pages/Session|
The majority of traffic comes from Google search. This was expected as the most of the solutions I have written about can be found directly from the top results when Googling for the problem. The blog has been optimized for search engines.
The next important source is Twitter (t.co). My Twitter account @RaivoL is strictly for technical subjects and I post my article links there.
Below Twitter is hired.im (now closed). It was a freelancer listing site (costing $1 per month). It did not become very popular but was well-promoted at some time. During that time I was listed on the front page, on top of the list. I guess I got very lucky with the time of registration. I got many good contacts out of it.
The next sites are a some kind of RSS reader (cyclur) and the SWI-Prolog's home page. I'm the author of many SWI-Prolog packs (some of them, like the Markdown parser, were crucial for operating this blog). I get Prolog job offers (this is a niche market, not filled well) thanks to this publicity. My current active project is an online CAD system using SWI-Prolog as a backend.
The last sources in the top 10 are search engines (bing, duckduckgo), my shell company's homepage (infdot) and Facebook. I had a Facebook page at some time (RLa Tech Blog) but I decided I wanted to keep Facebook only for personal matters.
No freelancer sites
Note that there are no freelancer sites (besides hired.im where I ended up in the absolute top, at least visually). I have an account on most of the large freelancer sites (Odesk, FreelancerMap, DoNanza, Zerply, FreelanceSwitch, GroupTalent, Freelancer.com, SimplyHired, Dice.com, Monster.com, Indeed.com, etc.) but I get no visits from there. Some of these sites do not even allow to post an URL to a site/blog that would allow to bypass them as middlemen. I added these accounts in 2013 while running out of money and desperately looking for job. The most time took a profile on Stack Overflow Careers. During the last 2 years I have only had a single referral visit from there!
|Browser||Sessions||Avg. Session Duration||Pages/Session|
The most used browser is Chrome. It is a bit surprising that IE (all version combined!) shows so little usage.
Top Operating Systems
|OS||Sessions||Avg. Session Duration||Pages/Session|
Windows usage is only 50% (the table does not show it but numbers add up to that). However, this might not be surprising for a tech site. Mobile OS representation is low.
|Browser||Sessions||Avg. Session Duration||Pages/Session|
The increase in session count is quite visible although the average session duration has gone down. From these numbers I can estimate the session count of 8000 to the end of this year.
Unfortunately I enabled integration with Google Webmaster Tools only a month ago and keywords (that are provided by it) would not be very meaningful as they only cover very recent searches.
Google Analytics contains a lot more data but I did not find anything else particularly interesting. If I find something important or get some data to compare, I will make a follow-up post.
Summary and the future
This is my first lookback on the blog. While there might be not many people interested in these numbers, this reflection gave me some insight about the direction where the blog is heading to. I like the current course. I enjoy writing and the complete feeling of solving something, writing about it, and making the solution available for everyone. And sometimes I get good contacts out of it. I'm definitely looking forward to write the next 100 articles.