Daegu has emerged to be the most trafficked page for us, other than the mainpage. We actually have a shitty ranking on Google for the keyword ‘Daegu’. I see that we are ranked #24, based on most Google Datacenters. But we do well in our long tail keywords, which are naturally more easier to move up in rank. If there is any reason not to fret too much is that ‘Daegu’ on its own, doesn’t get searched all that often. Last time that I checked Wordtracker, the keyword was only being searched 17 times a day. And depending on how you interpret site stats, we seem to be doing alright without top ranking on the cities main ‘keyword’.

Now, I don’t want to bore with explaining the pros and cons of this tracking tool vs the next. But no tracking tool shows the same results. And some are widely differently from each other. It’s like if you have 5 people watching a mystery drama that ends in a cliffhanger, each one will have different takes on whats going on and what info is left for the imagination to decide. Well, site tracking tools are not that scattered, but pretty close.

Tag-based tools like Google Analytics allow you to add html tags to all of your pages. This helps them provide detailed traffic analysis and how users navigate around. But its not without its drawbacks. Google Analytics (GA) code on the pages requires users to have Javascript installed, in order to be tracked by it. If you don’t have it, you won’t be counted. But the number of people who don’t have Javascript installed these days is not much. And then there are those who prefer to surf with Javascript disabled. And there there are other hitches that add up. If the user visiting your site is using a browser that doesn’t have cookies enabled, GA will still technically work, but the stats will be less accurate. Awstats and Webalizer are log-based and track EVERYTHING that hits your page, although Awstats filters out bots and spiders. Webalizer makes note of the referrer URL where you can see the footprints of the spiders, but you’ll have to manually subtract from the totals.

I haven’t tried Webtrends yet, which appears to use both tag-based and log-based means to measure your traffic. But the free stuff (Google Analytics, Awstats, and Webalizer) all need to be used in conjunction to get a ballpark verdict. The truth is usually somewhere in the middle of GA and Awstats. But as GA stat are more prettier to export, here’s a screenshot of the traffic that ends up on our Daegu page.

The green line shows where are traffic was last year and the blue one is traffic at this time. So we seem to averaging around 50 unique pageviews a day on the Daegu page (Webalizer minus spiders/bots says that we get 200+ pageviews to the Daegu page). So the 17 that we are not scooping from a high rank on ‘Daegu’ for Google searches doesn’t seem too terrible. This also doesn’t reflect our long tail keyword stats for Daegu pages that are off of the main one, such as Daegu nightlife (#3 on Google) and Restaurants in Daegu (#1 on Google), despite them not getting daily searches, but moreso weekly and monthly searches.

The Daegu wiki surely needs to be optimized more for SEO. But its nice to begin to see word-of-mouth disemminate, as I’ve met people offline in the city, who use it. And we had a some recent wiki edits and page additions from people in Daegu, as well, which is nice, as I’m in Seoul now and am out of touch with life down there.

I’ve noticed that the hair at the crown of my head is now getting sparse, despite thinking that the male pattern baldness gene passed me by, given my 35 years of a full head of hair. If it is not coming about genetically, then the past 3 years of server headaches may have been playing some role.

If I recall, our sites have been in as many as 7 different server environments since we started. We have a very finicky family of sites, where there always has to be at least one site that must bitch and complain in the form of irreconcilable DB or encoding issues. Well, nothing is every irreconcilable, if you have unlimited funds to press on and bring 2nd opinion and 3rd opinions in, and sooner or later after bring 10 people in to analyze the problem, someone is able to find the right recipe. We only have the luxury (which is still far better than most sites that have tricky setups) of having the in-house support team who governs the servers, plus we have a 3rd party server management firm who does work on a flat-rate monthly basis. If those 2 layers of support can’t find a way to resolve an issue, then the whole thing goes out the window and we re-approach.

The current issue is that we have recently moved from our dedicated server to a cheaper VPS (virtual private server). It’s a cost-savings move, given that I’m moving to Canada soon and a to-be-determined job situation to help pay down some site debts and maintain monthly costs. The VPS works well with all sites under our umbrella, except for the forums. The Korean language encoding was not working there and the 2 layers of support, plus my regular developer couldn’t solve the riddle. So the re-approach was to send the forums (along with theholygrillkorea, which also had an unrelated DB issue after moving to the VPS) to a cheaper, shared server run by the host who originally was managing the dedicated server. I figure that as their server environment and server specs are identical between their dedi and shared servers, that the encoding issue will most likely to resolve there.

So we just moved the forums and theholygrill to the shared server this weekend and I changed the DNS to point the subdomain to the new server. It’s not resolving correctly, as we are seeing a temp page, but the host is looking into it. Once the DNS is resolving correctly, we’ll be able to see the forums in their new home and see if the encoding is working correctly. Fingers crossed, as this is a difference of having to pay $270/mnth to have everything working fine on the original dedi server, to having the family of sites operating on the VPS and the shared server, for a total of $70/mnth (which includes the cost of both layers of support).

Edit: The forums and Korean encoding are working fine now on the shared server, so that’s stable. One hitch (of course) is the calendar of theholygrill site is crapping out and will cost about $110 in manhours to fix. Both developer and server admins are unable to stabilize it, even though it worked fine on the dedicated server. I’ll see if we can try another migration of those files to resolve the issue. But the dedicated server is going to be closing down next week, regardless.