Blog Server Migration Part III

January 13th, 2009

by John R. Sedivy

Introduction
Lessons Learned Blog Server Migration Part I outlined the difference between the commonly confused wordpress server (wordpress.com) and wordpress.org which we are using as the blogging software on our private server. In addition, the advantages of migrating from wordpress.com to wordpress.or g were provided and our rationale for making that decision. Blog Server Migration Part II outlined the challenges encountered with making this migration. Today’s article will focus on the results of our migration as well as the lessons learned that if employed, would allow for a smoother migration in the future.

Migration Results
Unbalanced Scale Our migration from wordpress.com to wordpress.org was nearly flawless and in the end proved to be a wise move – we are continually witnessing the advantages as outlined in Part I. First we made the decision to migrate on December 31, 2008 (New Year’s Eve) for a “go live” date of January 1, 2009 (New Year’s Day). This was a good decision for a number of reasons. We knew that there would be an interruption in traffic as discussed in Part II. During our first month of operation (December 2008) we found that our lowest traffic days, besides the first week of launch, was Christmas day and surrounding days. We had planned the migration to occur around New Year’s as we were anticipating history to repeat itself which in fact did occur.

Daily traffic has lowered since our migration date of January 1, 2009, however total traffic is increasing at a much more rapid rate. Consider this, from our launch date of December 1, 2008 on wordpress.com it had taken 22 days to reach 10,000 total visitors. On January 11, 2009 we hit the 10,000 visitor mark. After the migration we had reached 10,000 visitors within 11 days, or half the time of the month of December. If you take into account that I have not been blogging as frequently as in our first month, this is truly an impressive feat.

It may be noticed that I had listed six disadvantages in Part II as compared to five advantage for server migration in Part I. At first glance it may appear that the disadvantages outweigh the advantages, however in my opinion nothing could be further from the truth. As I mentioned previously, all advantages point toward control. That being said the advantage of control is tremendous and in my opinion, negates the disadvantages. The central theme of control and it’s related advantages should be viewed as being weighted much heavier than the listed disadvantages.

Control is scalable – as our business grows, our private server (wordpress.org ) capability can be scaled with it. Control allows for nearly unlimited possibilities and is really only limited by our imagination, willingness to work hard, and time. The theme here is short term pain for long term gain. The disadvantages outlined in Part II are all short term, perhaps a day or maybe a bit longer, and the advantages remain for the duration of the blog or business and compound over time.

So, in the end has it been worth it? Absolutely. At the time of this writing, we are only 12 days into the migration and are already noticing the advantage. We have able to make our blog more aesthetically pleasing, monetize our blog, and add widgets that increase our capability and automate some of our processes which in old form were done in a time consuming, sometimes inconsistent manner. And we are only scratching the surface. We learn more daily and as such increase our capabilities and as previously mentioned – are only limited by our imagination, willingness to work hard, and time.

Lessons Learned
Hindsight is always 20/20. In looking back there were some lessons learned, some things we did right, others which could have been done better. Here are our highlights for this project, the three that we found to be the most valuable:

1. Migrate On An Anticipated Low Traffic Day. This is the perhaps the single biggest thing that we got right in the migration. We were limited by the amount of historical metrics for web traffic as we had only been in operation for under a month, however we used this data to our advantage. By migrating on an anticipated low traffic day we minimized the anticipated traffic “hiccup” and as a result traffic is getting back on track at a much faster rate. Although there is some short term pain involved as Amy and I found ourselves simultaneously migrating our blog and sipping champagne on New Year’s Eve, but such is life.

2. Research Widgets. Widgets were admittedly a big disappointment. At launch we anticipated increased capability due to implemented widgets, and we did receive such capability, however there were some interruptions. For example, for roughly two hours our stat counter was not working, this was frustrating, especially given that our stat counter was starting at zero as a result of the migration, and we were already behind the eight ball. For most of the two hours we did not realize there was a problem, remember it was an anticipated low traffic day, however, once discovered we corrected the problem in relatively short order. I am not sure what we would have done differently here but having this knowledge would have at least been helpful. I guess a key take-away would be to list your key functionality, such as the stat counter, and test each item after migration.

3. Start With A Private Server From Day One. As with the widgets, this would have been nice to know and is perhaps the most important lesson learned from the overall migration experience. During our first month of operation CCB Buzz had over 22,000 visitors, this metric was lost during the migration. If we would have started on the private server (wordpress.org) on day one we would not have lost the 22,000 visitor metric and would not have had the corresponding “hiccup” in traffic. I see it as that we lost the 22,000 visitor metric, and not 22,000 visitors, as I am sure that at least some of those 22,000 original visitors have been redirected to our new site and are regular readers. Perception is reality and it would be much nicer for a visitor to see our stat counter read 30,000 rather than the 10,000 we currently have at the time of this writing.

Please note the following. When I say that I would have started on the private server (using wordpress.org) on day one it does not mean that I would have started CCB Buzz on the private server, it just means that if I start a new blog it will start on a private server (wordpress.org) from day one. Since CCB Buzz is my first blog, starting on the wordpress.com server was absolutely necessary. It helped us traverse the learning curve at a much faster rate. I am confident in saying that we would not have nearly the traffic growth if we would have started on a private server (wordpress.org).

I liken the wordpress.com server to training wheels. It provides a basket of tried and true capabilities and automated processes that allow for the blog author to focus on content, style, and blogging technique in lieu of being bogged down in the technical details. Having this foundation in place allowed us to focus on our blogging content, style, and technique – which essentially drove traffic growth. Once we had this in place we were able to explore pursuing the private server (& wordpress.org) without sacrificing the other qualities.

Now, if we were to develop a new blog the knowledge would be in place and the foundation established through our experience with CCB Buzz. This would allow us to start on a private server from day one and not experience the short term pain of a transition while experiencing the long term benefits documented herein. My recommendation is that if you are new to blogging or are using this strictly as a hobby, stick with the wordpress.com server. However, if you have grander ambitions such as full-time blogging or monetization pursue the private server (wordpress.org) path as early as possible. Since traffic = money, and time is money – the earlier you migrate to a private server (wordpress.org) the better off you will be.

A standard wordpress.com server configuration has definite short term benefits, however in the long run migration to a private server is key!