On Making More Website Changes

It was time.

Three months ago I implemented a radical redesign of the website. After pouring over usage statistics I noticed several broad patterns of the way readers, like yourself, were interacting with the content here.

First, there was a core group of readers who hit the website regularly.

Second, there were drive-by readers who were coming in via Google and other search engines, finding (or not) what they were looking for, and then either they would stay and read more, or they would leave. More often than not, though, they would leave.

The problem with number two is obvious. How does one hook the drive-by reader to stay?

I was using Brian Gardner‘s Vertigo Red theme, but one of the issues I had with it was that “landing pages” — those single-post pages that the search engines index and link to — weren’t exactly conducive to getting someone to stay by digging deeper into other things I had written.

What if I tried a different layout? One better designed for the “drive-by reader,” even?

So I made a change. I uploaded the venerable Hemingway WordPress theme and went to work modifying it to my needs.

For me, a compulsive tinkerer, Hemingway was golden. As I wrote a few weeks after activating the theme, the theme was no longer in active development, and there’s certainly no support for it. Which meant that if I wanted to do something, I had to code it myself. Previous post links. Next post links. A random post link. Related post links. Gravatar support.

The single post pages in Hemingway were inspired in their design — on the left was a “meta” area where things like links to posts forward and back could be listed, along with information on when it was posted, categories, tags, and the like, while on the right was a nicely designed content area.

The problem with Hemingway? It only gave excerpts from the two most recent posts on the front page. Excerpts. Not full posts.

Hemingway is a different sort of blog layout. It doesn’t feel like a blog.

There’s some obvious problems with excerpts of two posts on the front page. The major one being that content falls off the front page very quickly. If you make two posts in a day — not an uncommon happening for me — then previous content essentially vanishes from the casual view. Though Hemingway offered a link to recent posts in the footer, that’s not really a solution; it’s more of a stopgap. I recoded the index.php file to give the most recent post in full, and then excerpts of the next four most recent posts. Again, not really a solution, but it was workable.

Then I wanted to see how Hemingway would affect traffic, and the way readers approached the website.

Within a month I knew — switching from Vertigo to Hemingway resulted in a loss of traffic.

A forty percent loss in traffic.

I can think of multiple reasons for the loss in traffic. Content would get missed in a search engine trawl. Or, worse, what the search engine saw would no longer be there on the front page.

Last weekend I decided. It was time. The experiment with Hemingway was fun. I learned a lot. I really did like the way the website looked. But Hemingway wasn’t serving my needs.

Brian Gardner had released a new version of Vertigo, so I downloaded that and started working with it. I wanted something that looks lightweight on the screen because it’s lightweight under the hood. One thing I’ve found with the Vertigo code is that it’s lightweight. It’s not overly complex. You can look at it, and it makes sense. I used to use Tarski, and Tarski is not simple under the hood.

There were three things I needed to do for it to make it work as I needed.

  1. Rework the comments.php file to separate out the comments from the trackbacks. Admittedly, I don’t receive many trackbacks, but I like having them separate. Comments are for people talking about what I wrote here, while trackbacks are what people are writing there.
  2. Vertigo, at least the version I downloaded, is a three-column layout. I wanted to change the middle column — the left-most sidebar column — into a metadata area on single-post pages. In that column I would have links to previous posts, next posts, related posts, and the like. Essentially, I wanted to recreate the Hemingway single-post page layout, to make navigation within the blog easier.
  3. Finally, I needed to create some custom page templates. Not just for Tag Clouds and Archives, but I also wanted to create a special template for my bibliography page.

The latter, by the way, is interesting — at least to me. It’s not interesting for the content, but it’s interesting for how the content is generated. I wanted to take advantage of a WordPress feature I’d never really touched — the content management system built into the core. In my early WordPress days, static pages had to be created by hand and lived as files outside of WordPress, and I had never moved into the new system where static pages existed inside the WordPress database.

I fiddled with the files on Saturday and Sunday, and made the changes to the files that I needed. Monday and Tuesday I made the changes to the stylesheet, implementing a color scheme and header graphic I had designed for Tarski a year ago but had never used.

The end result? Something I’m happy with. 🙂

Will the web traffic come back? I don’t know. A forty percent drop is rather precipitous. So we shall see.

How long will this stay? Who can say? Who knows — maybe it will stay until I put the Wonderland theme back up in November, just in time for Christmas. 😉

Leave a Reply

Your email address will not be published. Required fields are marked *