Back in October last year this site experienced a sudden drop in page views, and as yet I’m still not certain what caused it. Since then, traffic has remained around the 400 views per day mark, and although I had slightly more hits in March than previous months, it’s not yet back to pre-October levels.
So recently I’ve been looking into ways of turning this around, and this blog entry is here to detail what I’ve done.
Improving site speed
Since switching from Movable Type, which outputted static HTML files, to WordPress, which uses dynamically-generated pages from the database, the speed of the site has gone down. Though I kept things lean by not using masses of plugins, pages were still loading quite slowly. Slow-loading sites are generally a bad thing – more than a few seconds and people are more likely to click away – but I’ve also heard that Google has started penalising slow sites in its search result pages.
I recently upgraded Debian Linux on the server from Lenny to Squeeze, which in itself brought some noticeable improvements, and also meant I could install APC – the Alternative PHP Cache – which improved things further.
I’ve always used WordPress plugins sparingly, and disabled any that I’m not actively using. I verified that plugins weren’t slowing the site down significantly by using P3 (Plugin Performance Profiler), a plugin by GoDaddy (ew…) that calculates how much overhead each plugin adds to your site’s load times. As it happens, P3 itself was slowing things down the most, so it’s disabled until I install any additional plugins. The next biggest hog was Jetpack – I do use a few features of Jetpack (namely the stats, email subscriptions and sharing buttons) so I need to keep this. In any case, it’s adding mere milliseconds to loading times which doesn’t really bother me much.
When I first switched to WordPress, I went for WP Quick Cache as you can pretty much install it and forget about it. But it doesn’t appear to offer the best performance when compared to other caching plugins. WP Super Cache is probably the most well-known, but I went for W3 Total Cache due to its better reviews. This will store static HTML files on your server’s disk, and will also run all of your scripts and stylesheets through a ‘minify’ system to reduce their file size and combine them into one file, thus reducing HTTP requests. I’ve only been using W3 Total Cache for a couple of days but it seems to have again made a difference.
Under Movable Type, I had a Google Sitemap XML file, which was then submitted to Google through its webmaster portal so that the GoogleBot crawler would know where all of my pages were. Unfortunately I forgot about this file when switching to WordPress and so it stagnated. Though the GoogleBot regularly crawls new pages on the site, having a stale sitemap is probably not a good thing. At first I used BWP-GXS but have since installed WordPress SEO which also provides this functionality, albeit with fewer options. I also hadn’t notified Bing of the existence of the sitemap, so I registered with Bing Webmaster Tools and submitted it. Though Google has 80% of the search market, I’ve had very few clicks through from Bing and crawler issues may have been the problem, so hopefully it will use the sitemap to add to its index.
I mentioned the WordPress SEO plugin earlier. I’d previously been hesitant to try such a plugin as I was worried about the effect it would have on the site – I didn’t want it to appear like a keyword-stuffed thin content site. But this plugin actually works well without impacting on your site too much. Unlike the more popular All in One SEO Pack it requires more configuring, but should therefore give you more control. I’ve used it to help to reduce content duplication between pages (or at least ensuring that duplicate content isn’t indexed by search engines) and set keywords for individual entries. I’m not going to start changing post titles to provocative things like ’10 ways to improve your WordPress blog’s search ranking’ as that’s not my style, but hopefully a bit of optimisation will help.
So this is what I’ve tried so far. Time will tell if it was worth it.