WordPress 2.7, Google Analytics and automatic upgrading

December 27th, 2008

I love the new version of WordPress – the new dashboard is immediately appealing; it looks like this:

The WordPress 2.7 dashboard in action

The WordPress 2.7 dashboard in action

 

As you may be able to see the primary navigation has switched to a vertical sidebar on the left, and the main area shows you a summary of activity, recent comments and even has an area that I’m using right now – Quickpress – that allows you to add a speedy post.

I updated from the previous version using an automatic update facility provided by Fantastico De Luxe, a facility that comes courtesy of my hosting provider.

I’ve now worked out that the update includes a change to the footer template, which is where I had put my analytics code. The automatic upgrading is not smart enough to pull out the bespoke changes to the file and put them in the new version, so I have to be smart enough to remember to do it myself. 

I suppose that the alternative would be to put the code in a separate file, and ensure somehow that all pages called that file. I wonder what other bloggers do?

Google analytics

December 21st, 2008

It just goes to show how little I am concerned about the traffic to this site, but I’ve just been into GA and I note that at around 5th October I managed to remove the GA tracking code from the footer of the blog pages, so that content is showing no activity for all that time. 

I’ve now put the analytics code on again, in the footer. This may well be a dull system message, and of interest to no one but myself, but I wish I’d written one of those when I did whatever it was that I did that caused the code to drop off the template. My suspicion rests with an automated WordPress update: if I were anything other than lazy I would not use automation for such things…

Yaab 1.1 release addresses issues

December 21st, 2008

The amazing Satheesh is at it again – perhaps encouraged by some feedback, he has already released an updated version of his autoblogging WordPress plugin Yaab.

I think that there may have been a misunderstanding previously about “duplication” – I have never seen the same item twice in one blog post – that would only happen, as Satheesh points out in a comment, where there was already duplication in a feed. 

My concern was that the plugin would publish the last X items (X is configurable) whether or not there were any new items available.

However, it seems that this problem has been resolved in the latest release. The relevant part of my log file is as follows (the most recent item appears first):

2008-12-21-08:45: BBC Main News 2 is fetched automatically. No posts published
2008-12-21-08:33: BBC Main News 2 is fetched automatically. 1 posts published

This is great. Yaab now checks in, and if it finds nothing new in the feed, it leaves it alone…

There are some other welcome new elements in the release, which I will assess in a future post, but it is all looking good.

Yet another Yaab postscript

December 21st, 2008

I’ve removed the multiple BBC-generated posts from the test site, as I like to avoid having duplicate content (unless I’m actively testing duplicate content issues). I left the 4th edition live, so that it is clear that there was originally more than one such post.

Yaab autoblogging test results

December 21st, 2008

And the results are in:

1) Duplication of content: I set up an autoblogger using the BBC news main feed, with an interval of 10 minutes. It published four separate posts, all with identical content. So, Yaab does not check for new content before publishing a new post. To be fair, the author doesn’t claim that it does – but this goes on the wishlist for the future. The perfect autoblogging tool would be something that you could leave alone entirely, without having to make guesstimates about how often the sources were likely to be updated.

2) The SEO Smart Links plugin works fine with Yaab. I have set it to publish only on individual posts and not the main page. I will post separately about this plugin, which is itself a fine piece of work.

3) The plugin worked beautifully with an RSS feed from Flickr, displaying pictures from the feed along with links and credits.

4) Multiple feed URLs worked well: although differently sourced RSS content displays differently. In particular, the content from a Google News RSS feed looks different. This may just be a feature of those feeds, but I think such issues could be ironed out with greater configurability of the data coming in from the feeds, as already suggested by another early adopter in his useful feedback post on Yaab, all of which I would agree with.

5) Adsense content units have gone on to the individual blog posts: they are mostly showing generic ads at present, but there is an advertisement for some sort of Goth/Emo dating site on the aggregated “glam rock” entry, so it seems that Google is picking up the content (as I would have expected). In time I am sure that I’ll see more relevant ads on all pages.

So, pretty impressive so far!

Yaab – next steps

December 21st, 2008

As mentioned in a comment to an earlier post, following the author’s advice I have now been able to get the Yaab Autoblogger to work successfully with a Yahoo Pipes feed, by substituting the “feed” prefix with “http”, which is great.

The next thing that is really great is that setting the interval for fetching new content and posting it also works a treat. In my impatience to see it working, I set the timer for only 20 minutes, and it produced another post right on time. 

Now, 20 minutes is probably not the best time to set, unless you are referring to feeds that are updated extremely often! What happened in this instance was that it fetched exactly the same items that it had got 20 minutes earlier, and created an identical post (which I have now deleted to avoid duplication). What would be a really great feature for this plugin would be if it were to check whether there were any new items since the last time it fetched the feed, and only create a post if there were any such items.

However, I should probably test this properly: I made some other changes to the feed (the name of the post in particular) and it may be that it would have handled the duplication better had I not done so. 

So that becomes the first task on my Yaab testing to-do list, which follows:

Testing to-do list

1) Create a new autoblogger on the test blog that updates frequently and let it run without making any other changes; assess how the autoblogger handles non-updated material

2) Install the SEO Smart Links plugin and see whether and how this works with Yaab – the ability to add in my own predefined links on certain words/phrases is a feature that I liked in WP-o-Matic, and has obvious potential SEO benefits. Satheesh mentions this as a future upgrade to his plugin also.

3) Test the plugin with an images RSS feed.

4) Try the plugin with multiple feed URLs

5) Test the output blog posts with Adsense to get some insight into how Google treats the output content

Yaab PS

December 21st, 2008

I should probably also mention two other things: I’ve not installed the plugin on this blog, but on another one, but there’s not much point looking at it at the moment because although the Yaab autoblogger worked when I fetched the feed manually, I removed the piece that it posted because it was a full-feed post, republication of which would be against the T&Cs of the Guardian newspaper.

Yaab Autoblogger – first impressions

December 21st, 2008

I am probably taking a risk by uploading a plugin recommended to me by the author and downloadable only from their blog, but following an earlier comment on this blog I was moved to install the Yaab Autoblogger plugin.

So far it hasn’t stolen my identity, crashed my computer or created any other kind of unexpected havoc, so while I can’t give any guarantees, so far it looks as though I was right to trust the instinct that told me that Satheesh (the author) was not the malicious type…

Download and extraction was fine, installation was fine, and the plugin works. There are some friendly cartoons in the settings to let you know where you are, and although the language used is a slight variant from the English that I use, I could clearly follow what was going on.

I have to say that the principal inspiration for downloading this plugin to give it a spin was the hope that it might be able to cope with a wider range of RSS feeds than WP-o-Matic, which had trouble with certain feed variants.

Now, I do need to give the major caveats that I am (1) not very patient and (2) not terribly systematic, but what I found was that Yaab gave me feed URL errors with quite a wide range of feeds. The main news feed for the BBC news website – feed://newsrss.bbc.co.uk/rss/newsonline_uk_edition/front_page/rss.xml – produced an error, as did a customised feed from the BBC: feed://newsapi.bbc.co.uk/feeds/search/news+sport/glam%20rock

A particular disappointment was the apparent inability of the plugin to recognise RSS output from Yahoo Pipes, for example feed://pipes.yahoo.com/pipes/pipe.run?_id=ivS59KeM3RGF_9CarbQIDg&_render=rss&textinput1=glam+rock

I did get it to recognise the general feed from the Guardian newspaper:http://www.guardian.co.uk/rssfeed/0,,12,00.xml.

However, I’ll do a little more playing around with it and see what I can come up with! I might even dig around in the code a bit to see if I can work out what is going on in the validation stage.

Causality part II

October 19th, 2008

I left it a little longer than a week, during which time I didn’t change any of the blogs, add or edit any posts, or even check the search results.

Today, the home page of the site is back up to #9 for the search term “search experiments”, which makes the previous change look something like the usual non-random churn.

Could it be that the initial dip was some kind of penalty, but the fact that no new splog posts have been published means that the penalty has diminished? It could, but there are a hundred other possible explanations. Beware of jumping to conclusions.

Causality, splogging and speculation

October 9th, 2008

Here is a common issue for anyone involved in SEO or Google-watching, and I suppose in many other areas as well.

You take an action X. Event Y follows. You can build a plausible hypothesis for a connection between X and Y. X therefore caused Y.

The SEO version of this goes: you released new feature/code tweak/section on your website. Traffic went up the following week. New feature was a success! 

My latest version of this runs as follows: I experimented with some splogging on another blog (the relationship with this one is not disguised: they are cross-linked and hosted on the same account). I created a number of automatic posts using RSS-generated content about Google, ensuring that every time the word “search” appeared in the posts, it linked to the home page of this site. Before I did this, the home page of this site was #5 or #6 on Google, which it had been for a while. Today it is #20. So it might be very easy to jump to the conclusion that Google has spotted my nefarious tactics, and has penalised my site. 

Is that a reasonable conclusion based on the evidence?