Having decided to treat blogging as a weekly habit, rather than an occasional dalliance, I've started taking a bit more notice of the page view stats available on the 'back end' of this blog, which is hosted by a company called Posterous. You can also see the often measly, occasionally huge, page views on the left of this very page.

Through this sporadic glancing at the numbers, a couple of things jumped out, so I thought I'd throw it into a Thunderfly five-minute special barchart to illustrate the issues.

Voila:

Clearly, there is an enormo-spike (technical term) in pageviews on the 4th August post, compared to those directly before and after. A strange looking anomaly on the face of it, but explained by some heavyweight web-celeb support; both Guido Fawkes and Ben Goldacre linking to my dissection of the data presentation in the Guardian's annual report.

So far, so good. Less easy to explain is the trend in the first half of the chart, prior to 27th April. Here, we have a downwards trend, but at a much higher level than subsequent posts. Indeed, if the Guardian post had been in this sequence, its pageviews wouldn't have looked out of place at all.

What's odd is that while the content of these posts is - of course - fascinating, my guess is that they have limited appeal. They certainly had limited publicity; many of the pre-April posts are links to training resources not circulated beyond a few attendees and a solitary tweet.

So was there really a drop of 80-90% in pageviews after March? As mentioned previously on Thunderfly, if you find a surprising result, the likelihood is the data is wrong. While I can explain the Guardian spike, there's no explanation through content or linkage for the inflated figures in the first half of the post. Ever the sceptic - my working assumption is that the pre-March data is in some way, incorrect. Unless someone at Posterous can tell us otherwise (there's no mention of any changes on their official blog).

The irony of this is that I have been receiving Google Analytics reports for this blog for a long time now but rarely looked at them. Why? Because I was already seeing pageview data every time I logged into Posterous. As my lizard brain followed the Principle of Least Effort, I was happy to stick with those figures, rather than open a weekly email attachment from Google, even though if I'd thought about it, I probably would have guessed the latter provide more accurate figures. *If* a change did happen in the way pageviews were calculated, I'm surprised no announcement was made to that effect, although other users have noticed the discrepancy.

A 'live' chart simlar to the one above might focus bloggers on their audiences more, enabling them to *easily* compare the relative success of their posts without delving into the Google Analytics.

This is also a cautionary tale of not taking the data at face value. Only when the methodology (apparently) changed, did it become clear that something might have been wrong with the previous data. Cross-checking with a different source - in this case, Google Analytics - would have highlighted this much earlier and provided me with a more realistic picture of this blog's audience.

1. Quality. Search engines will scour the internet for any images they can find which may be related to your enquiry. These will generally have been optimised to fit onto a web page. That is, they will be low resolution 'thumbnail' images specifically designed to be viewed on a small monitor.

When you blow that image up to fit on a Powerpoint slide, you'll find yourself with something blocky and distinctly amateurish which will detract from the information you're presenting.

2. Copyright. Google Image searches will list *any* images it can find. As such, they are likely to contain a lot of copyrighted material which you have no right to lift and use in your presentation. It only takes one email of your slides to an interested audience member and they could easily end up in the public domain, with or without your knowledge. If those slides contain copyrighted images, you could find yourself in hot water.

Even if you vow never to email your slides to anyone, there's no sense in getting in bad habits you'll just have to undo later. Learning how to use images properly gives you the skill you'll need for a publication or big conference presentation.

3. Ditch Google, get Creative. Images under Creative Commons (CC) licence are your friends. There are a variety of licences under the CC umbrella, but as long as you credit the image creator (generally, the photographer) and aren't planning to do any major re-editing of the original, you'll be OK to re-use any images with a CC licence.

Where to find them? Flickr has a huge range of images, many of which are Creative Commons. A good way to search them is through the Compfight search engine. Just click on 'creative commons' in the sidebar, enter your search term and away you go. Clicking an image takes you to its Flickr page:

Check the image licence (1, above) - it's summarised on the page but you can get more details by clicking through. This one says you are free to copy and transmit the work as long as you attribute the creator of the image, don't use it for commercial benefit and agree not to alter it.

Then move up to 'actions' and click on 'view all sizes' (2):

Then just download (4) and paste it into your Powerpoint with some text crediting the image. For this image I would suggest: Flickr user 'mobile disco': "I Love Peckham post-its" (there's no real name given for this user, if there is one, use that instead).

Ta-da! You have a high quality, correctly licensed image from the very creative Flickr community, rather than the lo-res, copyrighted fare of Google Images.

Further reading:

11 ways to use your image poorly - classic Presentation Zen post on the best way to use your image in a slide (hint: *fill* the screen).

Easy-peasy guide to crediting images - great flowchart to help you check you're doing it right.

Where can you find good images? - more Presentation Zen goodness, list of alternative sites for good, roylaty-free images.