Hello. My name is Tobias Schwarz (Noki on IRC). I’ve been using Openads for a very long time and I earn my living through online advertising. I’m friends with the Openads team and have been helping out with testing during the Openads 2.3 beta process.
I was pleased to be invited to guest blog on the new Openads blog. I’d like to share my thoughts on how you can optimize your website to generate higher revenue. I hope that Openads publishers will find this useful.
By optimize I mean improving the design of your website so that more of your visitors do what you want them to do. Today, I’m going to focus on the basic tools and techniques for measuring and experimenting with your website design.
This article does not discuss techniques for driving more traffic to your site. We might talk about that another day.
So, go get a cup of coffee and start reading!
1. Decide on the goal
First you need to decide what you are trying to achieve. What do you want each visitor to do? Your goal might be to increase the number of pages each visitor sees, the time they spend on the website or your conversion rate.
Conversion rates relate to a user action beyond just visiting your website, for example:
- Bloggers might want more visitors to click on advertising
- Amazon wants more visitors to buy something.
- Openads wants more visitors to download their adserver.
2. Get statistics
The next thing you have to know is that it’s all about statistics. You need statistics about your visitors and your website; this information will guide you in optimizing your site.
Depending on your goals you will be focusing on different statistics. Some key ones might include:
- Ad click-through rate
- Time spent on site
- Pages viewed by each user
- Movement of visitor around the website
One very powerful and free tool is Google Analytics. If you don’t already use it, or something similar like phpMyVisites, then install one now!
Openads reports on banner advertising click through rates which is perfect for people who want to optimize their website so that more visitors click on advertising.
By tracking several thousand visitors you will get “stable” statistics to use as a baseline for your experiments.
3. Start experimenting
Now that you know what you want to achieve and have statistics to measure how well you’re doing at the moment you can start experimenting.
A commonly used method for experimenting is a/b-testing. For a/b-testing you set up two different versions of your website and compare their performance.
How you do this will depend on what website software you use. Ideally, you would assign one of two templates to each new visitor and test which one works best. If your software isn’t flexible enough to do that you could use different templates each day then compare their performance.
Over time a/b-testing will help you find an optimized version of your site.
4. Optimize
There are a large number of things that could be optimized. I want to point out some common ones where a little effort can make a big difference:
- As you might already know layout is a factor. It does matter where you place content, ads and navigation elements. It’s also important to keep colours and contrast, fonts and font-sizes in mind. Make sure to use clear structures for all navigation elements. Set up different layouts and start to track the differences.
- Connection speed is a very important revenue factor. Sites that load faster have higher page impressions per visitor and also higher conversion rates. On an unoptimized site people with a faster connection produce about 50% more page impressions than people with dial up connections. On an optimized site (e.g. that delivers gzip-compressed contents) users with dial up connections produce slightly more page impressions. The age of users with different connection speeds is an important factor for this.
- Screen resolutions are also a revenue factor. If a site scales to 100% width, lines could get very long. For higher resolutions, page impressions usually decrease because content is hard to read. Work with fixed widths to take care of this problem. Use javascript to read resolutions and display additional banners on the right for higher resolutions.
5. A specific example for layout
I’m using these techniques myself at the moment. Alinki.com is my latest project and it’s a good example to look at. It’s a web-directory with lots of articles and it was designed having all the things I mentioned in mind (and a lot more). It’s a pity that there is only a German version online right now but since our discussion is largely about layout it won’t affect your understanding.
Check out this screenshot of one of its subpages.
I added several blocks and lines to it to point things out. When I designed the site I made sure important elements were visible without scrolling for almost all users. Most visitors have resolutions higher or equal to 1024×768 (95%). Everything above the green line is visible at 1024×768. As you might have noticed there’s a searchbox, breadcrumb navigation, links to register, login, help and contact (marked purple) and two boxes with google adsense (marked red). Depending on how many toolbars the user has installed, a small part of the content block could also be seen. The sidebar isn’t filled yet but it will contain teaser-texts.
Below the green line you can see Subcategories and an article with another adblock inside and some empty red boxes. Those empty boxes are locations that might be good to hold more ads. I say “might” because we haven’t tested those positions yet, but we will once we have more visitors, more content and the sidebar filled.
Less important content like imprint, another contact link and terms of service were placed at the bottom of the site (marked blue).
Of course, I’m using Openads to display the advertising on this site. When I experiment with the layout of Alinki I will use the Openads click through rate reports so that I can optimize the revenue I make from my site.

Tags:
Thanks for the tips. I always wanted to test out openads and see what happens.
Comment by Webanalyticsbook — July 18, 2007 @ 11:24 pm
Just to note - when Tobias says that he is Noki on IRC, he’s talking about how you can chat with Openads users on the IRC channel #openads on the server irc.freenode.net - as described on the homepage on http://developer.openads.org
Comment by Arlen Coupland — July 19, 2007 @ 11:27 am
Also I wanted to plug the useful ‘heatmap’ feature that some stat programs have. It lets you quickly view what links on your website receive the most clicks by displaying a screenshot of your website and a heatmap (similar to what a weather report looks like)
phpMyVisites just added this feature as a beta, quite nice!
Comment by Arlen Coupland — July 19, 2007 @ 2:12 pm
Heatmaps are especially useful if you have more than one link to the same target on a page. Using a heatmap you can track which one get’s more attention. Google Analytics has something similar called “Website-Overlay” but it only tracks targets, not the positions.
Comment by Tobias Schwarz — July 19, 2007 @ 2:31 pm
As arlen just said there is new heatmap feature in phpmyvisites : you can try the demo here
Comment by Matthieu — July 19, 2007 @ 3:08 pm
Does someone know if it is possible to add the google analytics goal tracking script in the page called when someone click on an add (ck.php)?
I would prefer follow the statistics from google analytics, it could be the ultimate combination!!!
Comment by Nicolas — September 2, 2007 @ 11:15 pm
Thank you for the tipps and tricks, in my opionion the knowledge is just the first step to succes. My problem is the process of doing it always right
Comment by Arno in Room 2012 — November 12, 2007 @ 10:30 pm
Server optimization and tuning also helps a lot. You may also offload your image serving to a different server or even a content delivery network - will lower your web server load. If you are hosting a lot of images, you should consider running your host on a server or host with 100Mbps full duplex uplink (if you hit 10Mbps in peak times often).
Comment by CDN — January 2, 2008 @ 11:36 pm
You should check out http://www.goingup.com as a (MUCH BETTER) alternative to Google Analytics or phpmyvisites..
It’s a free service, which offers the best of all the ones you mention, and offers also benchmarking, etc..
Check it out…
Nate
Comment by Web Stats, SEO, Analytics — September 26, 2008 @ 5:32 pm