In order to better understand in text advertising, some technological background is required. I’m not going to dive into the code structure here; in fact, I’m not the right person to attempt that… but I do think that some basic clarification could help when considering this method. Also, the leading in text ads networks utilize different technology and the following analysis may be helpful in choosing the right partner for you.
When you integrate in text ads on your website, all you’re required to do is insert a little piece of JavaScript code. This invisible code, when run by a browser, calls upon a bigger script to run on your website and identify your website to the in text ads provider.
The first job for this code is to analyze the content of the page and determine if it’s suitable for in text ads. Sometimes, if there’s not enough text or if the text is not suitable for advertising, the process will stop right there and no ads will be placed. If the page is suitable, the code will then analyze the page’s content and identify the terms which get the highest score balancing the contextual relevancy and the potential for advertising. The top terms will be highlighted as double underline links and will serve as hooks for advertisements.
As explained before, if a user is interested to learn more about a highlighted hook, he can hover over it with the mouse and a bubble would appear with a relevant ad inside. The code’s brain here is asked to match the ad to the reader and to the term in order to maximize the chances for a click. If a click was made, the code now needs to send the reader to the advertiser’s landing page and monitor behavior in order to charge the advertiser and credit the website publisher for the ad’s current price.
All this magic begins with that little piece of JavaScript code, but of course it requires a robust and smart system behind it. This system has several key differentiators among the leading in text ads providers.
When the JavaScript is loaded and the bigger code is called into action, the browser has some work to do. Therefore, the timing of the load is important. If this work is done in parallel with the loading of the page, it may slow down the loading of the actual content of the page. Such slowdown may result in visitors getting frustrated and leaving before the page completes its loading. The natural conclusion is that it’s important to make sure that the script that you’re using only runs after your page has completed to load the main content.
These days, with the numerous types of widgets and other scripts that run on a page, there’s another timing factor to consider – the timing of the in-text script among other scripts. The best example would be the common Google Analytics script, which runs in order to count the visitors and monitor their behavior on your website. Since the Google Analytics code may take some time as well, if you place the in-text script after it, the ads may appear a little later. The same goes for other widgets. It’s up to you as a publisher to determine which code should run first. My recommendation is to place the in-text JavaScript just after the actual page’s content, but before other scripts that are not money oriented.
The most challenging aspect of in text advertising is matching the best ad to the visitor and the hook. This process of ad placement requires analyzing many factors, including the context of the page, the content category of the website, the known characteristics of the user, the highlighted term within the hook, and the available advertising campaigns at that specific moment. Since the Web is very dynamic, some of these factors may change every minute or even within a given minute. Most important, available ads and their prices are mostly handled through open markets and they change all the time.
Some in text ads providers do not use dynamic ad placement. I won’t point them out, but it’s quite easy to check if the selected hooks and the shown ads change when you hit refresh. Ad networks that have a relatively old technology may find it difficult to dynamically match the ads, and they actually place the ads beforehand, without maximizing the revenue potential. In fact, some of them even choose the hooks for highlighting manually, and therefore would not deal well with dynamic content found in blogs, forums and news websites.
From my own experience with Infolinks, we’ve put atomization as a top priority. The smarter the algorithm and the more automated the in text ads placing process is, the higher is the effective CPM for the website publisher. There are two reasons for that. First, automated processes are more sensitive to changes in the market and to dynamic content; they simply get better results. Second, automated processes allowed us to manage the system with minimal staff, focusing our human resources on customer service. When the highlighted hooks require manual content analysis, the organizational structure would have higher costs that will limit the organization’s ability to share most of the revenues with the publisher.
Again, this is from my own personal experience, and I’m sure other networks are striving to do the same. When we set up our system to require minimal manual intervention, we’ve actually saved costs and enabled our organization to offer higher revenue share to our publishers. This is a good example for technology differences that have an immediate influence on the publishers.
Another example for technology differences is hidden within the IT infrastructure. Anyone who’s worked for a Web related company knows that sometimes the IT costs – that is, the prices of computers, servers, and network – climb so high, that they become a major factor in the company’s finances. Placing in text ads dynamically requires real-time analysis of content and real-time matching of ads in rapidly changing market conditions. If the system is not built taking IT infrastructure into consideration, then the IT costs can make it impossible for the company to share most of the revenues with its publishers.
While in text ads may appear to work as magic with a little piece of code, it turns out that there’s a lot of work behind it. Also, while most in text ads may appear to work similarly, there are major differences among in text ads network and understanding them may help you choose the right partner for you. If you’re looking for deeper insights into the technological process, drop me a line and I will try to do get you an answer from someone from our own tech team; and if you need more help in understanding the business implications of these technological differences, well, I might even try answering you myself.