By WorkZ Staff

Buying space online for your banner ad is often thought of as the last step in a successful online campaign. In fact, it is the beginning of a new phase. Once your banner ad is active online, it is critical that you track your ad's performance. The key information you need includes how often visitors click on your ad, and how well your main site has performed as a result of your online advertising. Two types of statistics can help you with this quest: ad-server statistics (those that tell you how often your banner ad was served), and site traffic statistics (those that show what effect, if any, your ad is having on traffic to your site). Keep in mind that statistics must be measured differently depending on the type of branding campaign you are running.

Ad-Server Statistics

The ad network (or publisher) that is running your ad can provide ad-serving statistics detailing:

• When your ad was live on the site.
• Which page(s) the ad appeared on.
• The number of visitors who clicked on it.

Additional details should be available. Different server packages have different reporting capabilities, however. Be sure to ask your publisher or ad network contact what type of information is available to you. At the same time, ask for a rundown of the reporting formats available to you. Unless you have a Webmaster or other technical person who can read logs, a straight statistical readout will be of little use. Ask if your publisher or ad network contact can provide the information in a more user-friendly format, such as a written report or even a simple spreadsheet. Also, be sure that someone will be available to you to answer questions and interpret information as necessary.

Ad-serving statistics can be imprecise, and, in the worst cases, inaccurate. This is especially problematic, and troubling, when ads fail to register legitimate "hits" (click-throughs) made by visitors. More often than not, this problem can be traced to browser caching or proxy-server caching.

Browser Caching

This occurs because many Web browsers, including Netscape's Navigator and Microsoft's Internet Explorer, store HTML pages and graphic, including your banner ad, on the user's local drive and then reload them from the hard drive rather than calling them up from the remote server. This practice, known as "caching," can cost you hits.

Proxy-server Caching

This occurs because Internet service providers (ISP) and corporations often store frequently-viewed Web pages on a (local) proxy server. This means their users don't have to go out to the remote sites and view the pages live, thus costing you potential hits to your ad.

On the flip side, it's entirely possible for false hits to be registered on your banner ad, thus artificially inflating your visitation tally. The most common culprits include:

Robots and Spiders

These software agents browse the Web continuously, cataloging each page as they go, and these visits can potentially register as a hit to your ad. That's fine if your ad is designed to appeal to computers, but it's not so great if you want your ad to be seen by humans. These agents can visit a page over and over again, vastly overstating your ad's true hit count.

Frames

Although frame sets are designed to fill a single browser screen, they are actually comprised of multiple HTML pages, each of which might be counted as a page view by the ad-serving software. The result: Even though your ad only appeared onscreen once, it might be counted as three separate page views (one for each of the pages displayed inside the frameset).

"Graphics off" Browsers

Surveys show that a portion of the Web surfing public turn graphics loading off while surfing the Net. In this case, the page containing your ad is delivered to the browser, but your actual ad is not.

Some argue that these understatements (and overstatements) offset each other. In other words, for every 100 free ad hits you get due to caching, you also (might) lose 100 hits due to bots, spiders, and visitors who didn't load graphics. The very nature of this milieu precludes ever coming up with a definitive answer to this question, so all you can do as an advertiser is stay aware of these issues, and find publishers and ad networks that can address them to your satisfaction.

Site Traffic Statistics

As surfers visit your site, your server can log basic data about their visit. By setting up appropriate log files and then using traffic-analysis software, your ISP or Web site administrator can help you analyze the success of your ad campaign. Much of this information will match the data you receive from the ad network or publisher that ran your ad, but your site statistics can reveal new information and patterns. They can also help you identify glaring discrepancies between the data you get from ad campaign reports and the data you get from your server logs.

These discrepancies are not necessarily due to mischief or incompetence on the part of the ad network or publisher. Take unfinished click-throughs, for example. These occur when someone clicks on your ad but doesn't actually follow the link and end up at your site. This can be due to a slow network connection, impatience, a phone call, or any number of other reasons. The click is measured (accurately) by the ad server, but your site is never visited.

Anecdotal evidence suggests that the difference between ad-serving and site-traffic statistics can range from 10 percent to more than 100 percent. Obviously, the larger the discrepancy, the greater the need to revisit your arrangement with your ad network or publisher.

Both ad-serving and site-traffic analyses can help you sort out, in a very detailed fashion, how successful you were in meeting the various targeting and placement goals you established at the beginning of your campaign.

Three Ad Campaigns Using The Stats

Let's look briefly at how you can apply these analyses to each of the three types of campaign plans.

Branding Campaigns

Your goal in a branding campaign is to get the number and type of  impressions of your ad for which you paid. In this case you are, for the most part, at the mercy of the ad network or publisher's ad-serving software to tell you how many times the ad was delivered. However, if you are also tracking click-throughs (and you should be in all your campaigns), then you could correlate your server's click-through data with the report from the ad network or publisher and extrapolate from that whether all your ad views were delivered. For example, if there is a 10 percent difference in click-throughs, then it might be reasonable to infer a similar difference in ad delivery.

A major problem with branding campaigns involves terminology. Some define an ad view as a request to a server for your ad, an ad request as an instance of a page with your ad being requested, and an impression as an instance of your ad actually being displayed. These terms get used interchangeably, and there is no universal agreement on which one to use when, so make sure you understand what an ad network or publisher means when they use these terms.

Traffic-building Campaigns

Your goal in a traffic-building campaign is to get as many surfers as possible to click on your ad and visit your site. Most ad-serving software uses click-throughs as a principal tracking mechanism. Since your server logs can detect every visit from a particular banner ad, raw click-through rates are fairly easy to confirm.

Click-through Campaigns

The process of converting a visitor to a customer almost always happens within your site, after the surfer has already followed the ad link, so the process of correlating click-throughs and subsequent sales is an internal one that you need to set up with your ISP or Web site administrator. Generally, this involves using cookies, member log-ins, or other mechanisms to track a visitor's movement once he or she has entered your site.

---

About the Author:
The WorkZ staff is made up of gurus in many areas of expertise including Sales, Marketing, all aspects of the Internet, Technology, even starting and running businesses.

Click here for more info from this author...