Because it is so easy to do just a little more research on the internet, many purchases are not made on the first visit to a web site. It is very common for someone to find exactly what they want to purchase, but not buy it immediately. They go to a few more sites to see if they can find a better price, a better deal or maybe free shipping. When they don’t find a better deal, they come back to your site and make the purchase. This is called a latent conversion. By ignoring latent conversions you may be working with misinformation in your web analytics.

One big impact of ignoring latent conversions is tha it makes your PPC campaigns seem like they are not as effective as they really are. For example, some can click on your ad, come to your site, find what they want but not make the purchase immediately. Instead they bookmark the product page, leave your site and later come back and make the purchase. Analyzing this case it would be a direct visit that leads to the conversion. But that is not correct. That sale should be credited to the PPC ad.

Most smart analytics packages look for this and do credit the appropriate ad and keyword with the conversion. The key is that either you need to be using a javascript based analytics package or you must use a persistent cookie. Since this data is cookie based, it will not be 100% accurate but you still should pay attention to this data.

It can give you insight into information that you are missing on your product page or other potential roadblocks to conversions. For this reason it is a good idea to segment out return visitors who make a purchase and review the paths that they take through your site.

I mentioned that the latent conversion data won’t be 100% accurate. That is true of most all revenue/conversion data in any analytic packages. That is not a criticism. It is a fact. So if you plan on comparing your web report sales data to your ecommerce database there are a few things to take into consideration.

Regardless of whether your web analytics package is javascript based or log file based there are factors that impact what gets written into the log files (either web server log files or log files generated by the javascript) which cause the analytics package to miss some sales.

It is also possible for your analytic package to overcount revenue data if someone hits submit a few times. You may want to read the information on how your analytics package handles this situation if you plan on trying to compare your sales numbers in your ecommerce database to your sales numbers in your web reports.

One other factor that analytics packages don’t generally handle is returns. The purchase may happen on your web site, but the return may happen through a order management system. This will cause your web reports to show more sales than your ecommerce database.

Web analytics are great for uncovering trends, determining which PPC and other marketing programs are effective and understanding a lot about the user behavior and experience on your site. They are not a replacement for your ecommerce database when understanding the full picture of your online sales.

Filed under: Analytics on Tuesday, January 22nd, 2008 by Joy Brazelle |
You can follow any responses to this entry through the RSS 2.0 You can leave a response, or trackback from your own site.