Multi-Channel Commerce & Retailing

Conversion rates are often on the forefront of the mind when operating a website.  In the simplest of models, increasing conversions can be lumped into those that increase the total number of visitors making it to the point of conversion and those that increase the conversion rate by increasing the probability that a visitor completes a transaction.  Ideally you would like to increase both simultaneously.  A useful way to look at and diagnose problems related to web conversions is through an acquisition funnel model.  This is the first post in a series that will be discussing this model in the context of a generic eCommerce site. 

The acquisition funnel model analyzes the macro behavior of visitors from the traffic drivers that brought them to the site up until to the transaction conformation page.  The premise is that at each stage in the model is associated with a probability that the visitor will leave the site or effectively not complete a transaction.  Thus the total conversion rate can be approximated by a multiplicative function whose variables are the individual probabilities of each stage in the funnel.  Although this does assume independence between the events that caused the visitor to leave the site or abandon the transaction it overall offers a reasonable approximation. 

The first stage of the acquisition funnel examines the relationship between traffic drivers and how effectively the website encourages visitors to stay on the site past the landing page.  This critical step affects not only the total volume of visitors browsing further into the website but also is a direct factor in influencing the effectiveness or cost per visitor to the site. 

An effective means of optimizing the cost per visitor and marketing spending is to link an individual traffic driver to the probability that a visitor will engage with the site.  This probability is typically referred to as the “bounce” rate on a landing page.  This allows one to correlate the effectiveness of a marketing campaign directly to the cost per click.  It also can be fashioned into a marketing effectiveness dashboard to clearly denote which campaigns are working and bring attention to those that are not.  Common reasons for high bounce rates typically include an inconsistent user experience and/or content that differs greatly from the users expectations when they clicked on the advertising link.  Tracking a specific traffic driver all the way through conversion can also be useful to understand the quality of visitor that a specific traffic driver is bringing to the site.

In my next post, I will be discussing landing page optimization in the context of the acquisition funnel model.