In order to improve your website’s performance you need to understand the factors which influence the chance of a sale. This area of study is called conversion mechanics. Conversion mechanics is a branch of applied statistical analysis which looks at the different types of visit and the factors which influence a person to make a purchase or not. Not all visits are alike. Different types of visit are affected differently by a common set of factors. It means different types of visitors are affected in different ways by the same influences. This article explains the different types of visit and the factors that affect their chance of conversion.

Different types of visit

There are three types of visit:

Browsing visits are made by people who are just surfing the site for ‘infotainment.’ Such visitors are more heavily influenced by elements of design than other people because any purchase they make will be an impulse purchase.
Research visits are made by people who do not intend to purchase during the current visit, they just want product information. Some will have the intention of buying later and are doing research in order to make a more informed decision. These are called focused research visitors. These visitors are typically beginning, or engaged in, a repeat visit cycle and may be actively comparing sites. They are very focused on finding product information and are almost impervious to the effect of design features. The only design feature which can really affect them is poor navigation when it makes it difficult for them to find the information they want. People looking at car insurance sites several months before their renewal is due are a typical example. They will purchase when their renewal is due, but not now. By contrast, background research visitors are simply be interested in the product area in general and are seeking to increase their overall knowledge. Hobbyists are a typical example. While they also show repeat visit patterns, the time between visits is likely to be much greater than visitors doing research in order to buy. These people are closer to browsing visitors in terms of the factors which will influence them to purchase as any purchase they make will be an impulse buy.
Focused Visits are made by people who intend to make a purchase on this visit. In most sites the majority of these people will be repeat visitors coming direct to the site. After them, the next largest group will be people who found the site via a targeted search in a search engine. The cost and nature of the product directly influences the proportions of these two. The more expensive a product is, and the greater the impact it will have on the purchaser’s life, the more likely the purchaser is to be a repeat visitor. People rarely take out loans on their first visit to a site because of the cost and length of the commitment. This person is likely to have visited the site before as a research visitor.

Most visitors start out as browsers or researchers, only becoming focused visitors on their final purchase visit. They then lapse back to browser or researcher until their next purchase, unless you loose them for good.

Conversion Mechanics

Conversion mechanics is about mathematically understanding the factors which are involved when someone converts and the relationship between those factors. These factors are the forces which influence the visitor’s decision. Some of them increase the chance of a purchase while others decrease it. There are four main factors involved in any visitor’s decision to convert:

The baseline purchase probability is the initial chance that the visitor will purchase on any given visit. This is closely connected with the type of visit. Someone who is just browsing is much less likely to purchase than someone who has come to the store with the specific intention of purchasing. Baseline purchase probability is also affected by the cost of the product and shipping costs. Dollar-for-dollar shipping costs have a greater negative impact than the product’s price. Previous purchase experiences also affect the baseline purchase probability.
The visit effect represents the impact visiting the site has on the decision to purchase. Positive visit effects, such as clear images, increase the chance of a sale; while negative visit effects, such as poor navigation, decrease the chance of a sale. Not only does each visit have its own effect, multiple visits create an accumulated effect.
The purchase effect refers to the experience the shopper had after previous purchases. Purchasing effects can have a positive or negative impact. For example, failing to deliver the product severely reduces the chance of a repeat purchase, whereas swift delivery increases the chance.
The purchase threshold represents represents the customer’s resistance to purchasing. Once the combined conversion effects listed above exceed the purchase threshold the visitor decides to purchase.

These factors can influence each other in different ways. Which factors are most important, and their relative weights, changes according to the type of visit. The impact of these effects also changes over time, so it is important to understand not just how these factors are working now, but what they have been in the past, and, most important of all, what trends indicate about the future.

Baseline Purchase Probability

The baseline purchase probability for focused visits will be high relative to the visit effect. The purchasing effect is likely to be low if negative, and high if positive. If someone has had a negative purchasing experience from your site, they are simply less likely to visit the site at all. If they do make a focused visit your site after a negative purchase experience, that experience wasn’t very important to them, and the importance of that negative experience is relatively low. By contrast research visitors will have low baseline purchase probabilities and purchase effects, but the visit effect will be high. Browsing visitors have similarly low baseline purchase probabilities, while the visit effect can vary strongly from visit to visit, and thus result in some occasional impulse buying.

Visit Effect

The most important effect is that of the visit. The visit effect directly reflects the quality of the website and is the area where you have the greatest influence. You need to know what percentage of your sales occurred on the first visit versus what percentage of sales comes from repeat visitors. Most sites find that the majority of their sales go to repeat visitors. Amazon, for example, does not break even on a new customer until the third purchase.

You also need to understand the cumulative effect of repeat visits. Does the chance of buying increase or decrease the more visits someone makes to your site? As people become familiar with the site the effect it has on them will change. For example, design features are more important in initial visits than subsequent ones, because the visitor becomes familiar with them. On the other hand, this familiarity means that elements which do not fit the familiar pattern have a greater effect in later visits than in early ones.

When it comes to the cumulative effect of visits, not all visits are equal. Some visits will have a greater or lesser effect than others. This is influenced by a number of factors, such as the nature of the visit (browsing visits have less impact than others), the time since the last visit, total visits already made, the effect of the last visit and the accumulated visit effect.

A very important, often overlooked, component of the visit effect is known as the Utility of Connection. Utility of connection looks at the relationship between the content being served and the connection this content is travelling through. Primarily this is about how long it took to render first a functional page (which has enough to use) and then the complete page. Utility of connection is actually decreasing for a large portion of the web’s users. This is because there is a general assumption in design circles that everyone has broadband. In actual fact, in the USA (most of Europe) around 40% of internet users are still on dial-up.

Purchase Effect

The chance of getting further sales out of an existing customer is most strongly influenced by how well you delivered the product to them. Problems with delivery do not automatically have a negative effect. A delivery problem can leads to a positive effect if your customer service response is above average. Sorting out a customer’s problems in a manner which shows you care and that the customer is important to you can have a more positive effect than simply delivering smoothly.

The purchase effect accumulates with multiple purchases. In other words, customers form an overall impression based on their accumulated experience of multiple purchases. The more purchases they make, the less effect each individual purchase has.

Purchase Threshold

It is important to understand the relative size of the purchase threshold and the factors of which it is composed. Companies often underestimate the resistance potential customers have. The main purpose of the site’s content is to address and overcome the factors which make up the purchasing threshold. Successful sites are created by a process of consciously identifying and addressing each element of the visitor’s purchase threshold.

The factors making up the purchase threshold are not just intellectual, but also emotional. People require information, but they also require motivation. No matter how much product information you provide them with, the actual decision to purchase is always an emotional one.

Conclusions

It is important to understand the different visitor types and ensure your site caters for each one. You need to understand how the different conversion factors are combining on your site for each of the visit types. The maths used to put these considerations together can be as simple or as complex as you like, but you need to do the work.

My thinking on this topic has been heavily influenced by “Which Visits Lead to Purchases?”, written in 2000 by two researchers in online behavior, Wendy W. Moe (http://www.smith.umd.edu/faculty/wmoe/) and Peter S. Fader (http://www.wharton.upenn.edu/faculty/faderp.html). If you want the full mathematics for such analysis, their paper (http://lcm.csa.iisc.ernet.in/scm/webmetrics.pdf) is a good place to start.