Data Mining
Log file analysis programs, although valuable for visitor behavioural analysis, can only provide relatively simple analyses of buyer behaviour. In contrast, data mining can provide very sophisticated customer research analysis by finding hidden patterns in databases.
Data mining requires the establishment and recording of visitor and item characteristics, and visitor interactions. These statistics accumulate when visitors interact with items, your web site, and your business.
Visitor characteristics include:
- demographics - attributes such as home address, age, income, home ownership, computer ownership etc.
- personality attributes such as highly protective feelings toward children, strong interest in local issues, impulse-buying tendencies, early adopters, keen interest in sports and recreation etc.
- attributes of the visitor's system, such as operating system, browser, domain, and modem speed
Visitor-item interactions include:
- history of purchases
- response to advertising, history
- preference information
- feedback
Data mining works best with clear, measurable goals such as:
- increasing average page views per session
- increasing average profit per checkout
- decreasing products returned
- increasing the number of referred customers
- increasing brand awareness
- increasing visitor and customer retention rate (such as number of visitors that have returned within 30 days)
- increasing conversion rate of visitors into customers
- increasing the average number of items sold to each buyer.
Data mining provides sophisticated intelligence and analysis of visitor and customer behavior but at a price. There are few "off the shelf" solutions and those that are available are expensive. On the other hand, if data mining enables a significant continuing increase in sales, then the cost will be more than offset by the benefits.
The Internet Marketing Engine will assist you to evaluate the costs and benefits of data mining and customer profiling and select the solution best suited to your requirements.
