Data mining
Questions this method answers: Highly effective in business-critical applications related to creation or re-design of large or complex web sites, web applications, and preparation for selection of new CMS systems. This methodology produces insights in many dimensions by examination of data files covering structured and unstructured responses of your customers or users. This process can also profile the relative effectiveness of your various document types and interfaces by determining how well-written and formatted your existing content is, based on the needs and expectations of your primary user profiles. The methodology provides fundamental data which is critical in the creation of complex, large-scale information delivery systems or upgrades to existing systems. This methodology provides deep insight into unstructured and structured user-generated content, such as message boards, blogs, forums and fixed document types. Also, this methodology creates detailed behavior profiles and knowledge maps of an entire system's content structure.
Statistical reliability: high; large sample size. Process can scan 1000-200,000+ documents or posts, individual content in HTML documents, PDF, and all other primary document types. Methodology makes use of large log file and data base information resources to model and examine user behaviors related to interactive systems of all types.
Functional Scope: Can conduct analysis of large complex sites and data repositories with or without concern for associated user task flows.
Fees: Moderate (no respondents utilized)
Lead time: 2-8 weeks for extensive top line report
Geographic reach: Complete world reach (anywhere there is high-speed internet access and approved access to web-based information resources)
Deliverable: Comprehensive report covering content analysis of your current system with a specific focus on imbedded content resident in the data and how that data is meeting critical user needs, expectations, and satisfaction.
Major advantage: Fast, reliable, scaleable; international in scope. This methodology is used to determine the business-critical user content and interactive behavior patterns embedded in large structured and unstructured comment files, log files, and behavior tracking systems including message boards, blogs and forums. This methodology is also used to create detailed knowledge signatures of all content on large complex sites.
Disadvantages: Requires detailed access to all content in your web site or data mine and sometimes custom programming to capture data in highly dynamic web sites. We often develop proprietary executive dashboards based on data from this type of study.
Quality of service issues: Methodology is difficult to set up and monitor and should only be undertaken by a firm with deep knowledge of tools being employed. Data analysis is complex and can be misleading without also mapping to proper cognitive model of users.
Applications
- Blogs
- Message boards
- Forums
- Log files and behavior tracking systems
- Data mines and executive dashboards
- User feedback, complaint systems
- Entire website document repositories (several hundred thousand documents)
- All document types including Lotus Notes
Frequently used in combination with
- Cognitive model development
- Web site information architecture
- Functional interface design and development
- New media business strategy
