David Housman

Ass Kickin’, Finger Lickin’ Data Collection Design

0 Comments

There are many ways to skin a cat, and multiple options to implement any data collection request in Omniture. When we‘re getting started, we ask “How do I collect this data?” When we mature, we ask “Which is the best way to collect this data” Our job is to know the different options are, then make decisions about which is appropriate. Design is dictated by requirements gathering.

Report design and solution design should be linked. The analyst should design data collection with a clear idea of what the report will look like, who will access it, how it will be used, etc. The resulting report and the different types of activity it supports is an important criterion for evaluating a design.

Our solutions can’t optimize all of these factors. There are always trade offs. The analyst must weigh these trade offs in his decision. Good solutions have considered the following factors:

Effort The time to implement different data collection features is an important consideration. Frequently we are faced with a choice between engineering and maintenance/reporting effort. Good solution design has considered the different options and costs and makes tradeoffs in the best interest of the organization.
Patterns Good designs solve the issue at hand, as well as similar or recurring issues. Where possible, the design can be used in other similar situations.
Sustainability The solution anticipates what will happen in the future, and can accommodate the site as it grows or changes.
Maintenance Humans are fallible. As the site grows or changes in predictable ways, it is preferable to avoid relying on manual effort to keep the solution in shape.
Reliability It is the responsibility of the analyst to protect stakeholders from inaccurate or misleading data. Any design should collect and deliver meaningful data.
Consistency Where possible, metrics and dimensions should be collected in similar fashion to support comparison across reports, systems, products, divisions, etc.
Tools The solution considers the different data collection tools and configurations available in the Web Analytics package. It also reflects a firm knowledge of JavaScript, and where necessary, taps the knowledge of engineers. Where appropriate, the solution incorporates research into the Web Analytics tool, or ways in which a common problem has been solved elsewhere.
Redundancy There may be other systems that do the same thing as this. Will this design add unique value to the data being collected that other systems do not? Is this system the appropriate vehicle for collecting this data?
The X Factor There may be other issues related to context of the application, the needs of the stakeholder, etc. that aren’t listed here. Good solutions fit the scenario. Sometimes it is necessary to think outside the box.

Leave a Reply