Tuesday, March 17, 2009

Are Mashups the Killer BI Application?

Do you remember mashups? Mashups are really composite applications, mixing and matching services, typically Web-delivered, to solve any number of business problems, including operational BI.

Mashups typically have to following characteristics:
  • They leverage two or more data sources which are combined within a new application or view of data.

  • They are built quickly around a particular business need.

  • The resulting information view is more valuable to the than the information that's mashed up.

  • They typically leverage information available over the Internet through APIs (e.g., Google Maps API).
For example, a friend had a client who was delivering beverages in the Bay area, and was troubled by delivery delays, typically associated with traffic issues. An innovative IT guy was able to create a mashup that combined both the delivery schedule for the day, and Google traffic. Thus, they could avoid traffic trouble. The greater delivery efficiency saved the company over a million dollars a year when considering saved fuel costs, labor costs, and also increased customer satisfaction.

Better, more BI-oriented applications included another client who was able to monitor sales patterns as related to customer demographics using a mashup. Thus, sales could be targeted at the specific demographics they were addressing, and they learned so much by using this mashups that they were able to increase sales by 50 percent in just a single year.

What's unique about mashups is that they have a tendency to address a tactical business need, and can do so quickly. Typically, mashups take only days or weeks to create, versus months for more traditional BI applications. Moreover, they have a tendency to leverage Web-based API, and thus you get access to valuable information you neither own nor host, and in some cases it's free.

The moral of this story is: Don't forget about mashups in the context of your BI strategy.


Post a Comment