Do you need a data warehouse to have mobile business intelligence?

In the realm of Business Intelligence and Datawarehouseing, Jonathan's points are spot on. In regards to the mobile part of the conversation, the question has to be asked. Is mobile just another platform, or is it a complete different use case?

In actuality many pure play Mobile BI solutions are being fed by users, with extracts from a spread sheet or even exported CSV files from an existing Business Intelligence platform. This "Sneaker Net" approach will make any IT organization loose sleep at night.

Business Intelligence companies are developing their own mobile BI platforms (some for the umpteenth iteration) Many have chosen once again to attempt to shrink their current offering onto a much smaller screen and are seeing little to no interest. (How history repeats itself?) The solution works well towards IT standards but gets no adoption.

Mobile BI makes the most sense in special scenarios. Like when a sales rep is in the field and want to show real-time substantiated data to a client/prospect. This selling experience is further advanced in my opinion by the "Intimacy" of looking shoulder to shoulder at a report. (At least for those who believe that a "face to face" beats web-ex anytime)

The minimal processing capabilities of the Mobile's CPU and the size of its screen really prevent anyone from performing deep analytics. One can argue that datawarehouse or not, a simple set of extracted data in the cloud will suffice for a mobile data source.

Jonathan's "It depends on the information that the mobile BI technology will be accessing" is absolutely correct. However, I believe you must focus more on the consistency and reliabilitycharacteristics of that information rather than limitations of specific sources. A data warehouse (or dependent data marts fed from the warehouse) is designed to meet business needs for information that is consistent and reliable. Therefore, if the Mobile App has demanding requirements in this area, then the data warehouse must be the source. If characteristics such as timeliness or innovative data are required, other sources may be needed. Timeliness, for example, may demand data from an "operational data store" or even from the operational systems themselves. Innovative data may come from Hadoop environments or even from spreadsheets.

Good conversation here. Another response to this question would be that the user requirements may be satisfied with a mobile application that accesses an existing data source (even in the cloud), not necessarily a data warehouse. As we see with the growing number of smartphones and tablets, the application may not always require a data warehouse to function at the level that meets the user's requirements.

For example, SAP has an OnDemand BI tool that allows you to upload raw Excel data to a web site/cloud, and then they have an iPhone/iPad (for example) app that you can use to slice and dice the data to show some charts that help provide information on the mobile device. The OnDemand information is at http://bit.ly/focusbiondemand and the iPhone app is at http://bit.ly/FocusBusinessObjectsOnDemandApp

By using a cadence for uploading a refreshed document to the OnDemand tool, you can provide timely data to the mobile users without additional data warehouse requirements.

Try these out and see if that doesn't address the key components of the users' requests.

Ryan

Yes. You need a data warehouse for BI, no matter how the information is visualized.