Many new to BusinessObjects Web Intelligence ask about the ability to mashup, or merge, data that comes from multiple sources. More suprisingly, others simply ignore the feature because it's unlike anything they've seen before. But, for those users who know about it, this feature is one of the heavily relied on in any given report. So, why is data mashup in WebI so valuable?
Use Cases
There are countless applications for mashing up data between sources. Let’s simply use one measure -- “sales” -- from a Sales DB. Many dozens of questions could be answered by combining this measure with data from other source domains. How are changes to sales correlated with amount of time sales people spent in the training course (Training DB) on effective sales? Does customer satisfaction (Customer Sat DB) have a positive effect on sales? How does seniority (HR DB) impact sales? How do sales changes relate to regional economic data (public economics data)? Etc. In fact, take any measure from any business domain and there are many ways this metric could be complemented and elucidated by mashing up with shared dimensional content between different data domains.
Why Business Users Need This Control
Some in IT might answer, “Such mashing up of data is not for business users, but is a job for IT through data cleansing and ETL processes.” There is no doubt that the value of IT processes to combine and cleanse data brings huge value to the organization.
We must recognize, however, that these IT processes take time, and while waiting for the final approved combined sources, the organization suffers from lack of analytical nimbleness. Many business questions don’t need 100% accurate data to get a valuable answer. Plus, business questions need to be answered in time frames of hours, days, weeks and not in months and quarters – the minimum time frame for an IT process.
Additionally, WebI exposes a relatively simple User Interface to enable the mashing up of related dimensions. Then combining those merged dimensions and related measures is automatically supported when building a report.
The image below shows how easy it is to merge two common dimensions from the list of data objects from two different sources in the same report: Ctrl-click and right click to merge.
A concrete example:
Fictitious Company A acquires Company B. Each company had their own data sources -- let’s say, BW and non-BW. Below's a snapshot of the customer sales data. Note some customers are shared between the two lists.
Imagine if this were 1000s of customers instead of a few dozen. Not so easy to eye-ball your way to insights nor to get common analytics out of the shared values between the lists.
Before any data integration projects had even been sketched out, let alone started, a non-technical business user is able to use WebI to build reports of customers from company A and combine with customers from company B sources. Just with the above columns, s/he can answer questions as such as which of the customers from Company A, and vice versa, are not customers from the other company? Which customers are the customers from both companies A and B? Which are the top customers from A that are also customers from B and vice versa?
Below you can see the resulting table showing just the 10 shared customers, with a total for the combined revenue. After the right-click mashup of the customer dimensions, this table was created with exactly 3 drag and drops and one simple calculation.
Even in this simplified example, this combined data becomes actionable right away -- perhaps leading to the identification of some new sales opportunities targeting these combined customers.
The IT project and processes can do all the magic behind the scenes to cleanse and integrate this data. But such a multiple month IT project shouldn't prevent a business from taking action on combined content today.
NEXT: What kinds of data sources can be mashed up? (And why personal data files are key!)
