net.work

The Way Business Is Moving

net.work published by
Issue Date: October 2005

DIY BI takes its first steps

1 October 2005

Nobody can deny that skilled and experienced experts are required to implement an effective business intelligence system. For some companies, however, simplicity is key for any information front-end.
Business Intelligence (BI) is one of the areas in IT that is continually growing. As more people in various jobs need information to be productive and deliver the results required of them, the greater the need for reliable data to assist them.
This is where BI begins to fall apart. There is plenty of information available from many sources, but unless users are experts able to extract what they need, getting the data required timeously is a never-ending battle - mostly lost.
True BI should deliver realtime information at the click of a button. Users do not want to bother with the details of where it comes from, how it got there and if it is reliable. They simply want to select their data, examine the outcome, drill down if required and assume the results are fine and get to work.
Bryn Davies, practice manager, Cape Town at the BI Practice notes that, from a tools and technology point of view, easy-to-use (or DIY) BI is most certainly possible, but two issues hinder adoption:
a. Problems in implementation and execution.
b. Human issues relating to how humans make decisions.
Bryn Davies, practice manager, Cape Town at the BI Practice
Bryn Davies, practice manager, Cape Town at the BI Practice
Problems in implementation and execution are due, in part to the selection of the incorrect presentation tool - usually selected or recommended by an IT person not fully acquainted with the needs of users. In addition, user training is often not sufficient to create independent users as it does not include training with the actual data sets that will be used.
Davies reports there is still a disconnect between end-user requirements and IT understanding. This problem is only exacerbated when the integrity of the data being queried is in question (due to deficient collating and cleansing of the data; and when companies have multiple sources of the same data, each with its own collation and cleansing practices, the same analysis on each data set may produce different reports.
Rian Durandt, head of Informatics at 3fifteen, the Microsoft applications solutions division in the Dimension Data SA Group, is a great proponent of DIY BI. Instead of coming out in support of a specific front-end tool or data warehousing strategy, Durandt suggests companies first make use of what they have before spending fortunes on new BI licences.
Rian Durandt,  head of Informatics at 3fifteen
Rian Durandt, head of Informatics at 3fifteen
Specifically, he mentions that if companies had to take advantage of everything they get in their Microsoft solutions - such as Microsoft Office - they would be able to go a long way in providing relevant information to the right people in an easy to use and understandable format. "Why must users be taught to use a new front-end tool when they are already quite capable of using Internet Explorer, or Excel, for example," he asks.
Of course the back-end functionality and the expertise required in creating the data sets specific to the user requirement still need experts to design and implement them, but this is a specific task to be done at the outset and perhaps at set intervals as the customer's business grows. The day-to-day information provided to users should be taken from these back ends as required and prompted by the user.
"Using the tools users are used to will give them more confidence and increase their productivity," Durandt notes. "Ensuring the front ends are simple enough for anyone to use - such as creating informative web parts in the Decision Support Panel. And information can be broken down further and presented in Excel with only a few mouse clicks."
Davies adds another factor to the argument. He says that no matter what technology is available, it also always comes down to the human element. Executives today tend to be more interested in simply being given information and using it in decisions rather than actually going through the process of extracting the data they need and applying it.
"The desire to be empowered must be there," he notes. "The majority of people just want to know about exceptions and/or want to spend their time and energy on making the decisions, not fiddling about with tools and numbers. These people will always ask someone else to get them the data or report."
Do you want to DIY?
Andrew Connold, managing director of Synergy Computing, expands further: "The interesting thing about DIY is that in many instances you can do it yourself, but would the results be as good as getting the professionals in - and would you be prepared to spend precious time on the task at hand?"
Andrew Connold, managing director of Synergy Computing
Andrew Connold, managing director of Synergy Computing
While some aspects of the technology are commoditised - and hence easy to set up and use - others remain fairly complex and require the services of professionals. BI falls into the latter category, as it can be a complex solution that requires a level of expertise and experience for reliable results.
"For the larger corporate, we have seen that most companies engage with a service provider for the initial set-up of their BI system, and then manage and maintain the solution in-house. This is a model that can also be applied to smaller organisations. However, the complexity of the IT environment, the available skills within the company and the complexity of the outputs from the required BI solution will determine if it is feasible and practical for an individual in the company to implement BI tools," says Connold.
Do not write off the concept of DIY by any means, he adds - but certainly consider the time and cost benefits of getting the job done properly and timeously.
The facility for DIY BI is naturally available to anyone requiring it, but one can ever get away from the issues of back-end design and implementation in support of the front end. Unless an organisation has only one data store, the data extraction, transformation and loading processes are complex and require expertise and expensive tools. (Of course, Durandt notes that Microsoft SL Server has many BI capabilities delivered with the primary database system).
Many companies provide the full suite of tools, from the front to the back end, with the focus on integrating everything. And these vendors would tell you extracting and using information is simple - and for someone trained in the particular toolset, it is. With systems as described above by Durandt becoming more common as well as more needed in the enterprise, it seems as if many companies will be opting for using the tool they know in order to ensure a large portion of the required intelligence gets to people who need it and will use it - mainly because of it being available in an interface they are familiar with.
On the whole, however, it is still too early to even think that BI is a DIY technology. Even with the complexities involved in preparing the data put aside; and even with the capabilities in Microsoft Office, for example, making it easier to manipulate and view a large percentage of the data workers need; the ability to get information and drill down for long term decision making and intricate analysis is still a job for specific tools. This state of affairs may not last for long, but the traditional BI tool vendors do not have too much to worry about yet.


Others who read this also read these articles

Search Site





Subscribe

Previous Issues