Nikhilesh Kumar Tiwari
MUMBAI, INDIA: Implementing a BI solution along with a data warehouse is a costly affair for any company and the cost runs in many thousands of dollars. A bulk of these expenses occur on licensing, services and hardware.
It has been observed that the old intelligence tools were unable to match up to the rapidly changing technology landscape. The current set of BI tools are no different, and offer the same run of the mill features which are more or less generic in nature.
As a result, the tools are not able to adjust to ever increasing business user’s requirement or technology landscape. Moreover, often a business user has to adjust his requirements around the features offered by the BI vendor, when ideally it should be the other way round. And that is because of these tools lack future-ready architecture.
I am listing a few instances below.
Complex licensing mechanism: Different BI tools have different and at times extremely complex licensing mechanism which is not very clear.
After purchasing the tools, there always arise more need to acquire more licenses for cores, users or developers. The development tools also come with their own inherent licensing. Incremental maintenance or renewal cost is an added burden.
But if you access or create dashboards by using a browser, the issue of separate licensing of components will not arise at all.
Unified access: With separate software for CRM, HR, finance, website analytics, social media, its becoming difficult for senior managers to keep tab of each activity.
Unified access can give them a bird’s eye view of the different URLs, user-id password. Currently, we can embed certain chart/report from BI into other segments. It can be really helpful if a BI can also have portlet kind of functionality, which could become a de-facto for accessing all the information.
BI resource crunch: The most challenging problem client’s face is in terms of skilled BI resources.
For open source BI tool this problem is much more severe. This makes me wonder why should there be a separate set of resource for BI at all. Why can’t BI tools be simple enough for a HTML/Java resource too to be able to work on the tool?
New data source type addition: Most BI tools support JDBC/JNDI database connectors or some limited number of databases. To add a new DB, an enterprise has to rely on the vendor to provide connectors. However, these days the onus of supplying these connectors has fallen in the hands of BI vendors.
Remember Hadoop faced the same issue for data storage! But this is an opportunity.
A developer friendly BI framework can allow a Java developer to write scripts and integrate those with the backend. This will definitely lessen the dependence on vendors.
Adaptability with new APIs to fetch data: Normally BI tools have connectors for commonly used APIs. But with new APIs coming into market, enterprises start scouting tools for new data acquisition.
This again is generally difficult with the current suite of BI tools. A developer friendly future ready BI framework could help in overcoming such situations.
Usage of advanced scientific and statistical charts: Normal BI tools support general tools like pie, bar, histogram, line etc.
But not all data or requirements can be fulfilled by using them. Especially the tools being used by data scientists or analytics should inherently support or be flexible enough so that advanced charts like bubble, snake, sunburst, radial can be integrated.
If the charts are not integrated, platforms like D3JS can be used which create new charts based on custom dataset.
Ways to represent data: What BI tools offer are generally reports and dashboards. But a BI framework developer can act as an artist, and with data he can create reports, dashboards, what ifs, infograph, mashup, cubes, scorecards, geographical dashboards and any other visual representing data.
BI software UI looks so very alien: BI have their own UI. They do offer some white labeling capabilities but that’s primarily on the header, footer, color, text which is very superficial and less.
A BI framework could empower the end customer to customize absolutely anything, right from colors, header, footer, text, icons, navigation, way of accessing repository, context menu, logo etc. It could go to the extent of customizing the extensions of the inherent files generated by the BI tool.
Browser could be used to for creation and consumption: All the BI tools currently use different tools—one for server, one for designing the solution, ETL, community plugins, extensions etc.
Hence, an end customer often has to download many different components with different versions of products. Interoperability issues can hinder seamless activity.
It would therefore, be ideal to use browser as a single place to create solution, edit solution, access solution. This will erase the need to download different components and worry about bandwidth, costing and compatibility issues.
Adhoc BI capabilities for business users: Adhoc capabilities allow business users to drag, drop and create their own reports and dashboards.
Many BI tools offer limited capabilities and do not allow or extend features to write custom scripts, add html, add visualization for adhoc, or custom calculated columns.
Extending the core functions of the BI tool: This is the most basic functionality but no BI tool offers. If we have a BI framework, then we will end up empowering the end-user to literally do anything.
He/she can add any kind of functionality by writing code on top of it. Ability to extend functionality and add features could change the way people view and use BI. Examples of extending functionality can include outlook plug-in, offline viewing, introducing new exporting options, rule-based system, custom alerting notifications and triggers, custom business processes. This could lead to a paradigm shift in the entire scope of BI.
An inbuilt workflow engine: An integrated workflow inside a BI tool could help in defining business processes and thus enhance capabilities.
Examples of workflow could be things like ‘run ETL AND create report AND mail to one set of users when value is between 0-50 percent, AND send it to other set of users when value is greater than 50 percent.
Have you faced any of these problems? How did you solve those? Overall, did your BI give you the desired RoI?
Nikhilesh Kumar Tiwari is all of 29 years and running his own enterprise in the Data Warehousing and Business Intelligence space called Helical IT Solutions. Opinions expressed in this write-up are his own