Investing in business intelligence is a costly affair. Even for larger enterprises, it is not something that is entered into lightly, and before doing so, there are a number of issues that need to be considered before any decision is made. A recent white paper from Birst Software looks at those issues and what business intelligence vendors should be offering.
According to Gartner research cited in the paper from business intelligence vendors Birst, and entitled "11 Key Questions to Ask of a BI Solution," only 20% of employees were using business intelligence as of the end of last year.
Before investing, enterprises should know what they are looking for. According to this paper, there are four basic requirements that need to be fulfilled before even starting to consider the finer points.
Four Basic Requirements
1. Historical analysis
Your business intelligence deployment should give you insight into not only performance, but also what is driving performance. Your system should be able to map data over a number of years.
2. Future projections
It should be possible to take that historical information gathered in the first step and project it into the future. This enables enterprises to align business goals according to key areas of success as well as understand why some business areas preformed less well than others.
3. Integrate information
Your deployment should be able to take information from your other business applications and integrate them so as to take meaning from across the enterprise.
4. Easily explored reports
It is pointless having reports about different areas of business if it can’t be explored horizontally and vertically. As a result, your system should have drill-down and drill-across capabilities to see information from every perspective.
11 Key BI Questions
These four capabilities are the absolute basics; if your vendor cannot provide these, then it’s probably a good time to move on. If they are present, then you have only reached the beginning of the process, as there are a lot more questions that you should be asking before proceeding. In this paper, those considerations are divided into 11 issues as follows:
1. Business views
To answer even basic business questions such as the effectiveness of marketing campaigns, your system is going to have to take information from dozens of places including third parties or partner sources.
Traditional BI solutions can do that, but often require long and difficult implementations using complex connections to data sources that bring data into one physical database. As a result your system should be:
- Able to have access to multiple data assets
- Able to analyze all kinds of data assets
- Access secured databases
- Manage the metadata of data assets.
2. Provides full features
The growing interest in business intelligence from the SMB market has seen a considerable number of solutions enter the market to cater for them price-wise.
However, companies need to ensure that they get full functionality, even if the price is lower than the larger systems. There is no point in investing in something that only gives half a product. Your solution should provide:
- A full BI solution: The license should provide everything including ETL, database management, OLAP slice-and-dice query generation, banded reporting and visual dashboards.
- Easy-to-understand pricing: Traditional solutions have many cost components including hardware, software, consultants, in-house IT support and ongoing maintenance, all of which can be difficult to track. SaaS solutions are increasingly popular as all these issues are covered by the vendor.
With traditional BI solutions, it can take as long as 18 months to see what the ROI is as a result of up front capital expenditure, IT resource requirements and development schedules. With SaaS, the time to use can be as low as a few weeks with ROI seen within 90 days. Ways to reduce the time to ROI include:
- Automation of standard processes
- Templates for typical reporting requirements
- Using existing data warehousing investments
- SaaS delivery models.
4. Data availability, security
Security and availability are two of the biggest considerations around business intelligence whether your deployment is on-premises or a SaaS. As data is coming from many systems and as your data is your competitive edge, it is essential that it is secure. As a result, your BI system should:
- Provide high availability: Considerations here include how much downtime you should expect whether it’s SaaS or on-premises
- High performance hardware: Ensure that the systems are built on hardware that will provide availability and scale of long periods of time
- Flexible security models: Needs appropriate levels of user access according to roles and needs in the companies, with the ability to restrict access to some data an essential factor.
5. IT resources
What kind of IT footprint does the system you are looking at require? Some of the older systems require almost constant IT monitoring and involvement, even to the level of creating reports or updating reports. However, there is no need for that now and potential buyers should be looking at systems that:
- Minimize IT requirements: This clearly leans in favor of SaaS deployments, but you also have to ask whether the SaaS deployment has enough IT resources behind it, in the event something goes wrong.
- Respect IT standards and expertise: It does this by meeting IT requirements for security, availability and compatibility with other systems.
- Empower the end users: It enables trained users to define reports, dashboards and alerts on their own, without any Java programming or scripting.
6. Risky integrations?
One of the problems with traditional solutions is with the large number of products and technologies that need to be put together to get a solution. An ETL (Extract, Transform and Load) solution is used to extract data and insert it into a warehouse. These are expensive and require specialized training. Again, there is no need for this currently and vendors now should be able to:
- Deliver all key integrated functionality in one solution
- Easy training with one set of commands for a single solutions syntax and data structures throughout
- Requires no custom coding as everything is integrated in a single solution
- Single vendor management with no cross-vendor diagnostics.
7. Business user functionality
Because business users are using the business insights provided by BI systems, it is essential that they are able to analyze and explore information without having to go to IT. The result is that the solution has to be not only powerful, but also easy to use:
- Easy to use: Makes it easy to explore data and new information with drill-down, and information filters for the information you require
- Quick response: Delivers quick responses even if the solution is analyzing gigabytes of information from across multiple tables and data sources
- Makes the complex easy: To make BI approachable for business users, the solution needs to manage complexity to make analysis easier to conduct
As you don’t know how your company is going to do, you need to know that your deployment can be scaled out as needs and userbase grow. SaaS solutions provide an ideal response to this; enterprises also need to ensure that the SaaS deployment:
- Can scale up and down to thousands of users without performance degradation if your company is an ambitious one
- Support multiple roles with room to differentiate between levels of users, including super users or occasional users as the case may be
- Grow without resetting, enabling enterprises to expand without having to make significant investments in infrastructure.
9. Changing business needs
Systems need to be able to rationalize data -- not just in the past, but also future data, even if that data is in new formats. It also needs to take account of changing business effectiveness. As a result, solutions should be able to:
- Add new data sources without requiring a major reset of the solution
- Be able to evaluate changes over time
- Offer business users self-service, so they can answer their own questions.
10. Powering ecosystems
To provide information to the right people at the right time, it is essential that enterprises can extend their BI system to their extended ecosystem. As a result, your solution should be able to offer:
- Multiple levels of access: Control over which type and amount of data gets seen and who can add or edit it
- Integrating data: For higher level partners it will be necessary to offer higher levels of visibility into your data
- Multiple devices: SaaS has an advantage here again as browsers can be adopted to access all information through all devices.
11. Vendor performance
As it’s just about certain that you are going to be dealing with your vendor over many years, you need a vendor that has an ongoing interest in your deployment. While the initial deployment in an on-premises version is the most important, you need to be sure that you will get the service you need after that deployment has happened.
These are only a breakdown of some of the issues that potential customers should be considering, but if you’re interested in a more detailed look at these issues, check out the white paper.