The choices available to companies looking to add business intelligence, reporting and analysis to their operations can be overwhelming. There are literally hundreds of product offerings to choose from: a Google search for “business intelligence software” will yield more than 290,000 responses! Further, most of the products in the Business Intelligence field are technically similar. Throw in technical jargon and marketingspeak and making relevant differentiations becomes quite difficult.
To cut through the clutter and streamline their decision-making process, most companies build a short list of preferred vendors based on some set of criteria, and then invite those short-listed companies to compete for their business. Often this involves an elaborate RFI process, with extensive spreadsheet-based checklists where different products are compared across potentially hundreds of attributes and factors. These processes tend to favor larger vendors – those with product scope necessary to achieve a check in the “yes” box for the most features. This type of process also tends to favor complex products – as those products are most likely to have the most features.
Needless to say, a long list of checked boxes contributes little toward understanding the benefits and complications created by using a particular product. The typical RFI process often produces an extremely homogenous short list – “winners” that all clump together with look-alike features and attributes.
So how to choose? Companies can both simplify and greatly improve the quality of their short-list selection process by taking a different approach – rather than creating a master list of all possible features and then checking off feature after feature in an attempt to “keep score,” companies should focus on the business value and ongoing expenses that different solutions create – measured across a few key variables.
This article is intended to outline key criteria that companies should consider when creating their shortlists – and making their ultimate selections.
Speed of Implementation for the Whole Solution – Analytics Applications To the Desktop
Just as product marketers talk about delivering the “whole product” rather than just a piece of the overall technology, companies should look at the time required to implement a complete working solution – one that makes it possible for business users to answer business questions. There are two measures to watch here: Time until the first application is deployed to an end-user, and total deployment time for the entire project.
According to a survey by DM Review and IDC, only 5% of companies were able to deploy a usable, working analytics application within 1 month – and the average time to first application deployment was 5 months. More than a quarter of the companies surveyed took more than 6 months to deploy a single working application, and the average time to completion for projects was 17 months – just shy of 1 ½ years.
Many of the BI solutions available today are “infrastructure heavy” with extensive pre-application deployment work paving the way to the actual end-user products. Companies should be aware of the timelines that their BI vendors require when they make their selections, and how those timelines match up with their expectations and requirements for the project. During vendor evaluation, companies should be sure to include at least one vendor with a track record of delivering rapid deployments (3 months or less, start to finish) so that they can weigh the benefits and tradeoffs in their evaluation process.
Key Metrics:
1. How long does it take to deploy the first analytics application?
2. What is the average or typical total project deployment time for this product?
Flexibility – Are We Building With Cement or Something a Little More Pliable?
There’s an old saw about ERP software: “It’s just like cement – you can pour it to look like anything you want, but once it dries, you’re stuck with it.”
Many of today’s business intelligence solutions involve extensive infrastructure projects – building rigid data warehouses and small armies of OLAP hyper-cubes over a period of several months – with actual applications reaching the end-users only at the very end of the project.
While these implementations can provide real business value, their success depends on a clear understanding of the business drivers and requirements at the beginning of the project. Unfortunately, analytical design tends to be an iterative process. For companies where the business changes rapidly or where no clear analysis program has been in place before, the ability to modify analytics applications quickly after they have been deployed will have a significant impact on the project’s success. The same survey mentioned above found that 75% of all analytics applications needed to be modified within 6 months of their initial deployment. Companies should look at their business intelligence system as a living, changing set of applications – and consider the time and cost required for making changes to existing applications as a critical element of their evaluation.
Key Metrics:
1. After an application is deployed, how long does it take to add a completely new data source to the application?
2. Once a new analysis area is identified, how long does it take to prototype and deploy a completely new finished production application?
Ease of Use – Can Our Employees Learn to Use It?
Your company can build the most powerful, sophisticated business intelligence system in the world. But if no one in the company uses it because it is too complicated to figure out, the system will be worthless. Ease-of-use takes two forms – for the IT department that maintains and modifies BI applications, and for the end-users who actually use the applications. Both have a significant impact on the long-term success of your business intelligence initiative. Despite this, ease-of-use is often overlooked during the business intelligence selection process. Perhaps because it is an attribute that is difficult to measure (check yes or no – does this product have “ease-of-use” in it?) the standard checklist RFI process rarely even attempts to address this issue.
Companies need to consider what it will take for their employees to use the systems they deploy.
Key Metrics:
1. How long would it take the typical line of business manager to learn to use this product? Measure in hours, days, or weeks.
2. How many days of training do other customers using this product recommend for end-users? For IT staff?
3. How long would it take a power user from our company – either IT or business analyst – to learn to develop new applications with this tool?
By addressing these three factors – speed of implementation, flexibility, and ease of use – during the RFI process, companies can significantly improve the quality of their business analytics and reporting product selections.
