In the previous Industry Insight in this series, a new vision for operational business intelligence (BI) was discussed. In this, the fourth in the series, the nettly issue of the costs that pertain to BI are tackled.
Tenders for BI systems range from R500 000 to R3 million. What can justify such a discrepancy for medium-sized companies?
Most of the costs in these bids are for implementation, as the software being recommended is often commoditised. In addition, companies tendering are often looking at a long implementation time (and the longer implementation takes, the higher the costs, as consultants' time is definitely not a commodity).
The cost of BI is clearly an issue, as ComputerWeekly reported last year: a worldwide survey of BI professionals found that "the perceived high costs of implementing business intelligence (BI) systems have prevented more than two-thirds of BI professionals from doing so throughout their firms, even though they overwhelmingly believe that their companies can benefit from the insights BI provides".
Simple assessment
Total cost of ownership (TCO) in BI is relatively easy to assess, once companies know what to look for:
* Hardware cost
* Software cost
* Software functionality
* Software architecture and scalability
* Time to implement
* User training
* Support and maintenance
* Data architecture
* Speed of adoption and speed of response
Each of these factors depends on some or all of the others, but it is relatively certain to say software acquisition, time to implement, user training, and speed of adoption and speed of response are the factors that will most affect TCO.
Software that costs millions to acquire will attract significant annual maintenance fees; software that is complex to use (and most of the BI software on the market now falls into this category) will hinder broad-based adoption, and limited usage means a high per-user cost; BI software that yields slow response times (most BI systems are stuck at around 10 seconds) will discourage users; and user training simply racks up the bills.
The traditional BI approach also mandates a number of tools: data warehouse, data integration or extraction, transformation and loading (ETL) tool, OLAP (online analytical processing cube), and front-end tool, such as a dashboard. Add up all the acquisition and maintenance costs and it's easy to see how BI can be unaffordable to many companies.
What skills?
Then there is the small matter of skills.
Skills are costly and scarce, a factor affecting everyone in the BI market, from tool vendors to consultants to end-user organisations.
Corey Springett is strategic business manager at Progress Software South Africa.
Skills are costly and scarce, a factor affecting everyone in the BI market, from tool vendors to consultants to end-user organisations. As a consequence, top BI consultants can and do charge a small fortune, and even entry-level skills are not cheap.
There has to be a better way, and there is. It's known as operational BI, and it offers profound features and benefits:
* Internet search-like ad hoc query across databases, data marts and cubes;
* Highly relevant search suggestions, made possible by advanced natural language linguistic query analysis;
* Report repository search and report search, allowing organisations to leverage existing reporting and BI assets;
* Personal customisable dashboard, featuring drill-down capability, scheduled queries and alerts;
* Fast and highly scalable read-only data store for large-scale deployments; and
* Vertical pre-built solutions, including vertical vocabulary, to enable quick deployment.
Operational BI goes beyond traditional BI, as it is not only meant for BI specialists, but also for executives, managers and line-of-business folk. Salespeople and customer-facing staff need information too, in a form relevant to their business tasks and activities. They need information and workflows and analysis to assist them in making sense of this information.
Operational BI must be aligned with business processes. Business performance management and business process management are moving closer together, which makes sense, as data and processes are closely coupled.
So, it really is starting to boil down to two options for customers: continue acquiring BI technology as they have for decades, or consider a move to operational BI.
* In the next Industry Insight, the focus will be on ease of use of existing BI systems, and a suggestion for an alternative approach.
* Corey Springett is strategic business manager at Progress Software South Africa.
Share