It is good business practice to keep tight control of spending and to minimise costs, but traditional business intelligence (BI) systems make it very hard to follow this rule, writes Roger Llewellyn, CEO of Kognitio.
To date, the major cost of implementing BI solutions has been the large up-front expense of buying the requisite technology and expertise. Successful BI is underpinned by a data warehouse that can store, recover and analyse the data needed.
If a project is successful, users will see more value and analyse more data. Therefore more capability is needed, more users are added, more analysis is carried out and the cycle continues, with the project always expanding thanks to its own success.
The up-front costs discourage many organisations from pursuing BI, and knowing that the project will demand more and more funding does not improve matters. In a recent Vanson Bourne survey, a quarter of respondents admitted to spending over £1m on BI implementations, with the majority reporting slow and unsatisfactory results.
Testing the water with SaaS
BI should be one of the easiest IT costs to justify, as it can return dividends on any investment. To overcome firms' wariness, IT departments need to "try before they buy". This offers a way of ensuring that a BI project can earn its keep while avoiding the large up-front and ongoing costs involved.
The best way to do this is to employ BI, and the data warehouse, as a software as a service (SaaS) model. This avoids the high initial costs while ongoing costs will also be strictly defined by the SaaS contract.
This means the project can be budgeted accurately and simply, the solution will be faster to implement, and it will scale up easily.
Once a BI project has proved its value, organisations will often want to take it in-house, both to reduce operating costs and to ensure the organisation has full control over all of the technology and data involved.
By starting with a SaaS model, organisations will know how to run and use a BI solution and, for budget and charging purposes, will know exactly which departments will make most use of the solution. As a result, the main expense in taking BI in-house will be to secure the technology, especially for the data warehouse.
The best way to cut costs here is to use BI and data warehousing solutions that run on standard, commodity hardware. Organisations can use any excess initial capacity for processing other jobs, and then shop around for more hardware as needed without being restricted to a single vendor's products.