Until recently, the only organisations able to access the benefits of Business Intelligence would be large ones with the budget to take on the enterprise-scale offerings of BI software giants such as IBM (Cognos), Oracle, SAS and Business Objects. These client organisations almost always have an in-house BI department, large quantities of data, diverse computer systems and interesting technical challenges to boot.

But lately things have started to change. Low-cost and easy-to-access options are now finally starting to appear. The cost-saving and competitive potential of BI are now accessible to organisations of all sizes. In this article I examine some of these ideas - and the benefits deploying them might bring to your organisation.

Idea 1 – Use Microsoft SQL Server
SQL Server now comes with a suite of Business Intelligence tools which allow you to both build and interrogate your Business Intelligence system. SQL Server Integration Services (SSIS) is an ETL (extract, transformation, load) tool – in other words one that allows you to collect and organise data in a data warehouse; SQL Server Analysis Services contains tools to analyse and report on your data. Although earlier versions of SQL Server included some of this functionality it has become much more user-friendly since the release of SQL Server 2005.

One big advantage of SQL Server is its price. At the time of writing, the standard edition weighs in at around $US6000, which includes both the database itself and the Business Intelligence functionality. If you are planning to use the database anyway, you could make big savings by utilising the other tools bundled in with it. However, remember that you will need a good understanding of Business Intelligence principles to do the development in-house.

Idea 2 - DIY Data Warehousing
This means writing the data warehouse yourself, straight out of the box, as it were. This means no user-friendly “wizards” you get with the likes of Microsoft SQL Server. To develop this way you need a high level of skill and expertise. The advantage is that you have complete control over how the data warehouse is written, without having to purchase specialist applications. This level of control also means you can fine-tune performance yourself and achieve significant gains in data throughput.

On the other hand, a data warehouse built this way can be difficult to maintain. Development may or may not be quicker than other methods covered in this article, depending on the experience of the developer – and could well be slower and so more costly. Standards are less enforceable, adding to potential maintenance costs.

I should emphasise at this point that using less expensive tools does not excuse the need for a good strategy (see my article Seven Common BI Strategy Mistakes posted on the Acestar Solutions web site). Strategic mistakes can quickly cancel out the benefits of low-cost access to Business Intelligence.

Idea 3 – Outsource your development
Outsourcing has been around for a while. Outsourcing development, in the context of cost-saving, usually means indirectly hiring skilled resources from places where wages are lower. These locations are typcially overseas, but they can also be in other parts of the same country.

Outsourced staff can either work on-site or remotely, depending on the needs of the project and other circumstances.

My own experience with outsourcing, as well as anecdotal evidence, indicates that it can work well but needs to be managed effectively. It is important to ensure that developers have the right level of experience, and specifications in particular need to be very clear and specific (as opposed to the scrawled notes on a Starbucks serviette I myself have often worked from!).

Idea 4 - Open Source
ETL applications integrate data, often from diverse sources, into a single place, so the most common use for ETL software is to build data warehouses. Open source ETL applications are now available from a number of vendors - notably Talend and Pentaho - and do away completely with licensing costs – although you still have to consider support and development costs.

To use any ETL application successfully you need to understand and apply best practice, getting the software for free does not excuse you from this! In a project of any size you need at least one data warehousing expert on board.

Being new – and free – the buyer should be concerned about not only what functionality is built into the product, but also how mature the product is in terms of robustness – in layman’s terms, how bug-free is it?

Open source reporting software is also becoming available through the BIRT (Business Intelligence Reporting Tools) project and a small number of software vendors.

Idea 5 - Cloud Computing and Saas
“Cloud computing” delivers computer services across the Internet (also known as the “cloud”). For example, you might purchase space on a server owned by another company – or several servers – without having to purchase or maintain the servers yourself. Saas stands for “software-as-a-service”, which makes the applications themselves available online, usually through the Internet. Some of these have been around for quite a while, for example Hotmail and more recently Google Mail. A number of factors make Saas interesting, not the least of which is ability of suppliers to tailor their pricing structures. For example, a contact management system vendor might offer a free fully-functional trial period, with an increasing price according to how many contacts you have. Contracts may be monthly and can often be terminated at short notice.

The advantages of Saas include

low capital expenditure (you don’t need to buy new servers and software licenses)
scalability
low entry-cost

Disadvantages might include

dependence on network bandwidth
potential system downtime
loss of control over performance
outsourcing of security standards
loss of overall project control
locked into a single vendor
does not fully address ‘siloed data’ issues

Despite its catchy name, Saas is in effect a way of outsourcing your IT delivery. After all, goes the argument, it’s not your core business so why be involved with it? But outsourcing has its advantages and disadvantages, like any other management approach (see above).

Of course, as with most things in computer-land there is as much air-ware as there are real applications. But that is changing rapidly and will continue to change as more and more providers see the benefits of joining the “cloud” market.

Like all new things cloud computing has an appealing gloss about it. It presses the right buttons: low-cost, distributed computing power etc. But it suffers from the ‘version 1’ syndrome – untried technology (that will no doubt be resolved when you upgrade to version 2). Every project feels a bit like a pilot project because you don’t fully know what’s upstream.

Summary - A Low-Cost BI Reality Check

Here are a few points you should consider before committing all your resources to these appealing low-cost options.

1.Bear in mind that low-cost solutions are not suitable to all projects and all organisations. There can be a number of reasons for this, including the desire to retain control over one’s environment, the unsuitability of the application, inability to access one’s data as one wishes and inability to find the required functionality.

2.Low entry-cost in the form of lower capital expenditure (servers etc.) and license fees does not necessarily result in lower overall costs, because development and expertise costs must also be taken into account. These will be dictated by a number of other factors, such as the solution chosen and the scope of the project.
3.A more specialised understanding of data warehousing might be required if using a relatively low-cost product such as Sql Server 2005.

Business Intelligence is truly coming to the small and medium-sized organisation. Low-cost alternatives to the traditional enterprise-scale solutions are now ready and waiting for you to take advantage of. The question is, are you ready for them?