Intelligent thinking

28 May 2007

Picture the scene. Ten metres from the summit of Everest, you’ve done all the hard work, the world is at your feet but there’s nothing left in your engine. You can’t make the last few steps and the moment of conquest eludes you. This is Michael Kearney’s analogy of what happens to companies when they fail to invest in business intelligence (BI).

As the Irish country manager of SAS, one of the global leaders in the field, he can be forgiven for making BI the difference between ultimate success and failure, but it’s an analysis he is more than happy to explain. “The vast majority of companies have invested huge sums in operational efficiencies. They have done all the hard work but they don’t get to have their photo taken at the top,” he says.

He describes how organisations are awash with data, veins of information that could be worth millions to the business if there was only an efficient way of mining it and turning it into business intelligence. “There are companies with massive legacy systems, for example, that presume there is no additional value they can extract from the asset because they are 20-30 years old and waiting to be replaced. I would argue very strongly that additional business benefits could be found without implementing anything else.”

What Kearney calls for is a change of mindset from his prospective customers. Sales reports from a previous quarter are not just barometers of success, they are also the keys to unlocking the future. If an organisation has a better idea of how it performed in the past, it will be more able to plan ahead.

“Managing directors want a crystal ball that tells them who to target in the next quarter, which customers they are likely to lose and which are the most profitable. They need to know if there is a product that needs to be enhanced or an inventory shortage that is likely to damage the business.”

The data explosion has been well documented and organisations have no shortage of information about their business. The problem is that senior management is wasting valuable time trawling through it to find what it needs, according to Kearney.

This is good news for BI specialists such as SAS, Cognos and Business Objects who have a variety of ‘push and pull’ software solutions for fast access to critical data. The problem, says Kearney, is that it can be a hard sell because many organisations think they have an adequate BI strategy in place. What they really have are simplistic reporting tools of questionable value.

“It is up to us to tell our customers about the possibilities of business intelligence and performance management. It’s more than reporting, it’s about having a holistic view of the business that is made possible by individual applications and tools,” he says.

Part of the challenge comes down to the way that BI software has evolved. Dr Brendan Goodman, HP Ireland’s head of business intelligence, argues that it’s been a process of evolution and we’re only getting to the point where forecasting is even on the horizon for most Irish companies.

He describes this type of data mining as fourth generation, a complex process confined to large organisations such as financial institutions and telcos. “They are now able to start predicting what is going to happen. Telcos, for example, want to be able to predict churn. And the business feedback will be used make customer-specific offers based on profiles, enticing people to stay with them.”

Evolving intelligence

The first generation of BI was pre-defined reports that came out of packaged systems. This was followed by ad-hoc queries, where data was exported to Excel, Lotus or specialist tools from companies.

This is the stage that most Irish companies are at now, according to Goodman. HP estimates that only around 20pc have moved on to the next phase, improving data warehousing and OLAP (online analytical processing) style analysis that works off pre-configured databases.

To put this in context, and to confirm suspicions that Ireland is falling behind, Goodman says that it’s two years since the US moved to the fourth stage of evolution. He attributes the slow uptake in Ireland to a perception problem. “Unless organisations can see the value in BI it can be seen as a luxury add-on,” he says.

There are signs that this is starting to change because of a number of drivers. “Companies see their competitors are doing better and they realise that a key reason is that they have good BI solutions. There is a profit and performance motive in the private sector that drives uptake.”

Another driver is compliance. The demands placed on US multinationals by the Sarbanes-Oxley Act of 2002 has filtered through to local Irish offices, spreading the awareness of the need for improved reporting tools.

The biggest barrier, however, and the one that is being most spectacularly eroded, has been cost of entry. Microsoft’s strategic desire to extend its fingers into every piece of the corporate pie took a new turn with SQL Server 2005, a database software platform that now incorporates BI tools at no extra cost. It integrates multiple data sources and provides a platform for complex business analytics.

This was Microsoft’s first major move into a market that is essentially divided into two halves. At the backend there are tools to extract information from databases to populate the data warehouse. At the front, there are the tools for querying and analysing the warehouse. “You must have a properly structured data warehouse to mine information properly,” explains Goodman. “If you don’t, people will use the tools to ask questions and get the wrong answers.”

Forrester Research highlighted this problem when observing that data volumes are growing while information extraction is declining. Source data is scattered throughout the organisation making report creation hard and time consuming.

Gartner says that 60-70pc of the BI challenge is about cleansing the data, getting it out, transforming it correctly and storing it in a properly designed warehouse, a series of processes tagged ETL (extract transform load).

“If you can do that and have reasonable control of the toolset, you won’t generally get bad information,” says Goodman. “But the classic scenario is still true: a fool with a tool is still a fool. You used to need people with PhDs to know how to mine the data, but that’s all changed.”

Richard Moore, business group manager with Microsoft Ireland, agrees:
“There was an issue in that it was hard to access the different types of data. We have solved that with SQL Server 2005 and made the solutions more affordable and more accessible.”

But an upgraded SQL Server is only half of the story. While the backend tools have delivered many of the features of its competitors at a reduced cost, Microsoft’s front-end reporting tools have not been as strong. Last year it acquired analytics specialist ProClarity to address the problem.

“It’s no good in having just the backend. What we are interested in is the democratisation of data mining,” says Moore. “In the past it was the job of a few analysts to take the data and make it available for the senior management. Now managers throughout an organisation can access intelligence and present reports to their teams in Excel or PowerPoint.”

He argues that Microsoft has made the whole realm of BI not only more affordable but also more accessible because it all takes place within the familiar Windows environment.

Unlike the components incorporated into SQL Server, the new tools are sold as an add-on, but still work out cheaper than many of its competitors, said Goodman. HP partners both Oracle and Microsoft, acting as a system integrator, delivering its customers a BI solution. He believes the Microsoft proposition heralds a significant change in the market.

“You no longer have to spend thousands to build a medium-sized BI environment, and it’s encouraging more people into the market,” he says. This is confirmed by a recent IDC report on business analytics (September 2006) that predicts 10pc growth in the market from 2006 to 2010.

While Michael Kearney, SAS country manager, agrees that BI is “opening up to the masses”, he is less impressed by some of the new pretenders that are muscling in on his space. Not just Microsoft, but SAP and Oracle are also selling BI as an add-on to their enterprise resource planning (ERP) systems.

“SAP and Oracle have a pedigree in transactional, not analytical, systems. They have saved companies money in operational efficiencies but they haven’t helped customers develop a strategy going forward,” he says. “The good thing is that they have helped validate the sector with their BI offerings, but the logical extension of having spent millions on world-class transactional systems is to spend a fraction of that cost on a world-class analytics system like ours. Putting in SAP or Oracle analytics is like putting furry dice in a Rolls Royce.”

A strategy for BI

Gartner offers two approaches to improving performance analysis, outlined here with the help of HP’s Dr Brendan Goodman.

1. Make the business case
Like many areas of IT, if the business fails to buy into a solution it probably isn’t going to succeed. More so than with other technologies, the driver for deploying BI software and developing a strategy must come from making a strong business case. An isolated pilot, run by the IT department, is not the way forward.

2. Find an approach
According to Gartner, a BI rollout falls into two categories, tactical or strategic:

A. Tactical
Identify a particular business problem, such as the pursuit of accurate data to support a direct marketing campaign. Identify the data sources required to make it valid and then look at the technology to deploy. Run a pilot and proof of concept.

HP’s Dr Brendan Goodman warns, however, against spending thousands of euro on new tools for a pilot. “Try to fuse what’s there already; do simple ETL that gets the data out of the source system relatively cheaply. If you have to make an investment at this stage, make it at the reporting and analytical end.”

Although few vendors would volunteer it, he says that some of them may even loan the technology, especially if they feel an investment is likely to come out of it.

According to Goodman, this tactical approach is the best way forward for most Irish organisations.

The next step is to build a data mart, where one or more databases are targeted for strategic interpretation. The next trick is to ensure that it can be easily populated on a regular basis with the data you need. A common failing with pilots is that they only work efficiently on a one-off basis and are not built with longevity in mind.

Complexities with this approach mean that the company would be well advised to hire a consultant, typically for a period of 5-10 days. “If you’re not prepared to do this it’s probably not worth undertaking the project,” says Goodman, unashamedly selling his company’s own angle on BI. “If you don’t, you risk ending up with a badly designed structure that can’t be expanded and gives you wrong answers.”

B. Strategic

Define strategic objectives for the organisation and develop a BI approach to support these objectives. Implement a phased approach.

According to Gartner, such a move has to be entrusted to a BI competency centre within the organisation. It must provide a strong, business-centric focus with clear performance management targets.

The people and processes have to be in place, along with the analytical tools at the front end, but the core foundation for success is the BI platform, the databases that run in the data warehouse environment. Gartner says that 60-70pc of a BI strategy is about the platform level.

“It’s all about getting the data together in a quality manner. If you don’t have that you could be wasting your time,” says Goodman.

Down to business

Irish organisations are seldom at the leading edge when it comes to implementing new technology and it seems BI is no exception. There is a reluctance in parts of the Irish market to implement it, says Ronan Stafford, CEO of Inflection Point, part of the Codec Group.

“There’s no culture of adopting a full, best practice strategy,” Stafford asserts. “What Irish companies are interested in is answering more tactical questions and less strategic ones.” As a result, they have tended to invest in point systems for solving a particular problem. Cost is not a barrier – BI systems are not typically expensive and Stafford notes that many organisations are happy to spend money on a disaster recovery system that helps them to manage risk.

Some Irish companies are starting to think strategically and linking that to budgets and monthly performance, but they’re not necessarily doing so because they see potential business benefits. “The reason why is not to run the business better but because they’re compelled to do so by compliance regulations,” he believes. “Doing it because you have a gun to your head is the worst reason to do BI.”

There are different approaches to implementing BI and Irish companies illustrate this. The construction group Kingspan is rolling out BI on a phased basis rather than going for a ‘big bang’, first implementing it for the annual budget and planning. This will be later expanded to provide key performance indicators, followed by strategic management and finally forecasting. Similarly Irish Distillers currently uses BI for budgeting and forecasting. “BI guarantees the numbers are correct and lets them analyse the budget to see what aspects are performing well,” says Stafford.

This pragmatic approach offers the comfort of quick wins and can save considerable time in tasks such as preparing financials, but there is a caveat, Stafford cautions. If a budgeting system is to be part of a longer-term BI implementation, this will have to be incorporated into the design phase so that elements to be added later can be integrated more easily.

Aer Lingus, by contrast, is running three BI projects in parallel: one for market analysis, the second for financial planning and the third for payroll and salary planning. “If the company is large enough and has an appetite for it, you can do that,” says Stafford. “There’s no approach that is better than the other,” he adds.

BI doesn’t have to be a long, drawn-out project either: the return on investment kicks in after an implementation period of two to three months, says Stafford. “It starts paying for itself quickly. For example, if it’s a sales application, the team can get visibility they’ve never had before.”

By Ian Campbell

This article appeared as the cover story in the May/June 2007 edition of Knowledge Ireland, the magazine for knowledge leaders. For more information visit