BI comes back into vogue

Business intelligence (BI) has come back into vogue, and it`s going mainstream. This is the message coming through loud and clear from both vendors and users that completed ITWeb`s Business Intelligence 2005 Survey.

One of the main drivers in the BI space is the current market condition. "From a business perspective, times have never been tougher. Our customers are faced with an ever-increasing number of strategic decisions. There have never been so many regulations to comply with and never has there been such an imperative demand for accurate information," says BI software and services provider SAS`s MD Bill Hoggarth, who spoke at the release of the survey results at an event hosted by ITWeb in Johannesburg recently.

This is the fourth consecutive year ITWeb has conducted this survey, which ran online for four weeks in July and August. This year`s survey received almost double the responses compared to 2004.

Over 30% of the 2005 respondent base hold managerial positions, another 16% are employed at executive level, 21% categorise themselves as general staff, 16% are consultants and 12% work in sales.

"The demand for this technology is starting to come from line-of-business managers who want answers to their problems through intuitive analytical software, dashboards, balanced business scorecards and key performance indicators," says Kevin Simpson, marketing director at SCS Africa, an IT solutions company.

There has never been so much of an imperative demand for accurate information.

Bill Hoggarth, MD, SAS

According to this year`s survey, top of the list of items being considered for adoption within the next 12 months by the respondents are dashboards (28%), followed by data quality management tools (19%), customer data integration hubs (17%), enterprise information integration offerings (16%), data migration/ETL tools (12%) and metadata repositories (8%).

"Many companies are now looking at implementing dashboards, but these will only provide real value if the underlying data is accurate and relevant," says Martin Rennhackkamp, strategic information manager at Prescient Business Technologies, a data warehousing and BI solutions provider.

"Poor data quality is still a growing issue, and people are starting to realise just how big a negative impact dirty data can make," he says.

Dr Pierre Mulder, chairman at software development house Redscreen, concurs, but says business decisions will continue to have to be made, regardless of the quality of the company`s data. "What one person sees as unacceptable data, someone else might view as important information. Generally speaking, people are looking for information rather than data, which is why the executive dashboard issue is coming through now."

Another speaker, BIPractice practice manager Estelle de Beer, says is it "shocking" to see the quality of some companies` data. Among the reasons why data has become so unreliable is the fact that data transitions and movement has almost become default behaviour - and things can go horribly wrong here.

Data migration and consolidation errors, and changes to source systems are also contributing to dirty data, she says. The impact of bad data includes the erosion of a company`s credibility with customers and suppliers, poor customer service, lost opportunities, significant compliance issues and the inability to detect - among other issues - fraud and overpayments, states De Beer. "The single view everyone is aiming for will never happen without consistent, accurate underlying data."

Asked about the impact of dirty data on their BI systems and processes, the 2005 survey respondents indicated that this resulted in poor decision-making (24%), opportunities not being identified (15%), lost or dissatisfied customers (14%), increased costs (14%), business processes disruption (13%) and lost revenue (12%.)

"Bad data is only good data when you know how bad it is," says Alastair Jacobs, solutions manager at Oracle and another guest speaker at the event.

The main reason for dirty data appears to be a general tendency to "pass the buck" when it comes to accepting responsibility for quality assurances. Over a third of the survey respondents said that within their companies the IT function is responsible for data quality, a quarter said business representatives are held accountable while 33% said both are charged with quality control.

However, the real issue is the failure of employees, right across the board in an enterprise, to ensure the information they personally touch or deal with, within their working environment, is accurate.

"Everyone uses the data, but no one wants to own it," says De Beer, adding that quality assurance should become part of the standard key performance indicators and a measured success criteria of all projects at local companies.

When asked "who makes the decision to invest in BI initiatives in your company?" four out of every 10 respondents said this was their "executive management team", rather than specifying a specific C-class executive. Given that almost every possible C-class designation was also given as a possible answer to this question, it`s clear the respondents aren`t even sure exactly who it is that champions the delivery of BI in their companies.

"Successful BI requires a `champion` within the company who is sufficiently senior to drive the implementation within the company, can ensure that a high level of data quality is being fed into the BI solution and also knows the business well enough to ensure the relevance of solution outputs to both operational and strategic decision-making," says Andrew Connold, MD of Synergy Computing, a BI solutions implementer.

Keith Jones, MD of BI software distributor Harvey Jones Systems, says dirty data is now being addressed as enterprise resource planning (ERP) systems are beginning to deliver "on their original promises. Data is now cleaner and more accessible and as a result, business intelligence is receiving more attention."

Hoggarth shares his view. "ERP systems have matured a great deal over the past 10 years. These projects are now delivering on their promises. People are waking up to what their ERP systems are good at and what they`re not good at. They`re great for BI - they provide us with a great store of raw data."

Caron Mooney, a director at customer relationship management (CRM) and BI solutions provider IS Partners, says: "As far as ERP systems are concerned, they now have to include business intelligence. It`s a tick in the box now."

Rennhackkamp, who says there has been a steady growth of about 12% in the BI market during the past year, has a slightly different view. "Most ERP systems don`t yet span the whole business, so although ERP vendors are including this functionality, it`s not exactly drawing all the required information across the enterprise."

He believes "people are fast realising the value of information more and more, and the gathering of BI is slowly moving up the company hierarchy. Previously it would have sat mainly in the IT department. But, he states: "It remains a challenge to get executives to accept responsibility for data quality and filter this responsibility down into the organisation."

Bad data is only good data when you know how bad it is.

Alastair Jacobs, solutions manager, Oracle

The survey respondents share his opinion. Among the major challenges they list, are obtaining buy-in from both end-users and management. Several respondents cited their frustration around this issue, stating there was a real need for education with regards to the importance of information to the business.

They indicated that the main drivers for data quality improvements in their companies were BI (35%), CRM (31%) and application migrations (14%).

Roughly half of the respondents to this survey are based at companies with less than 500 employees, about 40% employed at concerns with between 500 to 10 000 staff members, and 12% from businesses with staff complements over 10 000.

In terms of where they are currently at, almost a third indicated that their BI implementations are at the mature deployment stage. Just over 20% are considering adopting a BI product or solution, 15% have begun to roll-out an implementation, and 11% are still at the planning phase. Two percent are in decision mode, 7% have no plans to adopt, and 13% were not sure of the current status of BI adoption in their company.

There appears to be a fairly high number of new initiatives at local companies, with 21% of respondents reporting that their solutions had been deployed less than a year ago. Twenty-two percent said theirs had been up and running for between 12 and 24 months, 32% said theirs had been in progress for three years or more, while a quarter of the sample base were uncertain as to when theirs had been implemented.

The top three reasons the respondents gave for deploying a BI solution were to promote quicker decision-making processes, improve customer service and positively impact the company`s bottom line. Better utilisation of human resources came in fourth place, while improved time to market for new offerings was given as a fifth reason.

Four out of every 10 respondents said their company had received a return on investment. A third said they had not, while 34% were not sure. The results of last year`s survey indicated that 30% had received a return, 24% had not, and 46% were unsure.

Outsourcing is still not a big focus in the BI space, with only 10% of respondents indicating they had outsourced. Almost half had developed their solution in-house, while just under 30% had approached suppliers for assistance in rolling out a solution.

Forty percent of respondents said their investment in BI projects was being funded by the overall IT budget while 20% had drawn funding from their departmental budgets.

Sixteen percent of initiatives are funded through business development monies. The top three criteria for identifying an implementation`s success were given as efficiency (20%), productivity (16%) and a return on investment (15%). Ease-of-use came in fourth position, scalability fifth and the manageability of the product/solutions took sixth place.

Mooney says ease-of-use is "absolutely critical and this is an area that some companies just don`t get. They spend up to a million rand putting a system in and then only buy five licences. Basically giving as many users the right tools for the right job facilitates ease-of-use across the company."

"Lots of guys are now adopting a `pay as you go` approach, with deals starting at between 25 to 30 seats and growing into implementations covering from 400 to 500 users," says Jones. He cites as examples retailers Shoprite, which started with 25 seats about 18 months ago and is now up to more than 500, and Edcon, which is adding between 100 and 200 users every few months.

Another corporate that has gone the incremental route is First National Bank (FNB). It started its Siebel implementation four-and-a-half years ago, says its business intelligence unit head, Werner Swanepoel. "We then had about 100 seats across the country, largely in the bigger centres, which were deployed to the sales force." FNB now has between 700 to 800 seats in the sales, operations, back-end, call centres, support services and executive management departments.

Graham Mansfield, GM of Siebel Systems South Africa, says one of the main success factors was the "strong executive support" the initiative received within FNB. Siebel provides CRM, BI and customer data integration solutions.

FNB has also ensured that while the BI unit is responsible for measuring the accuracy of the data, each business unit is responsible for the quality of its own data. Touching on the issue of dirty data, Swanepoel says addressing this issue means first "understanding what you have and what you can do with it. If we wait for perfect data, we`ll never do business."

Mulder says the current legislative compliancy requirements are "making banks more equal, so they`re having to find other ways of being more competitive". He believes BI has a significant role to play here - first in attaining a single view of the customer across the bank, and then developing that into an "enriched" view that includes external information on each client.

If we wait for perfect data, we`ll never do business.

Werner Swanepoel, business intelligence unit head, FNB

According to several research firms, many corporates are now standardising across their enterprises. "BI is an absolutely essential component of any permanent management initiative, and standardisation makes this so much easier," says David McWilliams, MD of BI and performance management software provider Cognos. McWilliams also presented at the Johannesburg event.

"Companies are definitely standardising - there`s such a circus of products out there that they are finding it hard to manage and support them - in some cases, business intelligence products and solutions have become a management overhead for companies. Also, today most offerings can do 80% of what`s required, so a collection of different ones doesn`t make sense," says Rennhackkamp.

"Historically, different lines of business in organisations have had different information needs, access requirements and report preferences. Accordingly, business intelligence vendors delivered separate and distinct tools to satisfy them: one set of tools for production reporting, another for ad hoc query and analysis, another for multidimensional analysis, and yet another for data mining and statistical analysis," says Marc Scheepbouwer, MD of BI and business performance management solutions distributor Intellient.

Forrester says most Global 2000 companies and government agencies today have between five and 15 different BI solutions in production, often nothing more than collections of unconnected platforms and tools.

"This fragmentation is a huge barrier to pervasive business intelligence, which is a must for finding that treasured success. Business users cannot work the way they would like and even expert users are burdened by having to learn multiple tools. Fragmented business intelligence solutions place unnecessary demands on the IT professionals who must deploy and manage them," states Scheepbouwer.

Benefits here, says McWilliams, include fewer staff training courses being required, fewer help-desk and support centres, less time and resources spent on installing, upgrading and maintenance, and negotiating with vendors.

Asked whether advanced analytics or data mining played a significant role in BI in their companies, the respondents replied as follows: Yes - 55%, No - 28%, Not sure - 17%.

Corporates are slowly starting to use predictive analytics and data mining - mainly in the CRM arena.

Martin Rennhackkamp, strategic information manager, Prescient Business Technologies

"Corporates are slowly starting to use predictive analytics and data mining - mainly in the CRM arena. However, some of these analytic drives are not closely linked to the corporate data warehousing initiative, so we`re seeing a fair amount of wastage and duplication of effort," says Rennhacckkamp.

"We`re also still seeing a lot of silo-based initiatives and the business intelligence mandate often doesn`t reside high enough in the organisation for this to get corrected. The problem is aggravated in companies where there is more than one IT department," he says.

The next question in the survey: "Do you see BI as including predictive analytics and forecasting?" - drew the following responses: Yes - 76%, No - 8%, Not sure - 16%.

Respondents regarded query and analysis tools (46%) as the technology component most fundamental to the success of their BI implementation, with their data store ranked second (13%) and data quality tools in third place (10%).

Usage of BI end-user access tools reflects the same profile as last year`s survey. First up, is accounting/finance (17%), respectively followed by sales and marketing, at 14% each. In third place at 13% is corporate management, with technical/scientific/engineering ranked at 9%. External customer support and human resources got the next slot here (7%), with inventory/materials management and purchasing coming in at 6%.

Over 40% of respondents said the level of end-user literacy in their companies is poor, while a third feel it is "adequate". Fourteen percent said it is excellent, while 14% felt they are "not sure".

"There are a number of factors that affect a successful BI implementation, but the single driving factor is user adoption," says Jones. "If the solution is adopted, it will be successful. User adoption is driven by a number of factors - some political and some tangible. The users must trust the data and feel comfortable exploring it themselves. The solution must deliver valuable information in enough time to be useful."

Asked about future growth areas, Mooney says: "Government is going to be a big source of revenue in the BI space. Corporate governance is also starting to have an impact on business intelligence, as companies now have to ensure they are perceived as being as transparent as possible and able to consistently deliver accurate information. BI is also becoming of increasing importance in the retail and distribution sectors, where margins are tightening due to a stronger rand currency.

"Government is a non-profit organisation," says Jacobs. Current legislation is forcing government "to think through its budgets more carefully. BI will be critical for that analysis. A technology solution here will have to manage the impact of legislative changes, the need to migrate to a government e-standard, and complicated integration issues."

There will definitely have to be significant standardisation. "I don`t think there`s a [BI] product out there that hasn`t been implemented somewhere in government," Jacobs believes.

"The release of Microsoft SQL 2005 will push BI much lower in terms of the size of the business. This means that many people, who currently do not know much about it, will now get involved and do it badly. This is a standard response when anything is commoditised," she says.

Jones believes consolidation within the industry will continue. "A new wave of technology is also on the horizon, which will see new niche players being introduced. This will confuse the market for a while," he says. However, larger corporates, states Jones, will continue to cut the number of tools in their enterprises and rather focus on standardising their solutions.