Nine fatal flaws

Gartner analysts outline common mistakes that can derail a BI implementation.

Leading organizations seeking to reduce costs and transform their operations recognize the value of business intelligence (BI). A strategically deployed BI effort is key to success, regardless of the economic climate. However, projects can fall victim to pitfalls, which can prolong implementation, thus increasing costs, or result in an under-utilized system that doesn’t meet expectations.

To prevent these common mistakes from derailing a BI endeavor, IT and business managers must recognize and address them. To that end, Gartner’s Bill Hostmann, research vice president and distinguished analyst, and James Richardson, research director, have outlined nine fatal flaws in BI implementations. Generally, the flaws focus on people rather than technology. Whether alone or in combination, these flaws can be critical.

1. “If we build it, they will come.”

The value of a new BI system is not intuitively understood by business users. If IT solely sponsors, funds and leads an implementation without engaging the business side, adoption of the new technology will be limited. It’s not enough to build a data warehouse and BI application and expect employees to eventually grasp its benefits.

Therefore, organizations must explain the value of the new system clearly to end users. “A BI implementation must be business-led, planned, controlled and managed to ensure the new system’s value becomes obvious,” says Hostmann. The best way is to start with a business problem. Engage business users, analysts and knowledge workers from the beginning so they realize that they are part of the team developing the solution.

The project team should include significant representation from the business side. Explain how the system will improve the way end users do their jobs, and answer the key question: “What’s in it for me?” Adopt a BI competency center (BICC) to drive adoption of BI and to gather required skills.

2. “Managers dancing with the numbers.”

To maintain control of their data and sometimes to mask performance, some managers resist sharing their information more broadly throughout the organization. These managers often provide insight into their data and performance only via spreadsheets. Spreadsheets, and the level of non-standard calculations embedded within them, make it impossible to audit and compare relevant data with other managers, departments and information sources. More importantly, data silos could lead to a lack of corporate transparency, which can increase the firm’s overall risk exposure.

To solve the problem, organizations must limit the use of spreadsheets for formal reporting and management meetings. Use regulations such as the Sarbanes-Oxley Act to educate business users about the risks posed by spreadsheets and the unclear ownership of data. Establish a more rigorous reporting and analysis process, standards and governance that can serve a broad range of business managers’ needs and meet the requirements for auditability and access.

To make this work, business sponsors must believe in a fact-based approach to management and have the strength to cut through political barriers. A methodical approach to the project can expose the hidden agendas of people who create political barriers. A business culture shift to one of cross-organizational performance transparency is also needed and must be fostered by management. Publicizing the success of existing cross-team BI initiatives will help encourage the removal of data silos.

3. “Data quality problem? What data quality problem?”

Bad data leads to bad decisions. Employees won’t touch applications that use irrelevant, incomplete or unreliable data. It becomes counterproductive for them to try to decide which figures are correct or how to reconcile differences. So, if users do not trust the BI system, they will ignore or work around it, until a crisis forces the organization to overhaul it.

Companies must make it clear to everyone that corporate data integrity is a priority. They must set up checks and balances, as well as metrics that define acceptable levels of data quality to ensure that it is accurate from the outset and stays that way.

Organizations need to establish a BI platform framework that can handle updated corporate-wide business strategies and processes. The key is avoiding scope-creep.

That requires automated controls to identify data quality issues in incoming data and block questionable data from entering the data warehouse. Firms must also implement a back-end process for auditing and verifying data, including reconciling warehouse data with source systems. And to drive accountability, data quality must be included in job descriptions and performance reviews for the data stewards of the business applications that are sources that provide the data used in BI.

4. “Evaluate other BI platforms? Why bother?”

Many users believe that the primary vendor’s system is the only one they will ever need. One-stop shopping doesn’t necessarily lower the total cost of ownership or deliver the best fit for an organization’s needs. As a rule of thumb, most single vendors’ service suite applications can handle about 80% of a typical user’s needs.

As a result, some organizations, especially firms with market-differentiating processes and practices, need to keep their BI platform and applications flexible and adaptable to a wide range of application needs. That may include seeking best-of-breed applications to protect their competitive advantage. Start by examining the BI functions from your enterprise application vendor and comparing them with what else is out there. However, Hostmann issues a strong caveat: “Avoid point solutions that cannot be linked easily into the IT infrastructure or business applications.”

5. “It’s perfect as it is. Don’t ever change.”

Some managers assume the business will always be the same, so they want to leave everything as it is. But BI is an ongoing exercise. “Often, organizations focus too much on perfecting today’s business processes rather than anticipating tomorrow’s changes,” says Hostmann. “Successful implementations need to combine both—analyzing current problems while planning for evolution in the future. It’s a tough juggling act.”

Rather than treating BI as a series of discrete projects delivering a fixed set of applications, organizations need to establish a BI platform framework that can handle updated corporate-wide business strategies and processes. The key is avoiding scope-creep—projects that never end.

To do so, firms must introduce alternative solutions to fill in the gaps. This includes defining a review process to manage obsolescence and replacement within your BI portfolio. They must also examine possible innovations such as new algorithms for evaluating gaps in the legacy system.

6. “Let’s just outsource the whole darn BI thing.”

Firms struggling with BI efforts mistakenly believe they can dump their problems onto a third party and save money. But focusing too much on costs and development time often results in inflexible, poorly architected systems. “Organizations understand their own business better than anyone else,” says Hostmann. “But if outsourcers don’t grasp what makes you better than the competition or it takes them too long to learn, outsourcing may not be worth it.”

Instead, companies should outsource only non-core BI skills and competencies. Or, if necessary, they should rely on outsourcing only temporarily while building up skills within their IT organization.

In that way, by focusing on critical activities—growing and transforming their business—enterprises can continuously improve their processes while upgrading in-house skills.

7. “Just give me a dashboard. Now!”

IT departments will sometimes give in when managers lean on them to build dashboards quickly and cheaply. Increased demands on management are driving these types of requests. “Every day there are more and more regulatory requirements,” says Hostmann. “Senior executives need to have up-to-date status reports on all potential risks and their impact on the business since they may need to sign off on them.”

However, if dashboards aren’t linked to corporate objectives or the data sources used in business operations, they deliver little value. A more constructive response is to create scorecards and dashboards in conjunction with a strategy map—a cause-and-effect diagram of the objectives, applications and data sources. Otherwise, a scorecard offers little more than a collection of unrelated metrics and dials.

In the short term, to keep executives up to date, IT needs to make reports as pictorial as possible by including charts and visuals to forestall immediate demands for dashboards. Later, they can include plans for more sophisticated visualization tools in the BI adoption strategy.

The project team sometimes believes that they all know exactly what they’re supposed to do, so they don’t bother with a plan. That’s the final and biggest flaw.

8. “X + Y = Z, doesn’t it?”

A BI initiative aims to create a single view of the business, but many organizations haven’t even agreed on the definition of fundamentals, such as “revenue.” They end up creating siloed BI implementations that perpetuate the disparate definitions of their systems. In doing so, different data leads to deciphering, translating and interpreting information to establish a consensus among departments.

To achieve one vision of reality requires cross-departmental agreement on how business entities—customers, products, key performance indicators, metrics and so on—are defined. IT organizations must ensure that the standardized definitions, processes and calculation methods are all centrally coordinated and controlled. In that way, everyone can agree on the numbers and what they mean. They also need to align definitions across activities, even if they cannot execute a single, enterprise-wide BI implementation.

9. “BI strategy? No thanks, we’ll just follow our noses.”

The project team members sometimes believe that they all know exactly what they’re supposed to do, so they don’t bother with a plan. That’s the final and biggest flaw—the lack of a documented BI strategy or the use of a poorly developed or implemented one.

One solution is to create a BICC with members drawn from the IT organization and the business side to develop the overall corporate strategy and oversee the actual implementation. After the project is over, the BICC should remain intact to glean best practices from the implementation and start planning for Phase II.

By doing so, as market conditions change, the organization not only has an updated BI platform but also a path for renewing it. This will ensure continued fact-based analysis, enabling the organization to compete, grow and prosper.

No time like the present

Despite market uncertainty, now is an opportune time for far-sighted companies to initiate transformational projects such as BI implementations or upgrades. This will enable them to roar back during the economic recovery.

Adjusting to the current economy requires every organization to reconsider its approach. “All business information and analysis strategies and implementations need to be re-evaluated and most likely changed, given the major changes in the business environment,” says Hostmann.

Your Comment:
Your Rating:

Fuzzy Logix