Today, let's talk about BI Project, especially about why the project failed. In fact, BI projects fail at an astonishingly high rate – between 70 percent and 80 percent, according to Gartner. It is very interesting because I also have experienced "not good" BI project. Many factors that happened at that time. One of biggest factor I hate very much is "politic". My case may be an extra practice from the Article I would like to share here by Information Builders.
In my case you should have strong sponsor that can support your team. BI project involve so many stakeholders if the scope of work is enterprise wide. You gonna face people across division and departments. Many department / division means many boss that has their own purpose and ego. You need to have big big boss in your side to ensure the boss under the big big boss support you. What am I talking about? I talk about data privacy and ownership. Sometimes even in one company, the unit under them is not synergy and not trust each other. What a bad teamwork?!
Okay now let's speak about the main topic of this article : Top Six Worst Practices in Business Intelligence. Let's start 1 by 1 :
Worst Practice 1: Buying What Analysts Want Without Considering Other Users
Most companies make business intelligence (BI) purchasing decisions based on input from
just one group of users – the business analysts. This single perspective, however, creates many problems.
Analysts select tools that they are already familiar with, or ones that are similar to those they have used in past. This severely limits broad BI adoption across an organization, and minimizes ROI. Because analyst-chosen tools are heavily biased towards their own skills and needs, they are way too complex for the average business user. The majority of users within an organization require extensive assistance from either the analysts themselves, or IT staff, in order to access and interact with enterprise information.
Perhaps this is why, since 2009, BI usage rates have remained flat at around 25 percent. BI Scorecard’s Cindi Howson says that close to 80 percent of employees who make decisions still “lack the tools to make them on facts, relying instead on static spreadmarts or gut feel.”
Ease of use is a critical success factor for BI, but what is intuitive for one type of user may be complicated for another. Tools are for professionals, not for typical information consumers. Business users don’t need tools; they need different, more intuitive approaches to analysis. Companies that implement only on a handful of core tools that offer many powerful analytical functions through sophisticated interfaces – the tools that the analysts want – will alienate a large portion of their BI audience.
Empowering different types of information consumers with different analytical tools is the key to BI pervasiveness, and ultimately BI success. Extending BI and analytics to all users –particularly frontline and operational employees, customers, and business partners – promotes better decision-making enterprise-wide and aligns operations to strategic and growth goals.
The most successful BI strategies take all information needs into account, and ensure that
supporting solutions satisfy the requirements of not only the analysts and power users, but also many different kinds of information consumers at the strategic, tactical, and operational levels. The selection committee should include business users as well as analysts to ensure that all users can embrace the chosen solution, regardless of their skill set or technical savvy.
I think it depends on your project stakeholder. Is it enterprise wide or only 1 department? By the way, is it align to the IT strategy? Some products are departmental, some products are enterprise ready. Sometimes when we talk about the "future" plan, some customer will prefer buy enterprise ready product, but of course it will be "overkill" until the time is come. Otherwise they will choose the departmental, and when in the future the customer need to more, or other department want to use / leverage the BI, it will become problem too. In my perspective we should choose a product that can support small or big. Simple.
Worst Practice 2: Buying New Data Discovery Tools Without Changing the Excel Mindset
Many companies rely on Microsoft Excel to facilitate the analysis and sharing of vital business information. The problems with this approach are well known: multiple versions of the truth caused by limited version control, lack of auditing ability, and high error rates and data quality problems. These issues negatively impact planning and decision-making, and the damage increases as disparate and conflicting spreadsheets circulate throughout an organization.
Few realize, however, that the slick new data discovery tools they’ve implemented to supplement those spreadsheets are nothing more than prettier versions of Excel. These products may enhance the visual appeal of the information being analyzed, but they still create the risk of inconsistent insight and flawed decisions. Like Excel, they lack version control and auditing ability, and have no way to ensure data integrity. Users have their own data sets and their own means of manipulating information, so they’ll arrive at different conclusions even though they use the same tools.
The most successful BI environments keep people connected and ensure a single, consistent view of enterprise information – even if they use different features and capabilities to perform their analyses.
Companies need to move away from the Excel mindset, and implement a broad-reaching BI platform with a wide array of functionality. This eliminates the drawbacks of Excel and data discovery tools, while letting users conduct analyses and manipulate data the way they want to.
Whether they want spreadsheets, sophisticated visualizations, or pre-built BI apps that provide direct answers to specific business questions, a comprehensive BI platform will satisfy everyone’s needs, while preserving integrity, consistency, and auditing ability.
This is a fixed price. Excel is very good but they are not the best. You cannot change / replace it completely but you have to try what is impossible. The security and inconsistency are the biggest risks when we still keep using excel. This can be a good reason but don't forget to provide the replacement. Export feature is a must to address the need of excel.
Worst Practice 3: Making a BI Purchasing Decision Based on One Hot Feature
Companies often purchase a BI tool as a knee-jerk reaction to very specific, very narrow demands. A functional user insists on a certain feature to make her job a bit easier. A business manager wants a new analytical capability to solve a specific problem. An executive reads about new functionality and pushes for its purchase simply because it’s “cool”. These are typical scenarios, where one user or user group hijacks the entire evaluation process.
This tunnel vision is harmful to any business intelligence strategy. When BI efforts focus on just one requirement, broader-reaching analysis needs or future requirements are left out of the planning and solution-selection process. As needs change and grow, or as new requirements emerge, organizations struggle to evolve and expand the BI environment accordingly. They’ll purchase a series of disparate tools to address one need at a time, creating BI silos throughout the enterprise and driving up total cost of ownership. They’ll also frequently ignore the back-end information infrastructure to deliver the needed feature as quickly as possible, leading to a long-term maintenance nightmare.
Additionally, once they learn more about the selected tool, users who focused on a single feature in the first place often find the tool insufficient to address their higher-level needs. They go back to IT with demands for new tools, and the process starts all over again.
According to Wayne Eckerson, director of BI Leadership Research, some tools “satisfy the parochial needs of individual workgroups or departments, while BI suites provide an integrated experience and architecture that addresses the entire spectrum of BI needs in an organization and is thus easier to administer.”
The most effective and economical approach is to choose a flexible and extensible platform with a broad range of capabilities. Companies can deploy the most urgent features and functions right away – the dashboard that users say they can’t live without, or the customer-facing BI environment that will keep clients from defecting to a competitor – and then easily add predictive analytics, data visualization, enterprise search, or other advanced capabilities as new needs arise.
Even if the BI platform lacks the hot, new feature, the vendor will most likely add it soon. On the other hand, the vendor of the “popular” feature won’t be able to quickly catch up on all the other capabilities that are already missing from their tool.
The platform approach addresses immediate requirements, while future-proofing the BI strategy. It also avoids the selection bias highlighted in the previous worst practice, by meeting the needs of all present and future stakeholders – not just those who make the most noise to get what they want.
Buy a BI product that can satisfy all stakeholder and also has a clear vision and if necessary have a clear roadmap for their product. To make sure of it, you can see it from how fast the company release the new patch or upgrade. How fast the company follow the market or become the market leader. You can see the commitment on it because most of the product will force you to buy ATS / renewal. It is not cheap.
Worst Practice 4: Lack of a Concrete Data Quality Strategy
Most organizations want to give their analysts new business intelligence tools as quickly as possible. Yet, in their rush to rapidly implement and roll out bigger and better BI capabilities, they fail to consider the integrity of the information sources with which those analysts work. They either overlook data quality needs completely, or address potential data quality problems at a later time.
This oversight creates monumental problems. Sound business decisions depend upon optimum data accuracy, consistency, timeliness, and completeness. Data integrity is even more important when you deploy advanced analytics. One bad record can dramatically change a conclusion, forecast, or estimate.
Most companies don’t embark on a data quality initiative until after things have gone horribly wrong. A BI solution will only succeed if the underlying data can be trusted. Lack of a solid data quality management plan as part of a BI initiative will lead to poor results, and may actually do more harm than good.
Advanced analytics and BI tools are quite reliant on the “strong fundamentals of data capture, cleansing and governance.” According to a recently published Aberdeen research brief, best-inclass companies are three times more likely to adopt data quality tools, a decision that directly with “increased performance in data analysis, employee efficiency, and the speed and accuracy of business decisions.”
Many companies correct their data integrity problems by cleansing data as it is loaded into a data warehouse, data mart, or other repository, but this approach won’t tackle quality problems at their source. Corrected data is never reconciled with back-end systems, which means quality problems will still exist in real-time operational analytical scenarios.
A comprehensive data quality management solution, embedded directly into the BI environment, will ensure optimum integrity across all enterprise information assets.
Data quality is a must. Who want to view the wrong data and show wrong information. No one! Some BI tools already provide the tool to do ETL things. I think the real process should be conducted before the data enter the data warehouse. So it is a bit late unless you need to analyze the data not from "system", just like excel.
Worst Practice 5: Not Considering Mobile Users in Your BI Strategy
Mobile First is an emerging development practice that eliminates the problems and issues
that occur when IT teams slap a mobile interface onto sophisticated, graphics-heavy websites that were originally developed for desktop formats. In light of the growth trend in mobile consumption – people are now using smartphones more often than laptops – companies must consider the Mobile First approach for their BI applications. If they fail to address the needs of mobile BI consumers, they’ll experience low levels of BI adoption, and ultimately, diminished returns on BI investment.
The needs of mobile users cannot be an afterthought. They must be addressed as plans are being laid out. For example, organizations must take into account smaller screen sizes, bandwidth and connectivity constraints, and consumer-style expectations for ease of use and an engaging and interactive experience. Multiple devices must also be supported to drive BI pervasiveness. Users must be able to access BI content via their smartphone or tablet of choice. A BI solution that forces organizations to build native apps for each type of device in use will drain time, resources, and money.
Don’t ignore the needs of mobile users, or design a BI application for PC users and simply adjust it for mobile access. Organizations must ensure that all BI content is mobile-optimized. They must meet the demands of mobile BI consumers first, and then expand that mobile content to be incorporated into portals, dashboards, and other PC-based BI environments.
The key is to determine which mobile apps will have the greatest impact or generate the most revenue, and develop those first. With the right BI platform in place – one that is truly device agnostic – organizations can even let users choose which content they want to see on their mobile device, effectively using “crowd-sourcing” to determine priorities.
Yes, the mobile revolution is now! No excuse for this. For Indonesian market you need to make sure that the application is support Android. The management are also using Android, not only IOS. Maybe in the future it will changed because I believe Microsoft is on the way to make this change by it's Windows 10.
Worst Practice 6: Ignoring New Data, New Sources, and New Compliance Requirements
Today’s businesses operate in a new world order, one that involves rapidly growing volumes of data generated during increasingly complex transactions. Big data must be properly harnessed to drive business performance and ensure adherence to constantly changing regulatory guidelines.
New sources of information are also coming to light. Social media sites, blogs, e-mail messages, and other communication vehicles contain a wealth of vital, real-time business insight that can’t be obtained through surveys, focus groups, and other traditional forms of sentiment collection and opinion gathering. Companies can stay one step ahead by finding new ways to tap into these types of unstructured data and leverage it for competitive advantage.
But few BI platforms easily adapt to emerging requirements like these. For example, many BI solutions don’t scale, or require huge proprietary hardware appliances to support analysis against large data volumes. Some lack the ability to incorporate unstructured information into the environment or can’t retrieve data from important new sources like Facebook, blogs, and Twitter. And others can’t reconcile information across diverse infrastructures and big data environments to create a single, consistent view of key business information.
Partnering with the wrong vendor will hinder insight for decision-making purposes, and provide an incomplete picture of the state of the business. Organizations that choose a BI platform that can access all the information available to them – no matter how much of it there is, or where it comes from – will achieve true BI success.
Social media is the hot new data source. Organizations must seamlessly collect this data, and reconcile it with other enterprise assets. Furthermore, columnar databases and storage facilities for big data, like Hadoop, didn’t exist five years ago but are critical now. A BI vendor must not only provide these capabilities, but also ensure they are fully integrated into the environment in a cohesive way.
As the company grow, the data inside it will grow to. New data means new insight. Company need to leverage those things in order to keep on the right track. Similar to that case, BI need to be supported by data and new data. BI product need to grow and support new trend and technology.
Some of the worst practices mentioned in this paper may seem like common sense. However, high BI failure rates demonstrate that these worst practices are, indeed, put into effect more frequently than you might think. When trade journalists, vendors, and industry consultants are constantly promoting the “latest and greatest” technology and all its benefits, it’s easy to get caught up in the hype.
But now that you are aware of these six worst practices, you can prevent them from standing in the way of BI success in your organization. You’ll make the right choices, with the right goals in mind, to lay the groundwork for widespread user adoption and rapid, measurable return from your business intelligence investment.
Top Six Worst Practices in Business Intelligence
Artificial Intelligence Projects
Project Center in Chennai