9 Ways You're Failing at Business Intelligence
Big data is an unstoppable trend in the business world. More and more executives are coming to understand that business intelligence is essential to making strategic business decisions. However, BI efforts are commonly derailed by poor data practices, tactical mistakes and more. How can you, as a business owner, avoid these common mistakes during your own BI efforts? Check this article to learn how.
Solid business intelligence is essential to making strategic business decisions, but for many organizations, BI efforts are derailed by poor data practices, tactical mistakes and more.
Executives know they need high quality data in order to make sound business decisions. But getting accurate data in a timely, user-friendly format remains a challenge. Sure, there is a vast industry of consultants and vendors with business intelligence (BI) expertise.
How do you know if you are being “led down the garden path”? Is it time for an upgrade to your BI or to launch a new training program? To answer these questions, knowing where others have made mistakes is helpful.
1. Being an ‘order taker’ when building BI systems
“The customer is always right.” It’s a noble sentiment that has done much to improve customer service, especially in retail. But with technology, business users may not always understand what they are asking for. Even worse, they may try to dictate the technical details of the solution.
Implementing what users ask for instead of what they need is a recipe for BI failure. “Successful BI projects require elaborating and managing requirements, as well as the ability to properly validate BI results,” says Wolfgang Platz, founder of Tricentis, which offers a continuous testing platform to companies such as HBO, Toyota and BMW. The “five whys” technique — asking why five times about a single issue to attain greater depth — is one way to understand what users truly need.
2. Cutting testing time and resources
“Move fast and break things” is a key idea in the startup world. Established businesses, too, often have a need for speed. But in that quest to go faster, activities perceived as ancillary can suffer, like testing. Viewing testing as deadweight can lead to significant quality issues, especially if you rely on manual testing. Instead, look to testing and related “ancillary” processes as ways to deliver a higher-quality BI experience.
“Restricting testing, especially the only testing being done is manual, leads to a high number of defects in user acceptance testing that ultimately affect delivery times,” Platz says.
3. Short-sighting broader data integrity matters
Business intelligence tools are excellent at processing, displaying and analyzing data. But what if you are feeding corrupt data in the system? Or better yet: How would you demonstrate to an IT auditor that you have high-quality data guiding your management decisions? Focusing too narrowly on the BI tool and its configuration may mean you will miss these critical details.
“Today, BI isn’t being used only to support better decisions; BI is often embedded into operational processes. If you have errors in your financial or regulatory reporting — which are often supported by data warehouse technologies — BI can help bring those to light. But other processes can still fail. For instance, an insurance company with broker fees that are calculated even just slightly wrong can negatively impact your reputation and then increase customer churn,” says Platz. “Today’s businesses need to have a proactive, automated approach to BI testing to expose data integrity issues as quickly as possible.”
Making mistakes with financial and regulatory data can lead to expensive problems. Poor data quality also wastes money. In 2013, the U.S. Postal Service was unable to deliver more than 6 billion pieces of mail as addressed. That means lost or delayed customer statements, lost marketing opportunities and more.
4. Taking a reactive approach to upset users
No technology professional looks forward to dealing with angry users. System failures and frustration points will happen. Your response to those issues will influence whether your BI initiative succeeds or fails.
“The two biggest mistakes I see BI novices make is focusing too much on delivering requests and not involving end business users in the project,” explains Doug Bordonaro, chief data evangelist at ThoughtSpot, which focuses on search-driven analytics for retail, financial services and other industries. “When customers are yelling at you about long delivery times and service level agreements being missed, it's the obvious place to focus. Getting too involved in daily delivery misses the larger BI picture. Are you giving your customers what they need to make decisions? Do you understand what data they need? Is there a better solution to the actual problem than another report?”
It’s better to triage user complaints based on their relative importance to your overall strategy than to take a drop-everything approach to issues that arise.
5. Pursuing meaningless analysis
When you have powerful tools at your disposal, it is natural to seek an opportunity to put them to use. But business intelligence without direction wastes time. This problem is particularly common among relatively junior professionals.
“Novice and eager BI professionals are at risk of having tunnel vision and perform interesting analysis that are not guided by meaningful questions. The results can often lack a ‘so what’ finding and fail to deliver impactful insights,” says Mark Langsfeld, vice president of analytics at Anexinet, a consulting services firm with expertise in decision support and customer analytics.
Avoiding this mistake takes business knowledge and judgment. Asking yourself “how does this analysis contribute to the company’s goals?” is one way to prevent the “so what” problem. If you are unsure how to connect your analysis to company goals, there are a few fall-back points to consider. How does your BI analysis illustrate ways to improve revenue, cut cost or improve service? Those are perennial concerns that most business leaders have.
6. Assuming that data alone is sufficient
Can “more data” cure all of our business problems? That’s the unspoken assumption lurking beneath many discussions of business intelligence and analytics. Simply throwing data at an executive and hoping for the best is not going to work.
“If data isn’t presented and argued in a compelling way, it is ignored, or trumped by opinion. The value of having an argument and crafting a story component should never be underestimated,” comments Dan Sommer, senior director of market intelligence at Qlik. The implication of a data set might be clear to your frontline analysts. You can’t assume that point will be clear to others who are a few steps removed from the data.
To craft better stories, consider taking inspiration from other fields. In Made to Stick: Why Some Ideas Survive and Others Die, Chip and Dan Heath outline a model to explain what makes ideas “sticky.” Likewise, fiction writers and screenwriters have used the “Hero’s Journey” concept to tell stories for ages. If you want executives and customers to understand, remember and act on your insights, storytelling skills make a difference.
7. Too much trust in BI tools (and not enough in people and process)
Technologists know the right tool can make a tremendous difference. Think back to the first time you used a script to automate a repetitive task. Those early wins encourage you to constantly hunt for new tools to solve business problems. Unfortunately, too much emphasis on your BI tool tends to lead to disappointing results.
“As an industry analyst, I’ve underestimated how difficult it is to get truly broad and appropriate adoption of BI and analytics usage in organizations. Even if the tools are easier and easier to use, there are process, cultural and learning components needed to achieve success. That’s why we need to speak more and more about data literacy as a key component of enabling proper adoption of BI,” says Sommer.
If you are disappointed with your BI program, look beyond technology. For instance, does your staff know how to present data?
8. Ineffective vendor management
Your company may not have a BI department. In that scenario, working with outside experts makes sense. You might ask them to carry out the function of an outsourced service provider or assist on a project basis. In either scenario, you need to understand your vendor and provide oversight, especially with sub-contractors.
“My company didn’t have full control of one particular project where we worked with a sub-contracted vendor to the software company that sold the BI software. We had particular problems with data cleansing and data governance — “junk in, junk out” as my analysts love to point out,” explains Andrew Pearson, president of Intelligencia, a software consulting company based in Hong Kong. “Our analytical models weren't that useful because the data couldn't be trusted. We weren’t fully confident in their data cleansing capabilities and, since this is the baseline for good BI reporting and absolutely necessary for strong modeling, we weren't in a good position for success.”
If you are working with a third party, it is your responsibility to understand the project and who is working on your account. Otherwise, you may end up with a BI disappointment.
9. Ignoring loyalty to mainstream tools like SQL and Excel
Did you know that there are annual Microsoft Excel championships? Take the Microsoft Office Specialist World Championship for example — they have over five hundred thousand competitors and prize money for winners. That’s just one sign of how popular Excel has become in the business world. To a lesser degree, SQL has a wide following in the technology world.
Neglecting the human element in change and adopting new BI tools increases your odds of failure. “You can find plenty of people will SQL and Excel skills, but finding skills for products like Tableau, Qlik, Spotfire, SAS and SAP is more difficult. When you are introducing new software and new ways of doing business, there is an immediate pushback from business users who are used to doing everything in Excel and/or SQL,” explains Pearson.
Making a significant change to BI in an organization has consequences for careers. The art of change management and leadership cannot be ignored in guiding people through the transition.