Summary: Many businesses adopt Business Intelligence solutions with high expectations–only to be disappointed with the results. While many blame the problem on faulty software, the reality is, a successful BI project relies on a number of important factors. In this article, we explore 5 such factors that commonly hold back BI implementations.
We’ve seen a Business Intelligence (BI) explosion in the last few years. More organizations are adopting BI solutions to get a clear view of their business data.
How quickly is BI growing? From 2013 to 2018, analysts expect BI spending to increase by $100 Billion. The BI industry as a whole is growing at a 16% rate every year.
With all of this growth, you’d expect massive analytical improvements in businesses across the globe. As they invest in BI, their agility and decision making should improve…right?
The problem: For many businesses, BI doesn’t provide the expected results. It doesn’t give them a clear view of their business. It doesn’t offer the instant answers they expect.
In short, their Business Intelligence isn’t very “Intelligent.”
Why? For some, it could be a matter of inflated expectations. They assumed their BI solution could deliver results beyond its capabilities.
For others, the problem varies. The fact is, many critical elements go into a successful BI implementation. If you’re stuck with BI that doesn’t quite meet your needs, or isn’t “intelligent”, here are some of the most common causes:
1. It doesn’t really fit your business
One common underlying problem found in “unintelligent” BI solutions: They’re incomplete. Businesses try to work with a solution that doesn’t really fit their business 100%.
What causes this problem? Miscommunication is a common culprit. Details often get lost in translation (or overlooked) during the requirements gathering phase. When the IT department tries to implement a solution based on incomplete requirements, the solution will never fit the business.
What’s the answer? As explained below, the requirements gathering process shouldn’t be a one-time process. It must be an open dialogue.
“The biggest reason why BI fails is that companies don’t do proper requirements gathering,” says Heather Cole who is President and CEO of Lodestar Solutions. “Often times, IT might interview the business, create a list of requirements, and they start building. BI needs to be a collaborative project between IT and the business. IT is still running projects via a waterfall method, which doesn’t work in the ever changing business environment of today. Both IT and the business should be educated on Agile/Scrum methodologies and all BI projects must be run in some form of agile method, so there is constant communication between the business users and IT.”
2. The data isn’t clean
It’s a problem I see all too often. A business implements a best-of-breed BI solution and expect it to solve their data problems. But, they ignore their data quality.
How important is data quality? Well, imagine that your BI solution is a shiny sports car. It’s expensive, powerful, and comes with all the bells and whistles. Now, imagine that you filled the gas tank with dirty water instead of gasoline. Is that car going anywhere? Of course not.
That’s how important data is to your BI success. Messy, unorganized data creates all types of problems, from slow reporting to incomplete or inaccurate results.
“BI systems depend on good quality data with no way to capture that data automatically,” says Andy Byrne, CEO of Clari. “For example, looking at the sales world again, traditional BI systems pull data from the customer relationship management (CRM) system. But CRM data is notorious for being terrible because sales reps hate taking time from selling to update their CRM. On the other hand, sales reps live in customer meetings and in email. For BI to work in that environment, it needs to automatically capture all the email and calendar information. Systems that do that can show exactly what reps are doing and, more important — because meetings and email exchanges are with customers — how engaged and enthusiastic customers are.”
3. You have blind spots in your data
In any BI project, you must account for two important data differences: The type of data you collect, and the amount of data being collected. Let me quickly explain why that’s important.
The type of data you collect varies depending on where it’s stored. Sometimes data lives across different systems. Sometimes the data lives in a cloud host. These are often stored in different formats. This isn’t typically a big issue, since your BI solution should provide data blending or ETL capabilities to connect this data from multiple sources.
However, you must still account for the differences in the amount of data being collected across various departments. As explained below, these differences can create blind spots, which lead to faulty analysis.
“A key challenge to BI projects is understanding the blind spots of the collected data,” says Heinrich Hartmann, Chief Data Scientist at Circonus. “There will be parts of the company where high volumes of precise information are easy to collect, e.g. sales processes and accounting, whereas data from other parts of the company may be harder to quantify, e.g. human resources. As a consequence, all analysis of this data suffers from a selection bias, that is also commonly found in observational studies in health and social sciences.”
4. It’s not maintained/no one owns it
Many people get involved during the planning, building, and implementation of a BI solution. The IT department, business users, and executives all have input in the process. Sometimes the business even brings in outside consultants to help out.
But, what happens after the solution is rolled out? Is the project over? Not at all. The problem is, once the solution goes live, no one takes charge of its growth and adoption.
Many businesses fall into this trap. They focus so much effort on implementation, but forget about what happens once it’s built. Who owns the solution once it’s live? Who is in charge of ensuring that it grows and adapts to the business? Who will drive education and adoption? As explained below, much of the real BI effort occurs after the solution is rolled out.
“Many organizations use a build it and leave it approach,” says Pavan Singh, Senior Manager, Information Management & Analytics at Firmex. “They use contractors/consultants, and implement a solution in a long period of time. Once the platform is complete, they leave. The company does not spend any additional time on education, mentorship and focus on helping the power users/users utilize the business intelligence tool. Also the company many not spend time on adding more data and reports onto the platform on a regular basis.”
5. Users aren’t using it
“Although some BI environments are designed to be ‘self serve’ they can sometimes be overwhelming and complicated for end users to create/modify reports,” explains Singh. “If they do not have support and help from a competency center, or if it takes a long time and a lot of bureaucracy from IT in order to make changes, then there is a high likelihood that the end users won’t use the BI platform.
User adoption is one of the biggest challenges facing Business Intelligence today. How bad is it? A recent study puts BI adoption rates among employees at 22%. It goes without saying, but a solution that isn’t adopted will never give you the results you expect.
Why are adoption rates so low? It usually boils down to one of three problems:
1. The users don’t like the solution: Sometimes, a BI tool gets selected without input from the users. When it’s finally rolled out, the users find that it’s confusing, or doesn’t meet their needs. When this happens, they will likely revert back to the old way of doing things.
2. The users were not properly trained: Other times, a BI tool is implemented without proper user training. If users don’t understand the tool, and how it will help them, they won’t use it.
3. The users don’t want to change: You’ll run into this problem in every company. A handful of users will resist any new technology. They think that the old way works just fine. To combat this problem, training must be practical. You must not only explain how the tool helps them, you must demonstrate how the tool can solve a problem that they’re facing. For instance, build a sample report they need. Demonstrate its usage in their environment.
These are just a few reasons why your BI solution might not meet expectations, but there are plenty more. If you would like to add anything to this list, I’d love to hear it. Feel free to share in the comments.