With all the resources companies pour into finding the right business intelligence solution, it might come as a surprise that the most prevalent BI implementation challenges actually arise from within those enterprises rather than the solutions they adopt. An ongoing survey conducted by the Business Application Research Center (BARC) lists “lack of resources on project team,” “unclear requirements,” and “data migration” as the top three impediments to BI project implementations, all of which exist independent of the software being deployed.
Exago Senior VP Operations Mark Selinger says that, in his experience, these concerns crop up when companies aren’t decisive in their move to adopt BI. “The product could be a perfect fit,” he explains, “but if there isn’t clear communication within the organization about what to do with the product, the deployment process will stagnate.”
“The product could be a perfect fit, but if there isn’t clear communication within the organization about what to do with the product, the deployment process will stagnate.”
Exago helps prepare prospective customers for deployment success during the product evaluation phase by working with companies to develop a proof-of-concept based on their real-world use cases. This helps define the goals and scope of the launch before contract signing, which Selinger says is a crucial first step.
“Properly assessing each and every BI implementation at the outset helps us avoid challenges down the line,” he explains. “It’s our job not just to support customers with professional services and customer success agents, but to help them determine exactly what kind of guidance they’ll need for their first and successive iterations.”
In the spirit of education and expectation-setting, here are the top three reasons BI deployments fail and how to keep them from plaguing your project. (Please note that although BARC’s survey covers all types of BI, we’ll be focusing on scenarios involving embedded BI in SaaS environments.)
Lack of Resources
Exago prides itself on pricing transparency, but a BI solution’s licensing fee isn’t the only factor contributing to its total cost of ownership (TCO). SaaS providers embedding analytics into their products will need to divert both human and financial resources to the project, particularly during deployment.
While developers are working on integration and implementation, other areas of the company should also be preparing for the release. Product support personnel will need to be trained in the solution, as will end users, which means courses and other learning materials may need to be developed. Stock or “canned” reports will need to be built and compiled into a starting library. Client data will need to be primed for business reporting if it hasn’t already, and all servers in the network should be configured to handle the expected traffic. Marketing and sales teams will, meanwhile, need to prepare promotional materials and craft messaging around the feature.
In short, the deployment of a BI solution will impact virtually every corner of a company, so it’s critical for management to treat it like the large-scale operation it is. In practice, this means wrapping up or reprioritizing other major projects before beginning the BI implementation process and anticipating what additional resources — be it software, hardware, staff, contractors, or something else — will be critical to the rollout’s success. Clear communication between stakeholders before deployment will help ensure that the project is greenlit with all necessary backing.
It’s risky to adopt a BI solution before making absolutely sure it’s the right tool for the job, which is why we put so much work into the evaluation process here at Exago. Developing a requirements checklist to guide you and your team through a rigorous proof of concept is critical to the software vetting process.
But selecting the right data analytics application is only the tip of the iceberg when it comes to meeting your project requirements. Miscommunications can result in wasted time and resources, delaying deployment or derailing it altogether. It can be challenging to anticipate what your requirements will be — especially if your company is implementing BI for the first time — so partner with a company that can offer you guidance as you evaluate and continue supporting you through your implementation.
Exago BI Consultant Nick Cortina recommends clarifying requirements by breaking them down into tasks. A requirement like “cloud data connectivity,” for example, is difficult to quantify in terms of resource allocation whereas “get warehouse connection string” is much more concrete. Cortina finds analyzing requirements this way creates a sort of deployment roadmap and helps teams identify possible challenges before they can jeopardize the project’s target release date.
Poor Data Quality
You’ve heard the saying: garbage in, garbage out. An analytics application, unless it’s an end-to-end solution that includes Extract, Transform, Load (ETL) capabilities, isn’t designed to clean data. Sometimes organizations pursue BI without realizing that their data will need to be significantly reworked for optimal reporting.
Poor-quality data can take many forms. It could be poorly modeled and therefore inefficient to query. It could be sprawling, un-aliased, and confusing for users to navigate. It could be riddled with syntactical errors or stale information. It could be in sore need of row-level tenanting or other security measures. Rather than viewing BI as a cure-all for these issues, companies can view it as an opportunity to implement better data management practices.
Circling back to resources, it’s worth acknowledging that ETL and other data preparation methods generally require expert help from technologists who can code in a variety of querying languages. This is a good place to apply that talent, however. Investing in a low-code ETL application can help alleviate some of those responsibilities, but data quality is well worth the human resource allocation.
Remember, the length of time it takes to deploy a BI solution will vary from company to company. BARC finds that implementation time can average anywhere from a few months to the better part of a year, depending on the size of the organization and complexity of its data sources. A lengthy deployment can still be a successful one.
Nevertheless, the speedier the implementation, the sooner companies will start to see a return on their initial investment in the project. “Business intelligence projects can take up to two years or more to show impactful ROI,” says Solutions Review, “and for business stakeholders, waiting this long to see results can play a big role in why organizations abandon analytics after a short period.” This is why clear communication and expectation setting is so crucial to establish at the outset. Companies that adopt BI with a detailed understanding of their goals and a realistic idea of what it will take to accomplish them have an outsized chance of success.