1. Data Quality in a World of Excel
When data quality declines, it directly affects the reliability of reports, dashboards, and analyses. After all, reports and dashboards are nothing more than visual representations of that underlying data. The well-known BI principle remains as relevant as ever: garbage in, garbage out.
Many organizations still operate in what could be called the “Excel jungle.” This is understandable: the possibilities are vast, and its flexibility makes it highly versatile. But it is precisely that freedom that carries risks. Manual adjustments, inconsistent data entry, and varying structures can arise relatively easily. Formulas are accidentally overwritten, and version control is a major challenge. Anyone building a dashboard on a foundation of separate spreadsheets is essentially building a house on quicksand. The visualizations may look impressive, but the reliability needed for decisive analyses cannot be guaranteed.
The solution doesn’t start with new systems, but with behavior: establish guidelines for how data is entered and managed in Excel, and raise awareness about the consequences of incorrect use further down the BI chain. Stay focused on the output and maintain regular dialogue with the users of the source files.
2. Insight Ambitions versus Reality
Ambitions for data and reporting capabilities are often high and require the integration of diverse data sources. Consider combining financial figures with customer relationship data: this requires a system landscape in which data from different sources converges in a shared data lake or data warehouse.
When the current situation still relies heavily on standalone systems and files, the leap to fully integrated insights is significant. I’ve seen organizations make various attempts to get their BI project off the ground, and often they got stuck at the same point: the leap from 0 to 100 was simply too big. Whether it involves implementing new systems or adopting a completely new architecture, these changes can take considerable time in a large organization. If your project plan depends on that, you lose momentum and people lose interest.
The solution isn’t to keep focusing on what isn’t yet possible, but rather on what can already be achieved. Start small, deliver value, and work incrementally from there toward the bigger goal.
3. The language barrier between business and IT
A third challenge is communication between the demand side (the business) and the supply side (IT or development). These are two worlds that speak different languages.
When insight is requested, this sometimes results in technically correct solutions that do not functionally align with the actual information needs. Conversely, it is not always clear to the business side what technical limitations or dependencies are at play. Because no one in the department speaks “both languages”—understanding both the process and the data side—this leads to disappointment, frustration, and reports that miss the mark.
Building that bridge starts with translation: converting an information need into a clear functional specification and ensuring that a dashboard actually answers the original question. Project leaders who understand both worlds prevent misalignment in development and ensure more realistic expectation management on both sides.
I joined a BI project when communication was happening directly between the field and information management, with no translation step in between. The result: reports that never materialized, and that no one used to guide decision-making. The breakthrough didn’t come from better tools, but from investing time in understanding the work process behind the information needs. Only then did the dashboards align with the questions that really mattered.
Progress over perfection
The trick is to ensure that BI projects don’t get bogged down in endless meetings, waiting for all prerequisites to be met. Get started with what you have. An initial proof of concept may not yet be a robust solution, but it helps answer key questions: What do we actually want to measure, and what data are we still missing?
By starting small, you’ll see where data quality falls short and which processes need to be refined. These insights are invaluable for developing robust requirements for the steps that still need to be taken. Without these insights, the same questions will resurface later, at a time when making adjustments is much more costly.
Four tips to get started today
-
Start with the annual plan, not the dashboard. A high-quality annual plan includes measurable objectives that can be captured in key performance indicators (KPIs). Identify which of these can already be measured and monitored using the current data landscape, and start there. For the remaining KPIs, determine what’s needed to make them measurable, and prioritize them based on impact.
-
Invest in the translation role. Someone who can translate business requirements into technical specifications prevents reports that no one uses to make decisions.
-
Accept imperfection to gain speed. Don’t wait until the back end (data governance, data warehouses, new systems) is perfectly set up. That takes too long. Dare to start with a pragmatic solution to demonstrate the value of data, and use those experiences to better guide future IT initiatives.
-
Keep an eye on the back end. While you’re making rapid progress on the front end, you must continue to work systematically on building robustness. Data requirements should be part of all applications you plan to purchase. An Excel dashboard is a pragmatic solution for now, but in five years, you won’t want that foundation anymore. Applications must support you by providing access to their data.
The journey toward data-driven decision-making isn’t a sprint from 0 to 100, but a series of small steps in which you balance the urgent need for insight with the reality of system change. Start small, learn quickly, and keep the end goal in mind.
This article was contributed by Mobilee, a consulting firm specializing in digital transformation, strategy execution, and team development, for the readers of Data. Expo. Find more inspiration at www.mobilee.nl or visit Mobilee during Data Expo.
Author: Stijn Verlaak
