Organizations of every size and in every vertical have a vested interest in unlocking the power of data. Among other benefits, reliable and timely data delivery will reveal what business processes are (or aren’t) working well, help understand customers’ motivations and behavior, and make internal workflows faster and more efficient.
Harnessing and arranging the data is only half the battle, however. Once it’s in your hands, you also need to transform it into a digestible report with actionable insights that audiences of every technical background can parse. Just as importantly, this system of reporting and analytics must be trusted to be reliable and accurate.
Delivering Trustworthy Data and Analytics
Unfortunately, problems with data reporting plague even the most well-prepared organization. In a recent survey, 73 percent of respondents indicated that they had to wait days or weeks for the insights they required, while almost 70 percent of them frequently encountered questions that their current analytics setup had no response for. Even worse, another survey revealed that only one third of executives surveyed trust the data and analytics generated from their business operations.
Building trust in the data is critical for the success of any information delivery project. If the data is inaccurate, difficult to understand, or improperly shaped for the task at hand, the consumers of the data will lose trust in the data. When that trust is lost, users will find other ways to get the data they need (often by building their own offline data store from other sources).
Here’s a look at five key actions that you need to take in order to start creating reports that you (and they) can believe in.
Establish the Source of Truth
A symptom of an immature information delivery strategy is that there is no single and clear source of truth. This fact cannot be overstated: Unless there is a single trusted source of truth for every key business metric, your information delivery architecture cannot be trusted to deliver accurate, validated data. Defining the gold standard for each measure, KPI, and reference list is a prerequisite for any data delivery initiative. Without this, you’re just spinning your wheels rather than making real progress.
Once the source and/or formula for each key piece of information has been defined, communicating that to the wider audience of data consumers is essential. Just as it is important to define each term and definition, you must include in that messaging which sources are to be used to validate any derivative reports, dashboards, or manual queries. Lacking that, your reports will tell conflicting stories.
Get User Input
You can’t deliver high-quality reports or dashboards without understanding what it is that your users need from the data. Getting user input is critical to the process, but all too often this necessary part of the project is shortened (or skipped entirely) to save time. It’s not enough to try to infer requirements from existing reports or dashboards; spend time with those who will be consuming the data to fully understand their needs.
When gathering user input for data delivery requirements, you must understand the following:
- How do you use this data? Gathering input for information delivery requirements is a business exercise, not a technical one. When you begin gathering input from data consumers, seek first to understand how data consumption helps them do their jobs. With a clearer understanding of their business processes, you’ll be better equipped to deliver a solution that solves their business problem.
- In the current system (reports, dashboards, etc.) for data delivery, what is working well? What is lacking? It’s important to learn from successes and failures of the past. Invest the time needed to understand what is currently being used to satisfy data delivery needs, and how well those components are working.
- What are the business rules? Business rules describe relationships, data quality processes, exception and anomaly handling, and other structures and triggers that help turn raw data into actionable information. It is important to understand what business rules exist (or should exist) on the data, and at what point in the data lineage they should reside. Be aware, though, that non-technical users may be turned off by the term “business rules”. Instead, ask about data exceptions, manual clean-up tasks, and other open-ended topics that can reveal what business rules should be included in report delivery.
Make Things Clear
Context is key when it comes to data analytics. You wouldn’t give your users the number “8” without somehow contextualizing that figure–does it refer to the number of hours, the percent of hourly utilization, or one of a million other possibilities? Is that number good or bad? From where does that number originate?
Any data delivery solution must be built with clarity and transparency. For each data source, make clear its intended meaning, its origin and lineage, as well as any processing or transformation logic that has taken place. By providing this contextual information, you can help users understand the reliability of any given datum and find the source of errors or outliers.
Just as you provide precise and accurate data to your users, the language that you use to describe that information needs to be equally clear. Before you begin analytics and reporting, agree upon the terms that you use to measure outcomes, including metrics and key performance indicators (KPIs).
Even the most fundamental metrics should be defined. Something as basic as “a day” might be assumed to be a 24-hour period, but this definition can vary. For any essential metric, KPI, or descriptive term, be clear about what each represents.
Over time, data delivery solutions will evolve. The business needs will change, processes will get more mature, and occasionally the shape or grain of the underlying data will be modified. To maintain trust in your reporting or dashboard solution, be clear and proactive about any changes.
One of the worst outcomes for such a solution is for the data consumers to abandon it because they do not get consistent answers. Let there be no surprises for your reporting audience when critical changes need to be made. Communicating these changes and their impact requires extra time and effort, but that investment will help protect the trust in the data.