5 Mistakes in Data Discovery that Make You Look Like A Beginner

Anna Glorioso, NegosentroMany business intelligence (BI) applications in use today support data discovery, as well as some with advanced graphics tools for data visualization. Often data discovery is seen as a useful tool for presenting KPIs to business leaders, but requiring IT support on the back end. However, such enterprises fail to see the scope of their own data needs and the broader role BI could play. There are many industries where technology is rapidly changing operations.

What is data discovery?

This is a process for letting end users access centralized data from different sources, combine selected subsets to aggregate measures, and analyze the result set to identify patterns and trends for business insights. With the right tool there’s no need for structured data cubes or models; users can leverage company data without special training or complex procedures.

Graphic components that show key measures are engaging, but can be misunderstood. Data discovery tools producing visual elements only are unlikely to satisfy the different data requirements of the competitive enterprise.

But some data discovery tools may not meet operational demands in addressing real-time problems.

Before deciding on a BI solution, information managers must be able to avoid these mistakes.

  1. Reliance on visual elements

Information disclosure and visualization tools present data in a way that is clear to everyone in the business. As a consolidated information resource for exploration, specific needs within business can investigate and view the data they need to meet objectives.

Visualization makes data less demanding to translate, comprehend, and retain; when raw information is presented through pictures and illustrations, it’s easier to perceive measures, states, causes, and patterns that could hold business value. The simplistic appeal of graphical elements can sometimes mislead, as dashboards won’t answer more complex questions.

If an organization can augment their ROI through empowering more individuals to utilize it, then visual data discovery tools may be the right approach. But they must understand that they have limited use, don’t scale effectively, and can’t fulfill operational needs for continuous data.

  1. Customers are not onboard

Business experts need to be engaged with information technologies that allow them to make choices rapidly and effectively. But that should also include a level of transparency that engages customers. The more extensive the client base, the more important BI selection will become.

BI solutions that broaden how clients interact with their own information can build brand credibility. For example, BI professionals can work to distribute client-specific dashboards and reports to all customers. Non-technical business clients can access this information to answer their own questions. BI experts may utilize more complex data analytics on request.

  1. Over-use of dashboards leads to confusion

It’s always wise to limit most user activity to presentation rather than actual data manipulation. These policies should include rules for dashboard creation. There are many tools that allow dashboards to be customized based on data queries, but this may provide different answers to the same questions.

Data discovery solutions should include managing metadata (data about data), and data integrity processes, such as limiting basic use to read-only permissions. Having a large number of people produce their own data in their own way, without any BI knowledge and for very specific needs, will only lead to confusion.

Date security is also a problem. Fighting cybercrime cost companies $3 trillion worldwide in 2015. IT rarely has direct control over 3rd-party solutions. Allowing applications unrestricted access can lead to data breaches. Hackers are familiar with the weaknesses of many popular applications and can exploit them to gain access to sensitive data. Any BI solution considered must include some means for updated security.

  1. Poor data integrity and inaccurate conclusions

Even the most basic data such as a person’s age and address involves different data types such as number and text that are incompatible with one another. Data must be properly structured before it’s loaded into BI volumes and the integrity of the data ensured.

Data systems can only use clean data stored in a structured format. Data discovery tools can provide a variety of appealing visual elements, but they are only as useful as the quality of the data behind them. Even the simplest dashboards being used to measure performance and impact business management must be accurate and use valid, relevant data.

  1. High costs of memory-restricted data use

Whatever solution you’re looking at, it must be scalable as the business and its information expands. The value of BI is that it can extrapolate results from massive data volumes to meet different challenges. But data is exploding; by 2020 there will be more than six billion smartphones in use. As these volumes of data grow they become more of a drain on IT infrastructure.

BI systems must deal with huge amounts of data. One alternative is to use servers with excessive RAM so that the dataset can be cached there for faster I/O rates. With the relative limitations on the amount of RAM each computer can handle efficiently, this provides less performance improvement now.

In-memory solutions require a lot of costly server hardware. Strategies like multi-node servers may help, but must be regarded in terms of ROI to be viable. For many enterprises today, data volume and network throughput are so critical that they rely on constant archiving of older data.

The right choice in BI

There is a large gap in business value between data discovery tools and fully-featured BI platforms. Data visualization serves a purpose for providing a clear context of certain measures, but it’s limited to a narrow perspective. True BI solutions can offer much more, such as report distribution, both pre-programmed and ad hoc data querying, social media analytics, and forecasting.

A BI platform can also feed the same visualization elements while providing flexible and in-depth solutions for a variety of complex business needs.