This blog post is the second in a series of three about how big data can paralyze organizations from taking action. It was written by Brent Dykes, Director, Data Strategy at DOMO, a Riskonnect partner. This series was originally published as one long article on Forbes.com.
The three common traps individuals encounter when dealing with big data include:
- Data Distrust
- Data Daze
- Analysis Paralysis
While data daze may be less familiar than analysis paralysis or data distrust, you’re likely to have already experienced it personally or seen other people hampered by it. As data becomes more and more pervasive, data daze occurs when individuals become intimidated or overwhelmed by the sheer volume of data placed in front of them. Essentially, data daze is a form of information overload when someone is attempting to consume too many metrics/dimensions, charts, reports and so on at the same time. When an audience’s eyes glaze over during a data-intensive presentation or as they review a complex dashboard, you’re often seeing the effects of data daze.
The human brain can only consume so much information before becoming overloaded. In 1956, psychologist George A. Miller found that most people can only handle 7 chunks of information (+/- 2) at a time. In 2004, psychologist Barry Schwartz introduced the Paradox of Choice where having more options to choose from actually increases our anxiety and interferes with our ability to make decisions as consumers. Similarly, when people are inundated with too much data, they become stuck and unsure of how to move forward—even before any analysis is performed.
If less is more, we’re certainly our own worst enemies when it comes to data sharing and consumption. For example, rather than focusing on a limited set of Key Performance Indicators (KPIs), many organizations amass large collections of purported KPIs. Unfortunately, when every metric is flagged as being important or essential, it waters down the purpose and benefits of having KPIs. When this happens, noise and confusion are introduced rather than creating a strong signal for improving business performance.
In an attempt to be helpful, analysts often pack reports, infographics and dashboards with lots of juicy information and visualizations. However, they don’t always have a clear sense of what the people consuming the data really need or want. Throwing data at a wall to see what sticks is a problematic approach that can induce data daze. In these situations, analysts shouldn’t be guessing what business stakeholders need, and end users should clearly articulate what they’re trying to understand and achieve (which can and will evolve).
As follow-on questions based on the data arise, additional information is often appended to the existing reports and dashboards. However, simply providing more information becomes an easy but sloppy substitute for curating the existing data, removing what’s not relevant and offering up the right information.
A lack of knowledge or training can also contribute to data daze in a couple of ways. When individuals don’t know how to visualize data effectively or tell a compelling data story, the information they share can leave people more perplexed than empowered. As well, when people consuming data don’t understand the metric definitions or how to interpret the numbers, you again run into another roadblock to data adoption. Both of these contributing factors can be addressed through proper training and education—both for people producing and consuming data.
How to overcome Data Daze
- Re-visit your KPIs and supporting metrics to ensure they align with your corporate strategy
- Consolidate and curate your KPIs and supporting metrics to increase their relevance and potency
- Organize sets of supporting metrics under KPIs hierarchically or thematically
- Chunk data into more digestible layers or groupings that can be drilled into by business users
- Strive to simplify and improve data visualizations and report/dashboard layouts
- Train employees on how to properly visualize data, craft data stories and interpret data