Data is a wonderful thing. It can give you confidence in decision-making, clarity about the unknown, and it can bring enlightenment from pattern recognition and analysis. Yet all too often it can mislead us into seeing something that is not true, or providing false hope when we're actually looking at an anomaly.
Still to this day, the most common failure in any data-led initiative is insufficient or poor quality data. For five years in a row, Deloitte's Global CPO Survey has ranked poor quality master data as the biggest barrier facing procurement teams.
So how do we ensure we're building a data strategy on strong foundations to avoid these pitfalls and give us the insights we're looking for?
How Maslow's model helps us understand data
To help us understand how we should approach any project involving data, I'll use the framework Maslow outlines in his Hierarchy of Needs model.
In Maslow's model for understanding what motivates people's actions, physical and safety needs, such as having access to food and water, must be fulfilled before individuals can experience psychological needs such as love, belonging and self-esteem. You can't progress to the next level of fulfilment until the previous stage is satisfied.
The same is true of data. The numbers and figures that make up the data points we collect are the foundation of the pyramid, and the first step in any journey to extracting value.
Without accurate, well-structured data points that can be accurately collected and stored in a secure manner, it's impossible to build an effective data strategy. How can you trust your analysis if it's based on inaccurate data?
When I think about the effectiveness of data, I break it down into a model similar to the one Maslow outlines for human behaviour. In my framework the levels are:
Data is the most basic component and makes up the numbers and figures that need to be collected, labelled, and stored in a database. In themselves, these figures are not useful, but they form the essential building blocks for the entire pyramid.
There is a certain comfort that comes from knowing you have the data, even if at this stage you haven't turned it into something useful.
In order to turn data into information, you need to be able to query the data layer. By answering simple indisputable questions such as "who?", "what?", "where?" and "when?" you turn numbers and figures into insights you can apply to improve processes.
Analytics, pattern recognition, and the ability to combine information from different datasets helps us answer more complex questions such as "how?" and "why?" Now we're transitioning from basic informational needs into the realm of understanding, which allows us to make more informed and sophisticated decisions.
The final layer pools the wealth of knowledge built up over time to produce predictive insights based on things we've seen before. This is the ultimate destination for any successful data strategy.
Most businesses set objectives that are tied to 'Knowledge' and 'Wisdom', such as forecasting risks or responding to shifting trends in the market. We jump to the outcome, a world where we have greater understanding of what is happening and are more able to predict what our next move should be.
Very few of us think about the quality of the underlying data required to inform these insights, where it is stored, and how easily it can be accessed. Yet these fundamental attributes are the difference between insights you can trust and those that mislead. Getting the foundations right is the first, most important step of any successful data strategy.