Is the data correct?The number one advice we give to our clients is that they cannot trust their data blindly. In my opinion, data analytics is not about mastering data analytic tools and smoothly presenting and visualising the outcomes of it, but about building a thorough understanding of how the data is created, processed and changed throughout the process. It is essential to understand how the data reflect the underlying business and decision-making processes. In our practice, we see organisations too often base decisions on wrong, illogical or incomplete data. What is found in IT-systems or databases is in many cases an incorrect reflection of what is actually happening. I, therefore, believe that 80% of the analytics team’s work should consist of understanding the data they are working with, the business processes it supports and the decision-making it drives. The remaining 20% is about applying the right tools and analysis techniques. We have seen it in practice: a large company in the aviation industry has developed a culture in which analysts put a high emphasis on understanding the data, and it clearly pays off.
Strategies for better results
To achieve similar positive results, we see three improvements:
1. Dive deeper
The main advice is to continuously worry about the correctness of data, always and ever. Never assume that it is correct just because it is collected by someone from your organisation or with a commonly used software program. You should constantly check the data, understand the underlying logic of it, how it is collected and edited, build-in tests for assuring its accuracy (such as matching bottom-up and top-down results) and validate findings with business owners to check its relevance and reflection of the corresponding business processes.
2. Assign ownership
Most commonly, no one is responsible for consistently collecting and processing data. Someone from IT collects the information, and then the analysts do their magic without being accountable for the correctness of the data. However, data is so important that it should be owned by the business owners itself (and not by the IT-department or the data-analyst). It is a business responsibility, and clear performance metrics should be put in place to support this.
3. Find the right people
Proper analytics also means setting requirements to the people that work in the analytics team. In general, the best analysts are experts working with very complex excel sheets or more sophisticated analytics tools, day in, day out. On top of that, however, it is of great importance that they are pro-active, eager to understand the deeper background of the analyses they are performing, willing to step out of their daily routine to ask the business critical questions about the purpose of the investigation and capable of presenting their findings in senior management meetings. Typically this is a more all-round employee, often with a background in business.
Role of analysts
The third step – having the right people – is probably the most challenging. Finding adequate employees is difficult, but it is even harder to keep them. Being an analyst is not a temporary job, but an actual profession. It can take over a year before analysts understand the business processes and the relevance of the analyses they are doing. By then, they often leave the organisation or move up to a management position elsewhere in the organisation. Companies should find solutions to secure continuity and offer attractive career paths within analytics teams, e.g. by defining different roles (junior, senior, expert) and offering full-fledged positions in management teams.
It is evident by now that only analysing the available data is not enough. It should become second nature to be critical about the data. If analysts would spend around 80% of their time in genuinely understanding the available information and purpose of their analyses, the quality of the analyses and thus the decision-making will improve significantly.