Photo by Karolina Grabowska 

Going Beyond the Tools to Technique

The past few years have witnessed a significant evolution in the data environment, driven by incredible technological advancements, shifting consumer behaviours, and a related heightened appreciation for the power of data-driven decision-making. Most organisations have experienced staggering exponential growth in the volume and variety of data they collect or have access to, often prompting them to completely reevaluate their data strategies. Simply put, organisations are awash with data. In order to survive in this new landscape, businesses have had to respond positively by breaking down their data silos, merging disparate datasets where appropriate (and legal), and embracing real-time analytics, bringing about a more joined-up, holistic view of their data. They have also needed to recognise the importance of ensuring that their data is reliable, and have needed to prioritise data quality through cleaning, validation, and other processes. This necessary shift in attitude toward the data they hold, (alongside the introduction of new regulations such as GDPR) has helped to facilitate cross-functional collaboration and has in many cases enabled organisations to conduct more nuanced and insightful analytics.

Of course, the capabilities of available analytical and statistical tools have also experienced a remarkable surge during this period, helping to identify patterns and extracting insights from data that were once nearly impossible for businesses to obtain. The adoption and use of these tools is seen by many as being synonymous with efficiency and productivity. Consequently, organisations have been keen to integrate such tools into their business processes. While some organisations still only deliver data to others as a “self-service” offering, others are actively seeking to unlock more of the perceived hidden value within their datasets. Analytical and statistical tools, ranging from programming languages like Python and R to specialized visualisation platforms like Tableau and Power BI, have obviously played a pivotal role in this, as well as helping to democratise “data science”. They have empowered business users with varying degrees of technical expertise to independently explore and analyse data, often with a view to reducing dependence on specialised data teams.

However, amid all of this rapid growth, a major problem has emerged—the tendency of businesses to prioritise and value proficiency with analytical tools over a proper understanding of underlying analytical and statistical concepts and techniques. This attitude not only poses risks to the quality of analyses but also has profound implications for businesses, particularly when individuals are labelled as "data scientists" without a comprehensive grasp of the field (see my article on the labelling of data scientists here).

Amid all of this rapid growth, a major problem has emerged—the

 tendency of businesses to prioritise and value proficiency with

 analytical tools over a proper understanding of underlying

 analytical and statistical concepts and techniques. 

The attraction of analytical and statistical tools lies in their ability to simplify complex data processes, making them accessible to a broader audience. However, the temptation to rely solely on the user-friendly interfaces of these tools can lead to a distorted view of “data science”. Social media platforms are awash with “cheat sheets” for people using analytical tools so that they do not feel the need for any methodological training. Similarly, courses on the use of Python, R, SQL, and other coding languages are provided, often at high expense, by others proficient in coding rather than appropriate techniques. Yet these courses are always over-subscribed, and acquiring the skill on a CV is seen as key to advancement. In comparison, training in data literacy, analytical methods, and statistical techniques is often ignored. Even reasonably priced training is not taken up, and related qualifications are usually not valued by the business. The critical thinking skills, statistical methods, and data concepts that form the bedrock of meaningful analysis have been forced to take a backseat, resulting in analyses that, at best, lack depth and nuance but at worst are misleading and damaging. 

Critical thinking skills, statistical methods, and data concepts that

 form the bedrock of meaningful analysis have been forced to take

 a backseat

It is unclear why this is the case and seems to be a phenomenon peculiar to data analysis. If we looked at any other field and prioritised tools over technique we would be seen as naive, yet it is somehow accepted when looking at data. For example, in healthcare, we value the surgeon who has been appropriately trained, is familiar with the latest techniques, and has a good understanding of how the body works. We would not choose the surgeon based on whose scalpel was best, but rather on the skills of the surgeon behind the scalpel. Similarly, as someone who likes to listen to music, I would not listen to the person who had the best, most expensive guitar, but to the person who was the most skilled player, who had the best technique and played most emotively. The mantra is, and always should be “technique before tools”. Anything other than this approach can introduce significant risks to the business. 

The mantra is, and always should be “technique before tools”.

 Anything other than this approach can introduce significant risks

 to the business. 

Businesses that incorrectly prioritise tools over technique risk making important decisions based on superficial analyses. Senior management relying on hastily labelled "data scientists" may inadvertently steer their organisations in the wrong direction, as the analyses guiding their decisions lack the necessary depth and context. The dynamic nature of the data science landscape requires businesses to adapt continually to new challenges. A tool-centric mindset may hinder this adaptability, restricting the ability to learn and apply new analytical techniques. This can leave organisations ill-equipped to navigate the evolving environment and hinder their ability to leverage the full potential of their data. True innovation in data should arise from a deep understanding of statistical techniques and a creative application of methodologies. 

Businesses that prioritise cultivating skilled analysts with a deep understanding of analytical and statistical techniques stand to gain numerous advantages.

To address the imbalance between tools and data understanding, businesses must invest in comprehensive data literacy training for their entire staff (see my article on data literacy here). This goes beyond the realm of data analysts, encompassing all employees and fostering a culture where everyone understands the value of data and how it can be used effectively. Data literacy empowers decision-makers at all levels, enabling them to interpret analyses, question assumptions, and make informed choices based on data-driven insights. Providing the workforce with a solid foundation in data literacy will help enable interdisciplinary collaboration. Departments can work cohesively to analyse and interpret data, breaking down silos and fostering a culture of shared insights.

Data literacy empowers decision-makers at all levels, enabling

 them to interpret analyses, question assumptions, and make

 informed choices based on data-driven insights.

Data storytelling is a crucial component of data literacy (see my article about data storytelling here), translating complex analyses into compelling narratives that resonate with diverse stakeholders. Skilled analysts can use data storytelling to ensure that their insights are not only understood but also drive actionable outcomes. Effective data storytelling can influence decision-makers by presenting data in a way that is easily digestible and aligns with the organization's goals. A well-crafted data story has the power to drive action within an organization. It provides a clear path from analysis to implementation, ensuring that insights translate into tangible results. For example, a financial analyst crafting a suitable data story on cost optimisation can influence decision-makers to implement appropriate changes, ensuring that data insights translate into tangible cost savings for the organisation.

Businesses that neglect the importance of investing in skilled analysts and comprehensive data literacy training are short-sighted in their approach. Inadequate training increases the risk of errors in analyses, leading to flawed decision-making and potential financial losses. Organisations that fail to invest in skilled analysts miss out on opportunities for innovation, competitive advantage, and operational efficiency improvements. Inappropriately using data, especially without a solid understanding of statistical techniques, can result in biased analyses or privacy breaches, leading to reputational damage and legal repercussions.

Businesses that neglect the importance of investing in skilled

 analysts and comprehensive data literacy training are short-

sighted in their approach.

Striking the right balance in data science requires a holistic approach that goes beyond the pull of analytical and statistical tools. Businesses must prioritise the development of skilled analysts and invest in comprehensive data literacy training for all staff members. Those who recognise the value of this investment will not only make more informed decisions, foster adaptability, and drive innovation but also ensure their long-term success in a data-driven world. It's not just an expense; it's an investment that will pay dividends in the form of robust decision-making, a culture of innovation, and a competitive edge in the ever-evolving data landscape.

The journey toward unlocking the true power of data requires a concerted effort to bridge the gap between tools and data understanding. If businesses will embrace this holistic approach, they can pave the way for a future where every decision is informed, every operation optimised, and they achieve sustained success in the data-driven world.