The road to data-driven business supremacy

How data-driven companies turn data into action and action back into data to empower rapid success

Summary

• Data-driven decision-making must align with clearly defined business priorities
• A sustainable data-driven company culture requires a uniform source of truth
• Analytics can offer a scientific validation process to back up or discredit gut instincts
• Proprietary data sources provide the fuel needed to power a data-driven economy
• Artificial intelligence helps bring insight beyond the present with predictive analysis

We often hear of digital data being referred to as the new oil – the most valuable commodity in the world. However, much like an oil field, data analytics can be very shallow or very deep. By now, most business leaders appreciate the value of this mission-critical asset, and a rising number of them are heavily focused on becoming more data-driven.

Unfortunately, the reality is rather different. While most businesses use data to drive some of their decision-making, data analytics are still not deeply ingrained in their company culture. A recent survey found that global business is investing almost $40 billion per year on technology and services to power data analytics, yet only 72% of C-suite executives in large corporations claim they have not yet succeeded in forming a data-driven business culture.

A truly data-driven company is one that takes full advantage of the potential of its data. In other words, the entire infrastructure of the organisation is wired in such a way that it will turn data into action and action back into data, thereby empowering a cycle of continuous improvement born of informed decision-making.

In this article, we will look at the building blocks of a truly data-driven environment.

Clearly defined business needs

Data-driven culture starts at the (very) top. Companies with strong data-driven cultures tend have top managers who set an expectation that decisions must be anchored in data — that this is normal, not novel or exceptional. They lead through example.

David Waller, Harvard Business Review, February 6, 2020

For many organisations, there has long been a disconnect between business leadership and information technology. It is often said that the departments do not speak the same language. Data analytics is one such example, in that it is usually viewed as a purely technical challenge belonging under the purview of data scientists and other specialists. In other words, leaders do not know or understand the data, while IT departments do not know the business.

This lack of alignment is one of the leading reasons why so many organisations have yet to become truly data-driven. This problem can only be solved with a top-down approach. After all, to unleash the maximum potential of your data, there needs to be a clearly defined business need. In other words, you need to know what you are looking for and which data is important. Business alignment techniques must analyse business strategies and priorities and decompose them into the various ways in which data can help meet strategic goals.

Above all, data analytics requires businesses to determine the context and value of using the information to drive strategic initiatives. This way, they can determine which data is important and how to leverage it in informed and relevant decision-making. Usually, when companies start thinking about data analytics, they will address routine operations like operational and managerial reporting, but this is only the tip of the iceberg. Unleashing the full potential of data requires a deep understanding of what the company needs, before using these requirements to establish data analysis capabilities.

Technology must be implemented with a clearly defined purpose and context, and data analytics is no exception. A top-down approach starts with overarching business goals, which then need to be distilled into individual use cases and other statements of business requirements. Only then can you determine the key performance indicators (KPIs) needed to track performance across those areas. This ultimately tells data analysis teams what data they need to track and which data sources they need to implement.

Uniform source of truth

When employees argue that “my truth is better than your truth,” it’s a sign you’re masquerading as data-driven. Each team may be acting on data, but if they have different information, they are bound to disagree and some may even be misled.

Sudheesh Nair, Harvard Business Review, May 19, 2020

Many organisations still suffer from the silo mentality, in which different departments within the same company largely operate independently. In this sort of environment, information cannot be shared readily between teams, resulting in everyone having their own version of the truth.

Organisational silos usually start at the highest levels of the business, often with competition between senior managers who have different ideas on how to run the company. This ultimately results in different divisions of the company being unable or unwilling to share information and achieve alignment with top-level business goals. These silos can negatively impact workflows and reduce productivity and morale, both of which ultimately have an adverse effect on client experiences.

Successful businesses enable and encourage the free flow of information as such that every department has access to the same data. Agile companies often talk about systems of record (SORs), which is a source of data that is made available to other systems. In other words, all parties to whom the data is relevant will have access to it. This system of record is a canonical source of truth for storing and sharing common data throughout the organisation, as opposed to having everyone use their own data sources and methods of interpreting them.

Establishing a uniform source of truth is critical for refining insight and establishing data-driven decision-making. This allows for more accurate and better-aligned data analytics, without the risk of conflicting information. It fosters a more collaborative and communicative environment, rather than one that is powered wholly by self-interest. After all, the focus must be on outcomes instead of on ownership, which is another reason why becoming data-driven starts with people and business culture, rather than the technology itself.

Strict, evidence-based decision-making

Data-driven decision-making is about acting on facts rather than emotions. That said, it is only natural for business leaders to act on their gut feelings, and there is nothing necessarily wrong with that when focusing on initial areas of investigation. However, it is essential to have a data-based process to validate or disprove those instincts, and this must be applied across every business department without exception. This will allow decision-makers to identify the cause-and-effect relationships of their actions, rather than acting on instinct alone.

Evidence-based decision-making refers to making the most informed decisions possible using the evidence available which, in this case, is the canonical source of truth mentioned earlier. For example, senior executives might spend an hour before the start of a high-level meeting going over detailed summaries of proposals and the supporting facts needed to back them up. This practice should propagate throughout every level of the organisation so that employees from every department can communicate credibly with business leaders in terms that they can relate to.

Leveraging canonical data sources for evidence-based decision-making reduces conflict and helps establish a more collaborative and communicative environment. Opinions may still differ, but they do not end up driving top-level actions by themselves. Instead, having a diverse range of opinions can help identify which areas might be ripe for further investigation. For example, a leader in sales or marketing might have a feeling that their advertising campaigns are failing to get satisfactory returns on investment. In this case, they can delve deeper into the data to back up or otherwise discredit their claims. This, in turn, will reveal valuable opportunities for improvement.

Abundant proprietary data sources

Data is the raw material that fuels insight, and it is proliferating at a truly remarkable rate to the point the global datasphere has now well into the zettabyte scale. Still, the vast majority of this data remains unused for analytics and business intelligence. Moreover, a lot of data has little or no value anyway. The key to data-driven decision-making is not the quantity of data, but the quality of it. Good data is that which is relevant to the business and has been cleansed and enriched in such a way that it realises its true value.

Data broadly exists in two forms – open data and proprietary data. Open data can be used or shared freely for any purpose without restrictions. It is especially valuable for training things like machine learning and artificial intelligence models, but it is not usually a useful resource for business decision-making. This is because it is less likely to be relevant to the company in question.

Proprietary data is where the real value lies, since it is owned and controlled by the company. It concerns their operations, employees, processes, and customers alone, instead of being generalised for use among a limitless audience. In other words, proprietary data sources give organisations access to the features and functions they need to achieve alignment between analytics and business goals and drive meaningful and measurable success.

Examples of proprietary data sets include any source of information an organisation collects itself. For example, a bank may collect loan and other financial data, while a logistics company may collect geolocation data.

In the digital age, data sources are all around, in the form of systems like customer relationship management (CRM), enterprise resource planning (ERP), and financial data. This data should be collected by a universal system of record for cross-referencing and building context, which can usually be achieved through the use of application programming interfaces (APIs). These APIs serve as connectors to facilitate the interoperability between different systems used in a business, thereby breaking down the technical constraints that lead to organisational silos.

An organisation can never have too many proprietary data sources, and there are likely to be many untapped sources worth capitalising on. This is especially the case in traditional sectors, such as high-street and door-to-door sales, which are gradually being transformed thanks to the use of internet-connected smart technologies and other systems that collect data.

Foresight and looking beyond the present

“Data is locked in silos, inaccessible, poorly structured, and most importantly, not organized in such a way as to be used as the fuel that makes AI work. Instead, to reap the benefits of AI, companies need to create something called an ontology, a comprehensive characterization of the architecture of all of its data."

Seth Earley and Josh Bernoff, Harvard Business Review, April 28, 2020

The ability to look beyond the present with predictive analytics is the ultimate power of today’s data-driven businesses. In the age of big data, where data sets have become far too large for human interpretation, artificial intelligence and machine learning models have become critical for in-depth analysis. However, that is only just the beginning.

An organisation that only uses analytics to describe previous and current situations is forever bound to the past. One that has achieved data-driven supremacy, on the other hand, is able to garner valuable foresight by leveraging predictive analytics. By analysing vast amounts of data at machine speed, today’s machine learning algorithms can answer not only the question of ‘where we are now’, but also ‘where we are going to be’.

Unfortunately, a lot of data remains locked in siloed systems or lacks the structure needed to make it available for analysis. To truly reap the benefits of artificial intelligence and predictive analytics, it is necessary to create a flexible and agile environment that can accommodate the large and growing volumes of data, accept data from multiple sources, and manages the data throughout its entire lifecycle.

In reality, though, many organisations still struggle to distinguish between ‘bad data’ and ‘good data’. For example, bad data may be siloed, mislabelled, or fraught with privacy-related issues. Good data, on the other hand, is ready for machine learning, because it has been cleansed and structured in such a way that a machine can understand it.

That said, artificial intelligence has come a long way in recent years to the point it can now help derive insight from the vast amounts of unstructured data. This can help detect underlying trends and forecast the most likely future scenarios based on them.

Final words

Achieving data-driven supremacy can only happen with a strong foundation and a culture shift throughout the organisation. This remains enormously challenging, not least because of how inherently difficult it is to keep up with the constant and rapid proliferation of data. Becoming truly data-driven is not something that can happen overnight, just like any other area of digital transformation. As such, instead of viewing it as a destination, the shift towards data-driven decision-making should be treated like a journey of continuous improvement.