In a world of digital ecosystems, data is the raw material that fuels the digital economy. In order to be a part of this value creation, there is a need to have a clear data strategy. How does one go from the traditional approach to collection, storage, and reporting to strategic use of data?
In the past data was a necessary byproduct of existing processes and activities and has slowly through the evolution of functions like CRM gained tactical value in numerous organizations. Data went from being stored in the data warehouse for safekeeping, through the use of business intelligence for reporting and analytics towards real-time use of data in providing customer value.
Another distinction between past and present is the evolution from focusing solely on internal data to looking beyond the boundaries of the organization for relevant data through open APIs. The amount of data at hand has also grown at an exponential rate and is expected to grow even more as everything around us results in some kind of data output. As the amount of data grows at a rapid pace, so does the complexity regarding the use of data.
If things weren’t complicated enough, business-critical data usually correlates with privacy and personal information, meaning that it is imperative to stay compliant to data privacy regulations like GDPR and Schrems II as well as maintain bank-grade security for everything that contains personal data.
The first step towards making sense of this landscape is appointing someone responsible for designing and implementing your companies’ data strategy. Traditionally, data ownership is distributed across organizational silos, but the true value of data becomes actionable when you look beyond organizational structures and view the company as a whole. Ownership of a data strategy should not be based on data origination, but on the ability and necessary skills to make sense and take action based on available data.
However, hiring a bunch of data people and giving them access to play around with the data may be a good starting point in order to get a feel of the raw materials at hand, but will not make the boat go any faster.
Define the aspirations for use of data in your organization. What is the purpose of your data strategy? Are you seeking to gain customer insight and predictions that are more precise? Are you aiming for better operational efficiency through the use of sensory data in your production processes? Are you looking for a way to monetize data through new business models? The various possibilities are limitless and there is no one-size-fits-all when it comes to data aspirations.
According to a survey of more than 300 executives by Bain and company, two-thirds of companies surveyed stated that they were investing heavily in becoming data-driven, and 40 percent of the respondents expected significantly positive outcomes. Despite high expectations, 30 percent of the same executives had no clear strategy for embedding data and analytics in their companies.
While we often recite the success stories of how the best-in-class leverage data to drive profitability and user experience, it is easy to overlook all those who struggle to meet positive returns on their data investments.
A Mckinsey study states that the gap between leaders and laggards when it comes to data utilization is increasing, and embracing a data culture is playing an increasingly important role in succeeding. According to the study, you can’t import data culture and you can’t impose it, and most of all, you can’t segregate it. Data culture is company-wide, not something hidden away in the analytics department.
In order to lay down the groundwork, you should map your existing data sources and have a well-thought-out plan on how to collect, update and store your data. When do you need real-time data, and when is it good enough with a good old-fashioned overnight batch? The freshness of the data should also govern your overall principles of data storage and accessibility. Make sure you make sufficient use of your existing data before starting to collect external data. The more data you have, the bigger the complexity.
Data quality is everything. Poor decisions based on unverified assumptions will lead to unfavorable outcomes; poor decisions based on poor data quality may lead to systemic consequences and thus decimating the organization’s trust in data-driven decisions. There will always be parts of your data collection that is inferior and downright faulty. Use the data you know are correct rather than waiting for 100 percent of your data is going to be 100 percent reliable. As absolute uncertainty is near impossible, make sure to quantify your degree of uncertainty. Be explicit of uncertainty and include levels of confidence where uncertainty is present. Seek out drivers of uncertainty and attempt to run experiments that may shine a light on the root cause(s) of uncertainties.
Data quality needs to be based on a single defined source of truth, and just like the overall data strategy, someone needs to be accountable for data quality and governance of all time. Usually, you can find the answer to the same question from several data sources as most companies have siloed information scattered across departments. Many companies that depend on data have different “data tribes”, where have their own preferred source of information. As a result, the organization risk spending valuable time in arguing which variant of the same data point is more correct and/or attempting to reconcile subtly different versions of a metric that should be universal. Decide upon what should be considered the universal source of truth within the organization and stay true to it. Apply the same methods for data processing and modeling throughout the organization to avoid inconstancy.
If data is the raw materials of your digital strategy, APIs are the pipelines that provide those materials. Make sure your platform can handle data exchange through both internal as well as external APIs. When opening up for external data, cyber security is a crucial aspect.
API security is an increasing concern for companies deploying public-facing APIs, as they provide multiple access points to the underlying infrastructure. A study from cyber security company Imperva reveals that 69 percent of companies have public-facing APIs which offer a route to the sensitive data behind applications.
One of the key challenges with APIs is that well-documented APIs often provide a roadmap describing the underlying implementation of an application. Logic and data structures that otherwise would be buried deep in a company’s architecture. According to a CA white paper on API security, this can give hackers valuable clues that could lead to attack vectors they might otherwise overlook. APIs tend to be extremely clear and self-documenting at their best, providing insight into internal objects and even internal database structure – all valuable intelligence for hackers.
Having a strategy in place is a good start, but as everyone knows by now, there is a reason the term culture eats strategy for breakfast has become so popular, and when it comes to becoming data-driven one should not underestimate the cultural change required to embrace a shift in how data should play I vital role in decision making across all levels of the organization.
Set clear expectations on the purpose of becoming data-driven.The fundamental objective in becoming data-driven should be to take better decisions. According to Harvard Business Review, best performers have top managers that demand who expect that decisions must be anchored in data – not occasionally, but every single time, and often lead through example.
Set business goals before technology and data science goals. A company’s advanced analytics goals should reflect the company’s broader aims, allowing it to amplify its most profitable products, services, and processes. Coca-Cola has been using social listening tools to spot influencers who could help the company promote its brand to key customer groups. In the banking and insurance industry data is often used to assess risk and determine credit worthiness, thus pricing products more accurately.
Go from description to insight, and make sure data usage actually drives decisions rather than spending too many resources in verifying the past and/or describing the present.
Keep it simple. It is easy to get lost in the complexity and realize that everything is connected to everything when going down the data analytics rabbit hole. Like previously mentioned, more data will in many cases lead to more complexity rather than precise results. When starting your journey towards becoming data-driven, be obsessed with the business problem you are attempting to solve and seek the simplest solution for the problem at hand.
More data may lead to increased complexity and uncertainty rather than to drive better decisions. Rather than collecting more data, most companies would gain more from maximizing the yield on their existing data and becoming nimble enough to act on insights quickly. The value of becoming data-driven comes from the action and not the input. An accurate prediction of the future is of little value if all it does is to describe a potential future. Make sure to seek the shortest path from insight to action.
Measure, learn, and adjust. The frequency and speed of feedback have changed with digital, and the ability to test features and do adjustments based on real-time feedback from customer behavior is a valuable competitive advantage. However, make sure to choose your metrics with care. Choose the data points that provide accurate and actionable insights.
For companies that wish to stay relevant in an age of changing customer behavior and increased competition in the digital space, becoming data-driven is crucial. Having the right data and a clear understanding of how to utilize those data enables your organization to make the right decisions based on facts and insights rather than gut feel and assumptions.