Systems Thinking and Data Science: a partnership or a competition?

Systems Thinking and Data Science: a partnership or competing interests?

Information is pretty thin stuff, unless mixed with experience. – Clarence Day (1874–1935), American essayist.

Why do organizations get stuck with their data? It is such a fundamental question. Often, this problem can be due to the organization concentrating solely on technology and data. However, organizations can be supported by a synergistic approach by integrating systems thinking with the data strategy and technical perspective. 

Systems thinking is an approach to problem-solving which invests in understanding the system within which a problem or challenge is situated rather than targeting a specific component of the problem.

How can systems thinking and data science solve digital transformation problems?

Information can be hard to find. Understandably, organizations focus on the data and the technology since data retrieval is often viewed as a data problem. Data tends to run in silos throughout the organization. The questions usually focus on who can access the data and who can make the decision. Organizations often consider the ‘right’ time to access the data. As an aside, businesses often talk about ‘real-time’ data but often mean up-to-date data, which could be weekly or even monthly data. There is usually a focus on the correct data, which may be a matter of ensuring which one of the many different data stores is selected that has the data as it may be duplicated across various sources and analytical systems.

Actionable insight is a buzzword, but it is unclear what it means or how businesses can work to achieve the insight. Systems thinking helps organizations consider the proper context of the data. On its own, the data may be comprehensible by the person at the end of this extensive process. 

Systems thinking helps to build consumable data products, and it helps to make the data meaningful and relevant. To be very customer-focused, the organization must continually reflect on their strengths and the customer they serve. For example, for some organizations, no-code is an excellent way to show success early in the journey towards solving digital transformation challenges. No-code and automation solutions can help to accelerate the organization on its journey. There is a particular interest in the growth of Robotic Process Automation (RPA) which uses AI and metaphorical robots to deliver repetitive tasks and make digital transformation a reality. The RPA market may grow to $25 billion in 2025 according to Forrester, and it has the promise of supporting digital transformation through streamlining digital transformation (Reference).

For example, FPT Software, a global IT services and solutions provider, has a product in the automation and RPA space called akaBot, which has gained recognition on Gartner Peer Insights and other global review platforms. In the industry, there has been intensive growth in the need for low-code and no-code in terms of democratizing access within companies to go from simply using data towards a complete digital transformation strategy. Therefore, interacting with systems using minimal technical skills is very beneficial.

How is it possible to enable data-driven decisions in a systems thinking approach?

A crucial aspect of digital transformation is to enable data-driven decisions. The digital part of digital transformation is to arrange a digital foundation where the right person can access trustworthy, correct information at the right time in the proper context. The foundation should be well structured and have essential data quality measures, monitoring and good data engineering practices.

Systems thinking helps the organization frame the problems in a way that provides actionable insights by considering the overall design, not just the data on its own. Otherwise, how can the business move forward on a digital transformation program if they do not understand the issues? It can be easy to translate the problems into data or technology and believe that the organization has the answer. However, analytics is more complex than viewing a chart showing that sales costs have increased by five per cent. However, that five per cent is only part of the story. We have to add in the context to understand what that means in their business context and make a decision based on it.

However, the thrust here is not to diminish data science or data engineering. It is a massive orchestration of pulling data from multiple systems, organizing it, and ensuring that it is all categorized as data governance and policy. Then, there is an analysis that’s doing all that data that can convert it into digestible information to realize a benefit for the end customer. 

Ultimately, the goal is to empower a data-driven business to make meaningful and purposeful decisions. Of course, the findings need to add value, but how do we measure this success? After all, it can sound a bit woolly! The business teams are getting a value framework, which explains how the organization boils down the strategy into measures of success. It also demonstrates value to the team members who are implementing use cases. 

As the initiative travels throughout the organization, the strategy becomes relevant to team members as they can see how it ties to everyday work and a longer-term plan. Measures can be financial, tying in with the business strategy. Data helps organizations to find broad patterns and trends. However, aggregations can miss nuances and details due to removing outliers from the system. Systems thinking is a complementary methodology that can unearth crucial insights, but this happens slowly. It is resource intensive. Data has the power to inform how we describe and understand networks and communities. Analytics can help us to know what they are trying to do and why data is so hard for organizations. 

It can be challenging for organizations to harness the energy of data scientists to move in the right direction. From a data science perspective, it is possible to begin immediately by framing problems regarding machine learning or statistical issues. However, from the systems thinking perspective, it is vital to view it as a way of fixing problems with an impact on helping the user. 

Systems thinking brings people together from both fields to discuss some of the opportunities and challenges. Organizations are complex, which means no one person understands the whole. As a result, they are places where it’s often hard to see the full effect of our actions. Despite the most positive intention, people who try to improve the organization often exacerbate issues because they do not consider the environment properly.

Data strategy in a VUCA environment

VUCA stands for volatility, uncertainty, complexity, and ambiguity; these terms could be relevant to many data-based projects. It is easy to dismiss the idea of strategy when the data lives in a VUCA environment, and some people even dismiss the idea of having any possibility of any strategy implementation. However, this is where systems thinking adds value to the organization by considering it as a whole with moving parts that impact one another.

Data in a volatile environment

Data can be highly volatile in different ways. For example, data may not be appropriately stored on the disk or even be subject to security breaches. There is a technical understanding of data volatility, which measures how quickly data disappears from a system. Additionally, the data could be subject to large swings in one direction. 

Systems thinking would support the data strategy by looking at the organization’s volatile situation and contributing ways to manage the volatility.

Data in an uncertain environment

Data can be hard to predict. In data science, changing trends in data over time can reduce the accuracy of the predictions made by a model. This is known as data drift, and data scientists will monitor models for data drift to ensure that data science models continue to predict accurately.

Data as a complex environment

Data is highly interconnected and richly disseminated throughout the organization. However, it requires significant technical and people resources to manage properly. While data science can help to understand the data and make predictions from it, systems thinking sees the ‘whole’ system rather than just the parts, helping to simplify the circulation of data throughout the organization while adding meaning.

Data in an ambiguous environment

In an ambiguous environment, the relationship between entities is not always clear. Many different data modelling methodologies today help provide clarity in the data that lives in an imprecise, real-world environment.

There can be a cultural impact on developing technologies and data strategy implementation. This environment may impact the business in many different, unseen ways. For example, do data management styles change in response to varying interpretations of ambitious situations? How does the culture impact the ability of the organization to set out training plans for team members, and how does this impact technology adoption?

Systems Thinking and Data Science in Partnership

During the pandemic, the construction industry was severely impacted in a volatile and uncertain environment. To help one organization to move forward, FPT Software worked with one customer in the construction industry to help support the organization with digital transformation and a digital continuity plan for their UK-based, award-winning company. Due to considerable liquidity injection, the company needed to prepare for ten-fold growth expectations for sales and employees. The executive team worked with FPT to determine how to achieve operational excellence while scaling quickly. However, like many organizations, data silos presented obstacles to effective, data-driven decision-making and ultimately inhibited digital transformation efforts.

The FPT collaboration involved a series of design-thinking workshops to identify and prioritize three use cases where the organization could make the most impact: digital boardroom, bill of materials, and site inspection.

FPT worked with AWS, Java and React JS to build a successful AWS foundational layer with multi-account architecture, integrated lakehouse architecture, consolidated billing, centralized audit log, and centralized guardrail. AWS forms the data foundation as a single source of truth to ingest and integrate data from all enterprise systems and fragmented business data.

As a result of the collaborative, systems thinking approach of FPT’s DX Garage™, the engagement enabled the client to experiment and validate their ideas at a minimum budget along a fast execution plan. The client could reap the full benefits of data analysis, making enterprise-wide analysis possible. All the client’s operational teams now have a shared view of company project statuses, material inventory and a single dashboard for issues tracking.

Key Takeaways

The systems thinking approach helped the project’s success for FPT and their client. Considering context was crucial to the project’s success, supported by Value-Driven Analytics. There was an emphasis on understanding the user journey. There were also considerations over the drivers that impact the KPIs to measure success, such as cash flow, increasing scales or reducing the delivery time. 

A critical insight is that it is crucial to have a digital strategy that is entirely in tune with the corporate strategy. FPT used system thinking to focus on impactful case studies to help the client generate value from the new platform. Using system thinking, FPT helped the customer to see success early while achieving results with the technology to meet their growth objectives. Further, the organization has a digital continuity plan to support them in uncertain times. 

To learn more about how FPT Software can help your business, you can find more information here.

Leave a Reply