Data, Dashboards, Data Science, or Artificial Intelligence: Which Holds the Key to Enterprise Success?

Business executives, C-suites, and data leaders often ponder where they should invest their limited funds to create the most value for their company. Before delving into the reasoning behind where the most value lies, let's take a quick look at the practical definitions of each of these topics. The aim here is not to provide theoretical definitions but rather to offer practical definitions that will help us frame the reasoning around these insight vehicles that enable companies to become data-driven enterprises. If you already have a clear understanding of the meaning of these terms and why they are important, you can skip this part and proceed to the end to read my comparison and conclusion. However, I highly recommend reading the definitions, as I have also provided examples of value and how it is created. Also note there are investments required in technology, data literacy, operating model and change management that I plan to cover in my next articles. Let's get started.

DATA - In simple terms, data refers to the raw, cleansed, and transformed information that is directly used in business processes to run them efficiently. For instance, accurate labeling of components across sites and factories, precise linkages of purchase orders with invoices to avoid missing payments, and traceability of spare parts are all examples of utilizing data effectively. The presence of accurate data within and across processes is crucial for executing each step efficiently. Additionally, establishing a data thread across processes is essential as it forms the foundation for generating valuable insights.

DASHBOARDS - Another term for dashboards could be data visualization, self-service analytics, embedded analytics, or simply analytics. In its simplest form, dashboards present summarized data that typically reflects what has already happened, which the analytics community also refers to as descriptive analytics. For example, a procurement officer needs to know the remaining inventory before placing an order, and a sales executive wants to identify stores that have lagged in sales over the past week to determine if promotional activities are needed. Critical decisions are made based on these reports, making them indispensable for any company.

DATA SCIENCE - Data science involves using statistical methods, predictive analytics, and optimization algorithms to solve complex business problems. While the textbook definition may vary, the essence remains the same. For instance, optimizing the level of raw material or finished goods inventory to strike a balance between opportunity loss and stock-outs requires an optimization model. Another example is predicting credit card fraud or delivering relevant offers to customers who are likely to accept them. The likelihood of human experience or a dashboard outperforming an algorithm in this regard is minimal. Getting a few of these models right can give a significant competitive advantage by reducing costs or gradually switching customers from competitors without their immediate notice.

ARTIFICIAL INTELLIGENCE (AI) - This is the last and most advanced category, which, for practical purposes, I would define as algorithms that continuously learn from their environment and feedbacks. For instance, if an algorithm recommends this article to a number of profiles and, based on the read, impressions, likes, and comments, it starts recommending it to profiles that are more likely to provide a positive response. The goal of the algorithm would be to maximize readership and impressions. Companies capable of developing such AI solutions can change the trajectory of the company in both financial and non-financial terms. It is not necessary for an enterprise to have multiple successful AI algorithms in operation, as a few well-implemented ones can make a significant impact. For example, Netflix thrives on recommendation systems, FedEx optimizes routing to manage cost and delivery speed, and airlines employ dynamic pricing to maximize profitable occupancy. It is uncommon for an enterprise to have multiple such algorithms successfully implemented since mastering them requires time and effort.

A Comparison:

If you are a retailer, consumer goods company, or automotive company, where should you focus your investments to achieve the best returns?

When it comes to data, ensuring its accuracy within the business processes is essential. Many companies have already mastered this aspect as the consequences of inaccuracies, such as sending incorrect invoices to customers, paying incorrect salaries to employees, or mislabeling parts used in airplanes, can be severe. However, what truly sets companies apart in creating excellence is their ability to connect end-to-end cross business processes e.g, exposing upcoming demand with supplier stock availability and scheduling production in advance, thus avoiding delivery delays and preventing the accumulation of excess raw materials. This connection, which I like to call the "data thread," is often overlooked by companies. Addressing this challenge is no easy task, but if you are currently undergoing an ERP transformation, you have a unique opportunity to get it right for the future. Running this as one off initiative should only be done based on business case.

Dashboards serve as the decision engine for a company. More than 95% of decisions, regardless of function or industry, rely on key performance indicators (KPIs), charts, and trends. This applies to everyone, from floor managers to CEOs, without exception. Even if you face challenges in establishing a data thread across your business processes, you can still create connected dashboards to gain end-to-end visibility. Dashboards are a must-have for any enterprise; without them, you are not running your business efficiently, and worse, you might not even realize how inefficient you truly are.

When it comes to data science, it is crucial to prioritize certain areas such as consumer growth, pricing optimization, inventory management, supply constraints, consumer experience etc. If you operate in service industries like banking, insurance, or technology, you should also have additional focus on risk management. In the Nordics, most companies have fewer than 10 scalable predictive or optimization solutions in place, with the exception of banks and insurance companies. If executed efficiently, these initiatives can have a compounding effect on growth and operations that the competition will only notice when it's too late, recovery from such a loss might take significant amount of resources and time to regain lost business. To harness the full potential of data science, you need to implement and continuously refine solutions over several quarters to a couple of years in order to see real impact. Unfortunately, many companies give up before these impacts start to manifest. For instance, it is not uncommon to reduce inventory costs by 15-20%, improve the impact of promotional activities by 20-30%, or accelerate consumer growth by 30-40% if algorithms are improvised over a period.

Now, let's discuss the use of AI, particularly self-learning algorithms. I consider AI to be a disruptive and transformative lever that can change the trajectory of a company. It can lead to the development of new business models that scale to billion-dollar units or hyper-scale existing businesses into completely new market segments. For example, Netflix's success hinges on its AI-based recommendation system, and platforms like ChatGPT and DALL-E continuously improve content generation through interactions. The potential of such solutions is yet to be fully quantified, but we are talking about a trillion-dollar opportunity in the coming decade. Should large Nordics and global enterprises build and implement AI models? My recommendation is to leverage the foundational AI models already developed by leading tech companies and customize it to fit your business processes. Developing your own AI models would be a strategic decision that requires approval from the board and CEO. It is a significant undertaking that has the potential to disrupt the market, but it also demands substantial investment and a multi-year journey. Only a few companies in the Nordic region have made such decisions where C-suit are committed to supporting these initiatives.

In conclusion:

Evaluate your dashboards if it provides the right level of details for different personas to minimize intuitive decisions. When done well, this can give you a deep insight that is usually not available if you only measure KPIs. Harness the power of data science by identifying a few strategic areas where predictive analytics and optimization algorithms can make a substantial impact. Develop solutions and continuously refine them over time to maximize the benefits. If you can implement 10-20 data science solutions at scale, there is great chance you are doing better than competition. Consider the potential of AI, particularly self-learning algorithms, and evaluate whether leveraging existing models developed by tech giants aligns with your strategic vision and capabilities. Developing your own AI models is a transformative journey that demands significant resources and time which should be done only with C-suite commitment.

The objective of this discussion is not to provide a one-size-fits-all recipe but to give you a directional framework for thinking about where your company should focus to maximize the return on your investments. I hope this perspective helps you ask the right questions when making your next investment decision.

#nordics #artificialintelliegence #data #datascience #datavizualization #intelligententerprise #accenture #datadriven