A data lake is an intangible asset with growing value for the company

How much value the implementation of BI system will bring to the company ?

This question has always been ambiguous even for us: those who develop data lakes for companies along with Business Intelligence (BI) analytics.

BI, derived from the English word "intelligence," originally referred to obtaining data for more profitable trading on the financial markets of Europe in the late 19th century. Back then, Business Intelligence was translated as business espionage. Over time, BI became an umbrella term encompassing everything related to business data analysis.

The first challenge for assessment is that a BI system is a set of components: from transactional databases to ETL processes, from data lakes and warehouses to reports/dashboards based on KPI methodology.

Modern BI comprises four essential components:

1. ETL Processes: Collecting, transforming, loading, orchestrating data from initial sources.
2. Data Warehouses and Lakes: Storing data in one place with a data model and aggregation or without.
3. Visualization, Dashboards, Reporting: Delivering analytics to business users in the form of interactive panels.
4. Data Science (Intelligent Data Analysis): Statistics, time series analysis, and econometrics methods to identify patterns and forecast future indicators.

The second uncertainty (and the most complex one) lies in the fact that usage scenarios vary from company to company.

So, several approaches to assess the value of BI could be used:

• You could rely on research commissioned by BI system vendors. For example, a study by Forrester on the impact of implementing Power BI concluded that a company's margin profit grew by [2.5%]. The translated report from Forrester can be downloaded here.
• Or you could turn to BAIN's analysis done in 2017, stating that companies with advanced analytics make decisions five times faster, with a higher chance of profit growth.
• You might look at the effects of performance measurement systems such as OKR or the Balanced Scorecard.
• Or you could examine the effect verified through your own experience: cases from similar companies in the same industry. There are many such cases, both from companies themselves and from external analyses.

***

But it's much more crucial to understand the economic approach to assessing the value of BI for a company.

This approach emerged relatively recently and is more thoroughly detailed in Bill Schmarzo's book "The Economics of Data, Analytics, and Digital Transformation: The theorems, laws, and empowerments to guide your organization’s digital transformation."

We will analyze how to determine the value of BI for a specific company and, most interestingly, how to maximize it.

***

To achieve this, we need to go through a series of conclusions proposed in the book, which have also been confirmed by our practical experience.

The first conclusion: data alone does not bring significant added value.If we draw an analogy with oil, as suggested by The Economist, which once referred to data as the "oil of the 21st century," data by itself is valuable in the same way as raw oil was used for ship insulation and more in the past. Almost all companies today accumulate an increasing amount of data, but without additional investments in their analysis, i.e., using them as they are (at most for monitoring), only a small fraction of their value can be obtained.

The second conclusion: the high added value of data lies in predictive analytics.Here it is quite appropriate to consider the "Big Data Maturity Index" proposed in the book:

***

Many companies remain at the first stage because transitioning to the next level is challenging, as it involves a shift from an IT approach to an economic approach to data.

The transition to an economic approach consists of:

1. Shifting from an IT mentality, where data and analytics are used for business monitoring, to a business mentality, where data is used to predict future events or trends with a high degree of probability to prepare for them today.
2. Moving from high aggregation to analyzing all data and searching for new sources and levels of analysis.
3. Collecting not only structured but all types of data.
4. Updating data more frequently and transitioning to real-time analysis.

Note that the greatest benefit to cost ratio comes from points 1 and 2. Points 3 and 4 are not always applicable without restructuring a significant layer of IT architecture. Thus, it is business ideas that allow for much higher added value from data.

Returning to the analogy with oil, if it is processed and refined, the output is more valuable resources, from asphalt to gasoline. Now, after oil processing, it becomes valuable raw material for road construction and for the movement of most types of transport.

The third conclusion: insights must be created for specific use cases.That is, use cases dictate what insights, analytical profiles, and indicators are needed. If a company understands that there is a business initiative that, with the support of current analytics, including predictive analytics, can reduce expenses by X% or increase revenue by Y%, then it is through this business initiative that the value of the BI system for the business can be evaluated.

Characteristics of a strategic business initiative:

• Critical for success
• Documented
• Cross-functional
• Led by senior management
• Measurable financial result (reduce, increase, optimize)
• Execution time defined within the next 12–18 months

The cost of different oil products also varies significantly depending on their use cases, such as a liter of diesel fuel or 98 gasoline.

The fourth conclusion: the multiplier effect is possible if a unified data lake is created for all possible data analysis scenarios.Use cases can change, and over time, new data sources, indicators, and other methods of analysis become necessary. However, if a separate data infrastructure is created for each scenario or set of scenarios each time, the company will not get the maximum benefit from the funds invested.

It's like building a refinery for each oil field instead of transporting them to one.

By connecting data into one organized and supported data lake, the company can significantly increase the return on investment. New use cases can benefit from already collected data. Moreover, thanks to having data gathered together, new scenarios for their use may emerge, which were previously unimaginably expensive. Now, the cost of creating new analytics drops sharply.

***

Now, having created a digital asset in the form of a data lake, we can extract maximum value from the invested funds. And precisely the data lake can be considered an asset of the company.

Assets are resources controlled by the company acquired from past events, from which the company expects economic benefits in the future (this interpretation is contained in the principles of IFRS).

The data lake will bring economic benefits in the long term. The more it is used, the higher its reliability, accuracy, and completeness become, and therefore, its value to the company increases. In other words, the data lake is an asset with a growing value, making it somewhat unusual in finance.

Usually, the value of an asset decreases as it is used. However, what is called depreciation for tangible assets does not occur with digital assets.

Thus, the data lake can be compared to such an important intangible asset as the company's brand.

The final conclusion: analytical modules once again amplify the return on investment from the data lake.By solving a problem in one department, you can use the method of solving it for other departments. In particular, the data lake accumulates data, and the library of analytical modules accumulates solution methods. This allows for faster and better use of the collected data.

Analytical modules are composite, reusable, continuously learning analytical assets that deliver predetermined operational and business results. They help solve detailed predefined business problems: detecting anomalies, predicting consumer purchasing behavior, remaining useful life of assets, retaining disengaging customers, condition-based maintenance, assessing operator skills, and optimization schedule evaluation.

***