Becoming a data-driven business doesn't happen overnight. It requires focus, and there are no shortcuts. The important thing is to start the journey and make the best use of the experiences gained along the way.
It often requires a maturity journey for companies to become truly data-driven. The journey provides knowledge about where and how data can be used actively in the company and which data to leverage for what purpose. Often, there is a need to prepare employees for them to be comfortable in exploring data and incorporating the data-driven solutions into their everyday lives. Working with data is a process, which takes practice to become experienced, and must continuously be repeated as new technology paves the way for new business opportunities.
For the past 15 years, focus has been on IT technologies making data more accessible and not least presenting data in optimal ways. The focus is on reporting, preferably presented in dashboards, and most often targeting specific functions and audiences. Not least the top management.
Data for these types of tasks are event- or transactional based and collected from ERP and production systems. Using BI solutions, you develop event-based reports that provide insights into how the company has performed, i.e. information on what has happened.
The trend is shifting. Companies move from only using event data, telling what has happened, to use customer experience data and other relevant data, which can be analyzed to be able to say what will happen in the future.
When a company raises the bar and applies data intelligence, it gains an insight into what is about to happen.
New technologies make it possible to collect, combine, and analyze very large amounts of data and see relationships that haven't been possible before. In addition to transactional data, we collect data from sensors (IoT), our clients, and retrieve data from stream services. Using predictive analytics on the composite data foundation, it's possible to identify future events and act on them.
An example is usage of oil in the industry. A lot of information can be read out of the oil and help to identify the need for preventive maintenance of an equipment in which the oil is used. In this case, the proactive action may be to create a maintenance order securing that the oil is replaced due to exposure of high temperatures several times. However, another maintenance order, which inspects the equipment and secure adjustments, could also be required to prevent future high temperatures derived by unintended friction between components in the equipment.
When the company is working with data intelligence at this level, an algorithm predicts the necessary action to be taken. However, at this stage, it still requires a manual decision to act on the presented information.
In the example of the oil, it's the employee who decides whether two maintenance orders are to be created. The first order addresses the oil change, and the second is to have an inspection of the equipment carried out to prevent unintended friction.
As companies move one step further towards working with data intelligence, they begin to automate the identified actions predicted by the algorithm. This is possible because the company trusts its data and the results generated by the predictive analytics algorithm. The benefit derived from automating manual tasks is the opportunity to have the employees perform other or more complex tasks.
In the example of the oil, the employee no longer decides that the maintenance order should be created. It now happens automatically based on the data from the oil test, which is analyzed using predictive analytics. On the other hand, it turns out in this specific test that the confidence rate predicting the inspection of the equipment is too low to be generated automatically why in this case, it's for the employee to decide which actions should be taken.
Another benefit of automating manual tasks is the opportunity to bring more parameters into consideration in the decision-making process than a human being can handle. In this way, the company will be able to make more accurate decisions. Of course, the data results must be monitored, and the data model must be adjusted over time.
In the example of the oil, the company now decides to use a new spare part in the equipment. This doesn't impact the algorithm identifying when the oil needs to be replaced. However, we observe that the lead time for oil changes have prolonged.
The technology, skill requirements, and the possibilities for what and how companies can leverage data, mature over time. The same applies to companies' learning process and experiences with data. The most important thing is to be fully aware of where you are on the Data Maturity Curve, so that goals and means are balanced, when starting a data project. One step at a time on the journey to data intelligence.
The starting point for a good and efficient journey in the fulfillment of data intelligence, is that the company's Subject Matter Experts actively participates in the process together with a data scientist. The Subject Matter Expert must contribute by selecting the data set, which is the starting point for the analytical work. Although, a data scientist is good at analyzing data, it can take a long time to analyze all data correlations and causal relations. Therefore, it makes sense to start with the most important parameters in the data set as selected by the Subject Matter Expert.
The next step is to explore and evaluate the results in the development of the empirical data model. This requires a close relationship between the Subject Matter Expert and a data scientist. Although the Subject Matter Expert, with its many years of experience, has a qualified expectation of what the results should be, the task - and the challenge - is to make this connection visible solely based on available data.
An important benefit of the close partnership between the Subject Matter Expert and a data scientist is the indispensable knowledge being achieved. Companies gain insights into information they didn't know they could get. They see contexts they didn't expect to see. Also, they find that the contexts they are expecting don't exist in the data available.
The decisive step on the journey is to realize the gains on the investments made. It requires a transformation in the company to get the solution launched – and applied in the right way, quickly. A task that requires attention when it comes to required efforts, timing, and budget. It's crucial to keep in mind that transformation isn't just about employee training. This usually requires changes involving suppliers and customers because the solution affects existing workflows, products, and services in the value chain.
THE STRENGTH LIES IN THE PERSONAL MEETING
If you have a business idea and would like to know whether your data supports it, contact Christian Møller Dahl.
ABOUT Rooftop analytics
Rooftop Analytics can be your data partner on your journey to become data-driven. We clarify your data needs and develop algorithms for new products and services. We also support your business transformation and ensure that you reap the benefits of your data project investment.
The right framework for a data project is needed for it to succeed in creating data intelligence.
Bringing data into play opens for new business opportunities – also for small and midsize companies.