Business-driven data culture - What is the key to success?
Leonardo Mu?oz Mu?oz
Master of Business and Information Technology - Solution Architect - Business Intelligence Professor
In recent years, interest in data analytics at multiple levels of complexity has grown in the business and professional world, from “simple” descriptive analysis to machine learning and artificial intelligence projects. For this reason, data has been called “the new oil” but, unlike this widely and effectively used commodity, data analytics projects still have little impact or value.
Some technology specialized media such as VentureBeat in 2017 indicated that 87% of data analytics projects did not reach production. Consulting and research firms such as Gartner in 2019 estimated that, by 2022, only 20% of data analytics projects would give relevant results for the business. In addition, in its latest report published in March 2023, it indicates that only 44% of data and analytics leaders in organizations consider that their teams provide value to their organizations, this being a fact with high load of subjectivity. These and other relevant data about the success or failure of analytics projects are on the internet, confirming that something is still wrong when approaching and developing these projects.
And yes, there are also multiple specialized and academic publications looking for the holy grail of data analytics, many suggesting improving the culture of data-driven decision-making, improving resources dedicated to projects, educating business collaborators in terms of data analysis and many also suggesting improving the technical and business skills of those responsible for the development and implementation of data projects. As you can see, the responsibility for the success or failure of these projects is distributed and it is not an easy task to identify where we should pay attention.
In this short and informal article I will try to give a few keys that could help answer that initial question of What is the key to success? I do not promise to answer it, but surely you will have a clearer idea to start looking for the correct answer.
I have the hypothesis that we should change the factors order and first talk about business-driven data culture, where technicians and scientists learn about business and apply it to data, before talking about data-driven business culture, leaving to suggest that most of the solution lies in business areas that are literate about technology and data.
In order to do so, we will address the following key points when designing a data analytics project from the technical areas:
1.???The business:
Bearing in mind that every technology project wants to respond to a need from with or without profit business, the first thing any team that wants to approach a data analytics project must do is understand who they work for. Those involved in the design and execution must know and understand the mission, vision, objectives, strategies and tactics, the last three being measurable components to verify their effectiveness within the business. It will also be necessary to be in contact with a basic abstraction of the business model components, such as a CANVAS model that allows detailing a little more the organization areas, their specific objectives and their stakeholders.
2.???KPIs or key performance indicators:
A large part of data analytics is based on counts, sums, percentages, averages, in short, it is based on measurements. The KPIs will indicate what is important to measure within the companies or organizations, these must always be related to the identified business objectives and will be specific for each industry and business area. It is valid to initially assume that the objectives and KPIs are correctly defined by the companies, but, in real life, those responsible for the design of a data analytics solution must be able to suggest improvements in the definition of business objectives and the related KPIs.
3.???Business requirements:
To analyze data, it is necessary to get in touch with those who generate and work day by day with the sources, for this reason it is necessary to have direct communication with the actors in the administrative, operational and managerial areas of the organization. To achieve this, structured interviews must be conducted to understand their interests, processes, used technological tools and challenges set by mentioned business areas. These challenges or problems must be closely related to the objectives and KPIs, in addition, they must allow asking questions with answer in the past, present or future.
领英推荐
For the approach of the business requirements and analysis scenarios derived from these identified challenges or problems, it is advisable to use the data warehouse design structure, where we will identify facts and measures based on the KPIs and dimensions based on the characteristics and context of these measures, finally, obtaining the basic inputs for a multidimensional data model that will allow us to answer the questions raised from the contextualization of the data.
4.???The benefit:
Having clarity of the business model and what is important to measure from it, the real benefit (immediately verifiable) and projected benefit (expected in the future with risk) of data-driven decision making must be specified. As expected, the greatest amount of benefits will be related to money and therefore the team must be able to make financial analyzes that support cost reductions, improvement in resource use efficiency, production increase, sales increase, economic risk reduction, competitive advantages, etc., related to data-driven decision making.
These benefits will be a key element when carrying out the project feasibility analysis.
5.???The work team:
Data analytics projects address quite diverse problems, and this requires diverse or interdisciplinary teams. Therefore, the design and implementation team for a data analytics solution can NOT consist uniquely of data scientists and engineers. Although the need arises to broaden the business knowledge of the technical area, experts in finance, business architecture, project management, among others, will always be necessary to help support the justifications of the project to the interested business areas. Such diversity in the work team will make it easier to connect the worlds of business and technology, allowing to make possible sponsors among the decision-making areas.
6.???The scope and life cycle:
Finally, the scope of the project must be well established and limited to a real objective. In the ideal world of data analytics projects, the complete approach of the organization or business is usually considered, but real life must be less ambitious. Data analytics projects should have a clear focus on business units that require priority, have the potential to generate benefits, and have the necessary data sources. This does not mean that the project is limited to this scope, contrary to this, data analytics projects are dynamic, and their goal is the same goal of the organization, for this reason a maintenance and growth plan for the project must be in place, that includes generation and correction of data sources, dedicated technical and business personnel, technological infrastructure maintenance and plans to address new business requirements.
Building on these key project design points, the data and analytics teams will be tasked with continuing to search for the key to success. This is just the beginning of a long road, but a good start can lead to a good road that will end with a very good story told by data.
?
References: