The expectations CEOs attach to Data Analytics are very high, McKinsey adviser observes. In fact, most
companies are not yet benefiting as expected. In other words, only about one in twelve - eight percent - of 1,000 managers effectively scales their data analytics solutions. In the paper "Ten red flags signaling your analytic program will fail," McKinsey lists ten typical mistakes.
1. Vision is missing
The management does not develop a vision of what they want to achieve with Data Analytics: Many business leaders lack the basic understanding of the different uses of BusinessAnalytics, Reporting, Predictive Analytics and Machine Learning. Countermeasures: Chief Analytics Officers, Chief Digital Officers, Chief Data Officers - or whatever the executives are titled - hold workshops to coach management. These workshops ideally form the basis for later an "Analytics Academy" in the company.
2. Use Cases not calculated
The company does not calculate what value the first use cases should achieve within a year: enthusiasm is followed by frustration when companies do not operate on concrete numbers. McKinsey advises to set up three to five use cases, which bring in realistically fast results. This is preceded by a detailed analysis of the value chain, from suppliers to sales and after-sales processes.
5. Analytics translator is missing
"Analytics Translators" are missing: McKinsey literally speaks of "Analytic Translators," which every business needs. They come ideally from the business side and act as a mediator between technologists and business economists. Desired candidates for this role come from the company itself, know it, know the market and have a good understanding of mathematical models. Such candidates can train the company in their own academy.
6. Resources not properly involved
Organizationally, the resources around analytics are not properly embedded: If a data scientist complains that his work is wasted, this is probably due to the lack of embedding of data analytics. It belongs in the core business. McKinsey advises on a hybrid model: Parts of the company work agile. These agile teams include business executives and data analysts. Data governance remains a central task, but the agile teams should gradually gain more autonomy.
7. Wrong data cleansing
Companies Misuse Data Cleansing: Companies Distribute Data Cleansing Using the Watering Cane Principle, McKinsey Observes. That was a waste of money. Before starting Data Cleansing projects, it is important to clarify which are the most valuable use cases and what data is needed. At the same time, the data controllers must establish a data ontology (reference system with interference and integrity rules).
8. Matching platforms are missing
The right platforms for the solutions are missing: The importance of architecture as the basis of digital transformation is recognized. McKinsey thinks it's wrong to align them with the Legacys first. New data platforms can coexist with legacy systems. The task of the CIO is to enable the integration of data from different sources.
The management does not develop a vision of what they want to achieve with Data Analytics: Many business leaders lack the basic understanding of the different uses of BusinessAnalytics, Reporting, Predictive Analytics and Machine Learning. Countermeasures: Chief Analytics Officers, Chief Digital Officers, Chief Data Officers - or whatever the executives are titled - hold workshops to coach management. These workshops ideally form the basis for later an "Analytics Academy" in the company.
2. Use Cases not calculated
The company does not calculate what value the first use cases should achieve within a year: enthusiasm is followed by frustration when companies do not operate on concrete numbers. McKinsey advises to set up three to five use cases, which bring in realistically fast results. This is preceded by a detailed analysis of the value chain, from suppliers to sales and after-sales processes.
3. No strategy
Decision makers start use cases without following a strategy: Actionism without consideration also leads to frustration. Before initiating analytics initiatives, decision makers should address the following three questions: What are the threats to Artificial Intelligence (AI) and Advanced Analytics for our business? What opportunities do you offer us to improve your existing business? How can we tap new business opportunities?
4. roles not defined
The data analytics field does not properly define the roles: The term "data scientist" can encompass many activities and responsibilities. Data and digitization managers must develop concrete profiles together with the highest personnel decision maker.
Decision makers start use cases without following a strategy: Actionism without consideration also leads to frustration. Before initiating analytics initiatives, decision makers should address the following three questions: What are the threats to Artificial Intelligence (AI) and Advanced Analytics for our business? What opportunities do you offer us to improve your existing business? How can we tap new business opportunities?
4. roles not defined
The data analytics field does not properly define the roles: The term "data scientist" can encompass many activities and responsibilities. Data and digitization managers must develop concrete profiles together with the highest personnel decision maker.
5. Analytics translator is missing
"Analytics Translators" are missing: McKinsey literally speaks of "Analytic Translators," which every business needs. They come ideally from the business side and act as a mediator between technologists and business economists. Desired candidates for this role come from the company itself, know it, know the market and have a good understanding of mathematical models. Such candidates can train the company in their own academy.
6. Resources not properly involved
Organizationally, the resources around analytics are not properly embedded: If a data scientist complains that his work is wasted, this is probably due to the lack of embedding of data analytics. It belongs in the core business. McKinsey advises on a hybrid model: Parts of the company work agile. These agile teams include business executives and data analysts. Data governance remains a central task, but the agile teams should gradually gain more autonomy.
7. Wrong data cleansing
Companies Misuse Data Cleansing: Companies Distribute Data Cleansing Using the Watering Cane Principle, McKinsey Observes. That was a waste of money. Before starting Data Cleansing projects, it is important to clarify which are the most valuable use cases and what data is needed. At the same time, the data controllers must establish a data ontology (reference system with interference and integrity rules).
8. Matching platforms are missing
The right platforms for the solutions are missing: The importance of architecture as the basis of digital transformation is recognized. McKinsey thinks it's wrong to align them with the Legacys first. New data platforms can coexist with legacy systems. The task of the CIO is to enable the integration of data from different sources.
9. New metrics are missing
There is a lack of analytics quantitative value metrics: all data analytics stakeholders should join forces with the finance team to develop new metrics. They must then apply them.
10. Overlook ethical, social and regulatory issues
Companies overlook potential ethical, social and regulatory issues: McKinsey calls this the case of a company that wanted to get to the bottom of the root causes of absenteeism. Using data analytics, it was possible to see a connection to attributes such as gender and ethnicity of employees. That was not compliant. The data controller must ensure that the procedure always proceeds in accordance with the rules.
There is a lack of analytics quantitative value metrics: all data analytics stakeholders should join forces with the finance team to develop new metrics. They must then apply them.
10. Overlook ethical, social and regulatory issues
Companies overlook potential ethical, social and regulatory issues: McKinsey calls this the case of a company that wanted to get to the bottom of the root causes of absenteeism. Using data analytics, it was possible to see a connection to attributes such as gender and ethnicity of employees. That was not compliant. The data controller must ensure that the procedure always proceeds in accordance with the rules.
0 nhận xét:
Đăng nhận xét