In a post by McKinsey from 2013, we can see three main challenges in using analytics for the first time:
- Deciding which data to use (and where you should look outside your organization).
- Handling analytics (and securing the right capabilities to do so).
- Using the insights you’ve gained to transform your operations.
When starting an analytics project, it is important to differentiate between the two main sides of the project: organizational vs. technical. At first, a project manager needs to be allocated in order to build a project plan and get the team aligned. Before diving into the technical details, the analytics project manager should identify all of the relevant stakeholders who are interested in analytics. Mapping stakeholder needs, helps in understanding the overall goals of the project.
Once all stakeholders have been identified, a high level scope of the project can be shaped. In the case of an analytics project it could either be a simple project (like setting up a new instance of Google Analytics), or a robust analytics DWH with consolidated data from several systems (Google Analytics, Totango, Salesforce, Application database, etc.) depending on the needs of the stakeholders.
After the scope was defined and approved by all stakeholders, the most important decision of the project still needs to be made. Determining who will be the owner of the analytics services once the project is complete, will determine the long term success of the project. Without clear ownership a great project can turn into a poor service.
On the operational side, several decisions need to be made. In a BI/analytics services, Ownership is critical for data hygiene. If the data is not clean, you cannot trust the data. And if you do not trust the data, the analytics provides no value and may even cause damage. Analytics systems require an owner, an address that users can contact when they encounter issues. Ownership is more about who should be responsible for the analytics rather than who should actually do the maintenance and service assignments.
After establishing the ownership, a decision needs to be made in regards to how end users will use the system. Should the end users be able to run ad-hoc queries by themselves? Should the end users turn to the business analyst to run queries for them and update them? The answers depend on the nature and culture of the organization. In a small startup, simple analytics should be accessible to all. In the case of an enterprise, there are usually several common roles who will have access to the analytics.
The Analytics roles
The roles described in the post include the Data Scientist, Data Analyst, Data Architect, Data Engineer, Statistician, Database Administrator, Business Analyst and the Database and Analytics team leader. While many companies may look for a data wizard, the infographic helps understanding the small differences among all the new big data roles out there. Before assigning or recruiting people to an analytics project, it is worthwhile to understand those differences.
Data Scientists are more focused on advanced modeling and making the data models repeatable using programming.
The Business Analyst then uses the data models to provide results and insights which lead to actions taken by the business.