How to organize your analytics department to maximize effectiveness (draft)
Centralize vs decentralize model, which is right for your organization
What our data team looks like at the early stage.
Centralize Model
Description:
The team consist of 3 main components
BI Analyst: each BI Analyst responsible for one or multiple domains receive requests from stakeholders or BI manager and individually deliver analytics solution for each domain under supervised of BI Manager
Data Analyst: this role at the beginning is not really formed as a team. DA responsible to provide comprehensive BI solutions and maintain those solutions such as BI tools, cloud services, self-service authorization. DA can be seen as a technical foundation for the whole structure.
BI Manager: Working as a scrum master to receive tasks, filter, allocate and review tasks for both BI and DA before delivering the final results for the business side.
Advantages:
- This model working quite well with the limited requirements and less complicated analysis from stakeholders.
- Members can sync up knowledge between domains when having cross-domain analytics.
- BI manager can well manage the input and output of the team
Disadvantages:
- BI manager can become a bottleneck while at the same time receive and review tasks from both stakeholders and BI analysts.
- The analyst will not have a skill set of dealing with business and lack knowledge about rapid products or business change.
Embedded Model
Description:
Unlike, centralize model, each BI analyst is embedded in each team and working as a member of these teams.
BI Manager: Adviser for analyst and PO from the product team to ensure the best analytic solution for each requirement raised.
BI Squad leads: Communicate between team members to have a general analytic picture of all domains and recommend accordingly both workflow and business senses. Managing and allocated tasks
BI Analyst: Working closely with PO or PM to understand their requirement, get advice from BI Manager and Squad lead and deliver output directly for stakeholders (stakeholders later view)
Advantages:
- Analyst spend more time with business to understand products and receive requirement directly from PO (Product Owner) or PM (Project Manager)
- This model working well with rapid business changes and complex requirements from stakeholders.
- Members sharpen communication and negotiation with stakeholders about their task
- BI manager or BI squad lead can spend less time on review and filter tasks and spend more time on advice and guideline for each BI.
Disadvantages:
- Communication between BI Manager, squad lead, and BI Analyst can be challenging (because each analyst can work individually)
- An analyst may have a difficult time syncing up knowledge between each other while having a cross-domain analytics
How do we work together as a team (between DA and BI)?
For detailed process please refer to my GitHub repositories at bahanh/datateam: datateam structure (github.com)
How DA team work together
For the BI team, tasks requirement including various types, therefore, there is no fixed workflow. I will try to describe when I can create a workflow which is generalized and easy to understand as soon as possible.
Please noted this post is only a draft version.
Thanks for reading, if you have any comments, please leave a comment below.
