Recently, I did a presentation at a conference, and the topic was non-technical, unlike many other presentations that I do. I thought it would be helpful for you all to know about it, as I spent some time and gathered some information about it combined with the experiences I had with some of my customers. This article and video are split into two: this one will be about things needed to make a successful data analytics team, and the other one will be about the roles needed within the team.
Video
The golden era of technology
My expertise these days is mostly around Microsoft products, so I explain it from these products’ point of view. However, no matter what technologies, services, and products you use, you will still be able to use the concepts mentioned in this article.
I call this age the golden era of technology because there has been no time in the history of analytics when technology has been so helpful, and it is just getting better and better.
Tools such as Power BI are full-functional self-service tools that enable developers and business analysts to build an analytical solution. Combining those into a bigger umbrella of services such as Microsoft Fabric will make it even better, a fully SAASified experience that provides anything that a BI team needs to work with to complete an analytical solution. And AI agents such as Copilot are embedded within these services and products, which makes it even easier for the team. That is, as I mentioned, a golden era of technology that can enhance and reshape the way that the analytical team is built.
Things to consider when building a Data Analytics team
I categorize things to consider in building a data analytics team into these categories:
- Goal
- People and Roles
- Soft skills
- Team size
- Culture
- Standards
- Training
Each of the above categories and topics plays an important role in how the analytics team is built, works with the rest of the organization, moves toward success, and greatly impacts the organization’s decision-making process.
Mission and Goal: Business Centric
The first and foremost important thing in building a data analytics team is to define a goal for it. It is not enough to say we will do data analytics just for the sake of analytics. If you say we are building this data analytics project because we have data everywhere, and let’s integrate them and then develop some analytics and dashboards, that is not a good goal. This might mean that the analytics team spent 1.5 years integrating all the data from operational systems, but the business hasn’t gained any value from their work. The result of such a process usually leads to reducing the budget for the analytics team and sometimes even closing down that team entirely.
A good goal for the analytics team is defined from the business side. Assume that the analytics team is part of a bigger reseller sales business. The goal for the analytics team can be helping the business make informed decisions about whether new branches are needed and, if so, in what places, what new product sales lines would significantly impact the business, and so on. These are meaningful goals for the business. When you have goals like that, the business will support the analytical project.
Roles in the Analytics team
There are many roles involved in building the analytics team, some of which are external (outside of the team but still part of the mission and goal of the team). Here are some of those roles. Depending on the size of the team, you might have some or a combination of these;
- Architect
- Administrator
- ETL Developer
- Database Developer
- Data Engineer
- Data Scientist
- Data Modeler
- Data Analyst
- Project Manager
- Business Analyst
- Tester
- Deployment manager
- Analytics manager, team leader
- Project sponsor
- Application Developer / Programmer
- Consultant
In another article and video, I explain in detail about each role, responsibilities, tasks, etc.
Culture in the organization
The organization’s culture plays a role in defining the structure of the analytics team.
Things that you can ask yourself in this area are things such as below;
- Does my organization’s culture nurture business analysts in other departments (HR, Sales, Marketing, etc.) to become self-service users?
- Or does my organization’s culture prefer the analytical work to be entirely done by the BI team (or even external vendors), and the rest of the organization just be a consumer of those reports?
Depending o the answer to these, you may have two different approach in building your analytical team. Suppose self-service work is encouraged in the organization. In that case, your team will be able to achieve more with the help of self-service users who are also familiar with business processes. Then, you will need to build a set of datasets for them to work with and do their analysis based on it. However, if your organization wants everything to be done by the BI team, your core team will be bigger, and you will need more business analysts to help gather detailed requirements.
The size of the team
There are factors in deciding what is a good team size. Some of these are as follows;
Budget: You may know how to size the team based on the budget allocated to the team and the entire project.
Deadlines and Milestones: Does the project need to be completed within a 6-month timeframe, or is the deadline much later?
The current workload: How busy are the team members when a new project is defined, and how much of their time can be allocated to the project?
Skillset: What are the team’s skills? Do they need upskilling? Based on the deadlines and budget, it might be better to add someone with that skill to the team instead.
Culture: Would you have self-service users? Or would everything be developed by the analytics team?
There are some other factors, too. As the project progresses, these things might change. You may have more or less budget, deadlines, and priorities change, and as a result, the size of the team might change.
Standards, Guidelines, and Templates
Having standards and guidelines is essential for the team to work as a unit together and achieve a goal together. These can be formal documents; sometimes, they might just be a shared Wiki page in SharePoint online that the team collaborates to build. Here are some of those things;
Governance guide: How will users use the content produced by the analytics team? What are the medallion architecture guidelines? What is the Gold standard, and how is it different from Silver and Bronze? What process must be followed to publish content in a live environment? How can self-service users share their content with the rest of the organization? And many more questions like that are usually answered in a governance guideline.
Standards and Naming conventions: What are documentation methods? How are notes kept from the work done by developers? What are the naming standards to use? Having all of these would help with project resourcing; if someone new is added to the team, they will use these to understand how the work is done and can keep up with the rest of the team.
Templates and Themes: Is there a template for creating specific functions in Power Query? What about a template for visualization? Is there a theme that has company colors plus the fonts for each specific visual in it? Is it shared somewhere that the entire team has access to and is aware of its existence? These are critical to having the same quality of work from everyone on the team.
Meetings, Standups: Whether you use agile daily standups or not, it is still crucial for the team to have regular meetings to understand how the entire project is going, sync with each other, learn from each other’s experiences, get to know the new changes in the standards, and much more.
Training
Whatever the job function and role in the team, there is a need for training for that role. Of course, the training agenda will be different for each role. You cannot have the same training for the self-service users as well as the core developers in the team. You need to have this training agenda documented somewhere so that if someone in the team wants to step into another role, there is a clear definition of what is needed for that role.
The training is not necessarily classroom training—it can be that—but it can also be a SharePoint site with all available resources, such as books, blog posts, videos, etc., that the team found helpful and collected in one place as a learning library.
Training is not a one-off process, either. It has to be ongoing. In this era of technology, if you are not spending time learning, you are staying behind. The technologies won’t wait for you to catch up.
The training agenda must include technical skills, soft skills, and learning about the processes. For example, a Data Scientist needs to learn how to use Notebook and Python in terms of technology but also needs to know how to communicate with the rest of the team, how to present the work, and what is the general process of machine learning in terms of training models, evaluation of the results, changing parameters, evaluating again, and coming up with a reasonable outcome and operationalizing it.
The role of AI Agents (ChatGPT, Copilot, etc.)
It is not realistic and reasonable to fight AI agents. These are not here to take people’s jobs. They are here to help. But it is important to understand their level of help.
There is a reason that Copilot is called Copilot, not the Pilot itself. It is there to help you achieve something faster; however, the final decision, the review of what has been suggested to you, and altering it whenever needed are still your choice. You need to inform your team about it, and this would be the right way to use something like Copilot rather than blindly accepting a complex DAX measure that it suggested to you.
It is also important to teach the team how to use agents like that using prompt engineering and make them available using Copilot in Fabric, Github, Teams, and any other helpful places.
Last but not least, learning how to customize and have your own version of Copilot also helps in some scenarios. Learning how to use Copilot Studio or AI Skills in Microsoft Fabric are some of those methods.
Soft Skills
Whatever your role in the team, either technical developer or business analyst, you will still need these soft skills. The whole team functions much better when every member has these skills.
Communication
Communication is not only for user-facing roles. It is also for the developer who spends all his/her time behind his/her desk from 9 to 5. learning how to communicate with other team members and users outside the team is essential. This includes but is not limited to the following;
- Run a workshop (meeting) with business stakeholders
- Discuss the requirement
- Discuss the obstacles in the project
- Be transparent with users with the project update
- See the risks in advance and communicate as early as possible
- Communicate with other team members and developers
As a technical person, this is a key skill to have to be able to turn the dial to the business side and talk business with end users, but then turn it back to the technical side when talking with other developers in your team
Can-do Attitude
It might be necessary that, for some specific reason, you are asked to do something that is not in your skillset or your job definition. Remember that you are acting as a member of the analytics team, and the goal is to achieve whatever the team is set to achieve. If you are asked to do something like that, then approach it with a can-do attitude. You might get help from others in the team to do it, but at the end of the day, have this attitude that you can do it. Here are just a few of those instances;
- You may need to dig into a C# code
- You may need to read a third-party API documentation
- You may need to connect to a non-familiar analytical source
- You may be asked to produce ideas and suggestions
- You may be asked to embed the report into a custom application
- You may need to understand a source system which has no documentation
Presentation
Again, knowing how to present is important no matter what role you have. It might be the work that you have completed, it might be an issue or risk that you have faced during the project, it might be an idea that helps the team to have a better outcome. The ability to present it and how you present it can make a big difference in the outcome. Here are some of those instances that you may need to present something;
- Explaining ideas and methods
- Demonstrating techniques and solutions
- Conveying the right message
- Presenting to a big or small group
- With demos or without
- Public or for an organization/customer
Summary
Building an analytics team requires a goal, the people in the team and their roles, standards, and conventions, an understanding of the culture, continuous training, soft skills, and some other aspects mentioned in this article. The other article will discuss the details of each role within the team.
Have you had any experience with any of the items mentioned here? Tell me about your experience down in the comments below.
Reza is author of more than 14 books on Microsoft Business Intelligence, most of these books are published under Power BI category. Among these are books such as Power BI DAX Simplified, Pro Power BI Architecture, Power BI from Rookie to Rock Star, Power Query books series, Row-Level Security in Power BI and etc.
He is an International Speaker in Microsoft Ignite, Microsoft Business Applications Summit, Data Insight Summit, PASS Summit, SQL Saturday and SQL user groups. And He is a Microsoft Certified Trainer.
Reza’s passion is to help you find the best data solution, he is Data enthusiast.
His articles on different aspects of technologies, especially on MS BI, can be found on his blog: https://radacad.com/blog.
Thanks. Waiting for the next article.